text stringlengths 454 608k | url stringlengths 17 896 | dump stringclasses 91 values | source stringclasses 1 value | word_count int64 101 114k | flesch_reading_ease float64 50 104 |
|---|---|---|---|---|---|
kig
#include <bogus_imp.h>
Detailed Description
This ObjectImp is a BogusImp containing only a double value.
Definition at line 89 of file bogus_imp.h.
Member Typedef Documentation
Definition at line 98 of file bogus_imp.h.
Constructor & Destructor Documentation
Construct a new DoubleImp containing the value d.
Definition at line 43 of file bogus_imp.cc.
Member Function Documentation
Reimplemented from ObjectImp.
Definition at line 197 of file bogus_imp.cc.
Returns a copy of this ObjectImp.
The copy is an exact copy. Changes to the copy don't affect the original.
Definition at line 58 of file bogus_imp.cc.
Get hold of the contained data.
Definition at line 108 of file bogus_imp.h. 162 of file bogus_imp.cc.
Reimplemented from ObjectImp.
Definition at line 92 of file bogus_imp.cc.
Set the contained data to d.
Definition at line 112 of file bogus_imp.h.
Returns the ObjectImpType representing the DoubleImp type.
Definition at line 269 of file bogus_imp.cc.
Returns the lowermost ObjectImpType that this object is an instantiation of.
E.g. if you want to get a string containing the internal name of the type of an object, you can do:
Definition at line 244 of file bogus_imp.cc.
Definition at line 122 of file bogus_imp.cc.
The documentation for this class was generated from the following files:
Documentation copyright © 1996-2020 The KDE developers.
Generated on Sat Mar 28 2020 07:52:04 by doxygen 1.8.7 written by Dimitri van Heesch, © 1997-2006
KDE's Doxygen guidelines are available online. | https://api.kde.org/4.x-api/kdeedu-apidocs/kig/html/classDoubleImp.html | CC-MAIN-2020-16 | refinedweb | 251 | 54.39 |
I'm basically trying a bunch of online problems that I found on the web and I've been stuck on this guy for 2 hours.
string array1[3] = {"He", "She", "They"};
string array2[3] = {"Ran", "Ate", "Sat"};
srand(time(NULL));
string array1[3] = {"He", "She", "They"};
string array2[3] = {"Ran", "Ate", "Sat"};
srand(time(NULL));
int random1 = rand() % 3;
int random2 = rand() % 3;
cout << array1[random1] << " " << array2[random2];
He Ran, He Ate, He Sat, She Ran, She Ate, She Sat, They Ran, They Ate, They Sat
You have nine possible combinations. Create an array that has the indices
0-
8. Use
std::random_shuffle to shuffle the array randomly. Then, use the elements of the array as indices to the combinations.
int indices[9] = {0, 1, 2, 3, 4, 5, 6, 7, 8}; std::random_shuffle(indices, indices+9);
Complete Program:
#include <iostream> #include <algorithm> #include <cstdlib> #include <ctime> #include <string> int main() { std::string array1[3] = {"He", "She", "They"}; std::string array2[3] = {"Ran", "Ate", "Sat"}; std::srand(std::time(NULL)); int indices[9] = {0, 1, 2, 3, 4, 5, 6, 7, 8}; std::random_shuffle(indices, indices+9); for ( auto index : indices ) { int i = index/3; int j = index%3; std::cout << array1[i] << " " << array2[j] << ", "; } std::cout << std::endl; }
Sample output:
They Sat, They Ran, He Sat, He Ate, She Ate, He Ran, They Ate, She Ran, She Sat, | https://codedump.io/share/jARZOmttQF0b/1/printing-out-all-possibilities-for-each-array | CC-MAIN-2017-09 | refinedweb | 230 | 55.61 |
Some of you may have tried it, but there are still many developers who haven’t used this hidden gem — Sencha Ext.Direct. It was announced in 2009 and named one of the best features introduced in Ext JS 3.0.
In this article, we will introduce you to Ext.Direct and provide examples of an Ext.Direct connector and backend for Node.js.
Why You Should Consider Ext.Direct
In essence, Ext.Direct is a language and platform agnostic technology to communicate between client and server. All calls are initiated from the client-side and offer features that make coding fun again. I love to call it RPC on steroids. The most important feature is that Ext.Direct is bundled with Ext JS, so you don’t need any extra libraries on the client-side to get it running. You also have the ability to retry calls, upload files and batch transactions. You can find the full specs here.
Ext JS is server-side agnostic, so you can consume a wide range of web services. That’s why it’s such a powerful client-side framework. The most popular ways to communicate are REST, regular Ajax calls using JSON or XML data and JSONP. Although these are indeed popular and widely used, there are times when you want a more flexible and framework-friendly stack. This is where Ext.Direct shines.
Ext.Direct is used to call exposed or predefined server-side methods. The beauty is that calling any server-side procedure feels like executing any other JavaScript method locally. For example:
ExtRemote.DXStaff.create({firstName:’John’,lastName:’Smith’});
You don’t have to think about what’s in between and how to treat the backend. It’s just another object with methods you can call and get a response, regardless of whether it’s a data selection from a database or a mathematical computation. Another benefit of Ext.Direct is the very tight integration with data stores and forms. In this case, all you have to do is specify endpoint methods for CRUD actions and load/submit methods for the forms.
With Sencha Touch 2.3, Ext.Direct specs are now fully implemented with support for forms along with CRUD functionality for data stores. The same syntax can be applied for both Ext JS and Sencha Touch. Extensive documentation is provided as part of the following examples and backend connector as well as in the Sencha docs.
Current Server-side Stacks
Now, let’s look at the server-side. You may wonder whether you can leverage Ext.Direct, if your company is running PHP or Java. Ext.Direct comes with a simple implementation in PHP, and there are lots of other implementations available that are supported by our community members. For example, there are Java, Perl, Ruby, .NET, ColdFusion and Python implementations to name a few. Recently, we got a new addition to this growing list — a backend for Node.js.
Ext.Direct Connector and Backend for Node.js
If you plan to adopt Node.js for your backend solution and would like to try Ext.Direct features, you will need the Ext.Direct Connector for Node.js. It can be found in the npm repository here. You can download and install it via npm or grab the source code from github.
To help you get started, we created an extensive set of examples that you can find here. They cover the most common scenarios and widgets that our frameworks support.
In order to make it easier for developers who are not familiar with writing a Node.js backend, we also provided a full-featured web server to serve static content. As for database choices, we supplied both MySQL and MongoDB examples as well as a client for Ext JS 4.2.x and Sencha Touch 2.3.x.
The backend provides the following features:
- API auto-discovery
- Flexible server port
- logging
- Development / Production environment
- Ext.Direct router and static page serving
- File upload
- Session support
- CORS support
- Flexible response headers and age
- Both http and https protocols
After you download the Node.js backend, you will have to configure it. If you’re using MySQL, you will also have to import the DB schema (optional and used only for examples). Unless you plan to make any changes to the server scripts, everything you need can be configured from the two scripts named
server-config.json and
db-config.json. Once you’ve configured them, run this command in the terminal:
npm install
This will install all of the required modules and dependencies.
Let’s take a closer look at config options for Ext.Direct, as all of the others are self-explanatory.
"ExtDirectConfig": { "namespace": "ExtRemote", // Direct namespace "apiName": "REMOTING_API", //Direct API name (both entries should match at client side) "apiPath": "/directapi", // URL endpoint where API will be served "classPath": "/direct", //location of your Ext.Direct classes "classPrefix": "DX", // For advanced cases,You may have multiple class instances in the same folder, this way you can separate and list prefixed classes only "server": "localhost", //RPC server domain "port": "3000", //RPC server port "protocol": "http", // [http|https] "appendRequestResponseObjects": true //if true, every API call will receive full request and response objects as part of the argument list }
Now that your server is configured, you can start it up: issue this command from the terminal (we assume that you’ve already installed the Node.js server):
node server.js
By default, the server is configured to respond to requests via port 3000. If you intend to serve from port 80 (standard http), you have to change that inside the config file and run the server with superuser privileges.
sudo node server.js
If you plan to serve the application from the same server, then place its files inside the
/public folder.
This is what the simplest HelloWorld API method would look like:
var DXHello = { wave: function(params, callback){ callback({ success:true, msg:'Hello world', params: params }); } }; module.exports = DXHello;
You can call this method from both Ext JS and Sencha Touch using the same syntax:
ExtRemote.DXHello.wave(‘Hi!’, function(res){ console.dir(res); } );
This is what the common function signature would look like:
// method signature has 5 parameters /** * * @param params object with received parameters * @param callback callback function to call at the end of current method * @param sessionID - current session ID if "enableSessions" set to true, otherwise null * @param request only if "appendRequestResponseObjects" enabled * @param response only if "appendRequestResponseObjects" enabled */ happyNewYear: function(params, callback, sessionID, request, response){ // Your code here }
We hope you find the examples useful. Try them out and share your feedback with us on the Sencha forum. Happy coding!
Great post, but if someone is interested in a little bit deeper overview, here () I described how to combine Ext.Direct with moongose, some unit test tools and simple ST app.
@dcima We have identified a cause and it will be fixed in 4.2.3
Very interesting topic.
I’m trying your examples and everything it’s working fine except the FormFileUpload on some browsers.
It works ok on Chrome (32.0.1700.76m), Safari (5.1.7.7534.57.2), Opera(12.16.1860)
but on Firefox (26.0) and IE (11.09.9431) has a strange behaviour:
the first time that i select a file from local file system clicking on the “Select Photo…” button produces no effect and the filesubmit event is not fired on server.
Following selections are ok.
I’m using ext-4.2.2.1144 on server (Ubuntu 13.04) with client Windows 7 64 bit SP1.
No problems with the upload examples of the framework (ext-4.2.2.1144/examples/form/file-upload.html).
We’ve translated this blog article into Japanese here:
Also, just so you know this is the Sencha Japan User Group:
Sweet!! I would love to see another post that explains how to interact with models/stores. | https://www.sencha.com/blog/writing-an-ext-direct-backend-in-node-js/ | CC-MAIN-2015-32 | refinedweb | 1,323 | 66.03 |
Oh, what a tangled web
This is the sixth in a series of posts on how to build a LINQ IQueryable provider. If you have not read the previous posts you might want to file for a government grant and take sabbatical because you've got a lot of catching up to do. :-)
Complete list of posts in the Building an IQueryable Provider series.
Broken? How is this possible? I thought you were this bigshot Microsoft uber developer that never made mistakes! Yet, you claim you gave us shoddy code? I cut and pasted that stuff already into my product and my boss says we 'go live' next Monday! How could you do this to me? Sniff.
Relax. It's not that broken. It's just a little bit broken.
Recall in the last post I invented these new expression nodes; Table, Column, Select and Projection. They worked great didn't they? Sure they did! The part that is broken is that I did not handle all the cases where they might appear. The case I handled was the most obvious, when Projection nodes appear on the top of the query tree. After all, since I'm only allowing for Select and Where anyway, the last operation must be one of those. The code actually assumed that to be true.
That's not the problem.
The problem is Projection nodes may also appear inside selector expressions themselves. For example, take a look at the following.
var query = from c in db.Customers
select new {
Name = c.ContactName,
Orders = from o in db.Orders
where o.CustomerID == c.CustomerID
select o
};
I added a nested query in the select expression. This is very different than the queries I wrote before, which were only tabular. Now, I'm basically asking my provider to construct a hierarchy, each object is going to have a name and a collection of orders. How in the world am I going to do that? SQL can't even do that. And even if I wanted to disallow it outright, what happens if someone does write this?
Bang! I get an exception. Not the one I thought though. I guess the code is more buggy than I realized. I expected to get an exception when trying to compile the projector function, since this lovely query should leave a little ProjectionException fly in my selection expression soup. Didn't I claim that it was okay to invent my own expression nodes because no one was going to see them anyway? Ha! Looks like I was wrong. (The actual exception I get is from assembling a bad expression tree because I was mistyping the Projection nodes when I constructed them. That's going to have to be fixed.)
Assuming I fix the typing bug, what am I going to do about these nested projection nodes? I could just catch them and throw my own exception with an apologetic disclaimer that nesting is a no-no. But then I wouldn't be a good LINQ citizen, and I wouldn't have the fun of figuring out how to actually make it work.
So on to the good stuff.
What I want to do when I see a nested ProjectionExpression is turn it into a nested query. SQL cannot actually do this, so I'm going to have to do it or something like it in my own code. However, I'm not going to shoot for a super advanced solution here, just one that actually retrieves the data.
Since the projector function has to be turned into executable code, I'm going to have to swap some piece of code into the expression where this ProjectionExpression node currently is. It's got to be something that constructs the Orders collection out of something. It can't come from the current DataReader since that guy only holds tabular results. Therefore its got to come from another DataReader. What I really need is something that turns a ProjectionExpression into a function that when executed returns this collection.
Now where have I seen something like that before?
Thinking...
Right. That's what my provider already does, more or less. Whew, I thought this was going to be difficult. My provider already converts an expression tree into a result sequence via the Execute method. I guess I'm already half way home!
So what I need to do is add a function to my good ol' ProjectionRow class that executes a nested query. It can figure out how to get back to the provider for me in order to do the actual work.
Here's the new code for ProjectionRow and ProjectionBuilder.
public abstract class ProjectionRow {
public abstract object GetValue(int index);
public abstract IEnumerable<E> ExecuteSubQuery<E>(LambdaExpression query);
}
internal class ProjectionBuilder : DbExpressionVisitor {
ParameterExpression row;
string rowAlias;
static MethodInfo miGetValue;
static MethodInfo miExecuteSubQuery;
internal ProjectionBuilder() {
if (miGetValue == null) {
miGetValue = typeof(ProjectionRow).GetMethod("GetValue");
miExecuteSubQuery = typeof(ProjectionRow).GetMethod("ExecuteSubQuery");
}
}
internal LambdaExpression Build(Expression expression, string alias) {
this.row = Expression.Parameter(typeof(ProjectionRow), "row");
this.rowAlias = alias;
Expression body = this.Visit(expression);
return Expression.Lambda(body, this.row);
}
protected override Expression VisitColumn(ColumnExpression column) {
if (column.Alias == this.rowAlias) {
return Expression.Convert(Expression.Call(this.row, miGetValue, Expression.Constant(column.Ordinal)), column.Type);
}
return column;
}
protected override Expression VisitProjection(ProjectionExpression proj) {
LambdaExpression subQuery = Expression.Lambda(base.VisitProjection(proj), this.row);
Type elementType = TypeSystem.GetElementType(subQuery.Body.Type);
MethodInfo mi = miExecuteSubQuery.MakeGenericMethod(elementType);
return Expression.Convert(
Expression.Call(this.row, mi, Expression.Constant(subQuery)),
proj.Type
);
}
}
So, just like I inject code to call GetValue when I see a ColumnExpression, I'm going to inject code to call ExecuteSubQuery when I see a ProjectionExpression.
I decided I needed to bundle up the projection and the parameter I was using to refer to my ProjectionRow, because as it turns out the ProjectionExpression also gets its ColumnExpressions converted. Luckily, there was already a class designed to do that, LambdaExpression, so I used it as the argument type for ExecuteSubQuery.
Also notice how I pass the subquery as a ConstantExpression. This is how I trick the Expression.Compile feature into not noticing that I've invented new nodes. Fortunately, I didn't want them to be compiled just yet anyway.
Next to take a look at is the changed ProjectionReader. Of course, the Enumerator now implements ExecuteSubQuery for me.
internal class ProjectionReader<T> : IEnumerable<T>, IEnumerable {
Enumerator enumerator;
internal ProjectionReader(DbDataReader reader, Func<ProjectionRow, T> projector, IQueryProvider provider) {
this.enumerator = new Enumerator(reader, projector, provider);
};
IQueryProvider provider;
internal Enumerator(DbDataReader reader, Func<ProjectionRow, T> projector, IQueryProvider provider) {
this.reader = reader;
this.projector = projector;
this.provider = provider;
}
public override object GetValue(int index) {
if (index >= 0) {
if (this.reader.IsDBNull(index)) {
return null;
}
else {
return this.reader.GetValue(index);
}
}
throw new IndexOutOfRangeException();
}
public override IEnumerable<E> ExecuteSubQuery<E>(LambdaExpression query) {
ProjectionExpression projection = (ProjectionExpression) new Replacer().Replace(query.Body, query.Parameters[0], Expression.Constant(this));
projection = (ProjectionExpression) Evaluator.PartialEval(projection, CanEvaluateLocally);
IEnumerable<E> result = (IEnumerable<E>)this.provider.Execute(projection);
List<E> list = new List<E>(result);
if (typeof(IQueryable<E>).IsAssignableFrom(query.Body.Type)) {
return list.AsQueryable();
}
return list;
}
private static bool CanEvaluateLocally(Expression expression) {
if (expression.NodeType == ExpressionType.Parameter ||
expression.NodeType.IsDbExpression()) {
return false;
}
return true;
}
public T Current {
get { return this.current; }
}
object IEnumerator.Current {
get { return this.current; }
}
public bool MoveNext() {
if (this.reader.Read()) {
this.current = this.projector(this);
return true;
}
return false;
}
public void Reset() {
}
public void Dispose() {
this.reader.Dispose();
}
}
}
Now, you can see that when I construct the ProjectionReader I pass the instance of my provider here. I'm going to use that to execute the subquery down in the ExecuteSubQuery function.
Looking at ExectueSubQuery... Hey, what is that Replacer.Replace thing?
I haven't shown you that bit of magic yet. I will in just a moment. Let me explain what is going on in this method first. I've got the argument that is a LambdaExpression that holds onto the original ProjectionExpression in the body and the parameter that was used to reference the current ProjectionRow. That's all fine and dandy though. The problem I have is that I can't just execute the projection expression via a call back to my provider because all of the ColumnExpressions that used to reference the outer query (think join condition in the Where clause) are now GetValue expressions.
That's right, I've got references to the outer query in my sub query; chocolate in my peanut butter. I can't leave those particular calls to GetValue in the projection, because they would be trying to access columns that don't exist in the new query. What a dilema, Charlie Brown.
Thinking...
Aha! I've got it. Fortunately, all the data for those GetValue calls is readily available. It's sitting in the DataReader one object reference away from the code in ExecuteSubQuery. The data is available in the current row. So what I want to do is somehow 'evaluate' those little bits of expressions right here and now and force those sub expressions to call those GetValue methods. I wish I had code that could do that. That would be perfect.
Wait, isn't that what Evaluator.PartialEval does? Sure, but that won't work. Why? Because those silly little expressions have references to my ProjectionRow parameter, and ParameterExpressions are the rule that tell it not to eval the expression. If I could only get rid of those silly parameter references and instead have constants that point to my current running instance of ProjectionRow, then I could use Evaluator.PartialEval to turn those expressions into values! Life would be good.
How to do that? I need a tool that will search my expression tree and replace some nodes with other nodes. Yah, that's the ticket.
Here's something, I call it Replacer. It simply walks the tree looking for references to one node instance and swapping it for references to a different node.
internal class Replacer : DbExpressionVisitor {
Expression searchFor;
Expression replaceWith;
internal Expression Replace(Expression expression, Expression searchFor, Expression replaceWith) {
this.searchFor = searchFor;
this.replaceWith = replaceWith;
return this.Visit(expression);
}
protected override Expression Visit(Expression exp) {
if (exp == this.searchFor) {
return this.replaceWith;
}
return base.Visit(exp);
}
}
Beautiful! Sometimes I amaze even myself.
Okay, great, now I can swap out those nasty references to the ProjectionRow parameter with the real honest-to-goodness instance. That's what the first line in ExecuteSubQuery does. And it only took a few dozen lines of English to explain it. :-)
The second line calls Evaluate.PartialEval. Just what I wanted. The line after that calls the provider to execute! Hurray! Then I throw the results into a List object. Finally, I recognize that I might have to turn the result back into an IQueryable. Weird, I know, but the type of the 'Orders' property in the original query was IQueryable<Order> because that's how IQueryable query operators work, so C# invented the anonymous type using that for the member type. If I try to just return the list, the projector that combines the results together will be none-too-pleased. Fortunately, I already have a facility to turn IEnumerable's into IQueryables; Queryable.AsQueryable.
Wow! It's almost as if someone designed this stuff to work together.
Full disclosure: I did cheat a little bit. I had to modify the Evaluator class. I had to get it to understand my new expression types. I know, I know, I said no one else needed to know about them, but it is my code too, so I guess that's alright. I'll save that one-liner for you to view in the attached zip file. I only post mega-long code snippets, not measly one-liners.
I also had to invent a new CanEvaluateLocally rule for Evaluator to use. I needed to make sure that it would never think it was possible to evaluate one of my new nodes locally.
So now let's take a look on how DbQueryProvider changed
public class DbQueryProvider : QueryProvider {
DbConnection connection;
TextWriter log;
public DbQueryProvider(DbConnection connection) {
this.connection = connection;
}
public TextWriter Log {
get { return this.log; }
set { this.log = value; }
}
public override string GetQueryText(Expression expression) {
return this.Translate(expression).CommandText;
}
public override object Execute(Expression expression) {
return this.Execute(this.Translate(expression));
});
return Activator.CreateInstance(
typeof(ProjectionReader<>).MakeGenericType(elementType),
BindingFlags.Instance | BindingFlags.NonPublic, null,
new object[] { reader, projector, this },
null
);
}
internal class TranslateResult {
internal string CommandText;
internal LambdaExpression Projector;
}
private TranslateResult Translate(Expression expression) {
ProjectionExpression projection = expression as ProjectionExpression;
if (projection == null) {
expression = Evaluator.PartialEval(expression);
projection = (ProjectionExpression)new QueryBinder().Bind(expression);
}
string commandText = new QueryFormatter().Format(projection.Source);
LambdaExpression projector = new ProjectionBuilder().Build(projection.Projector, projection.Source.Alias);
return new TranslateResult { CommandText = commandText, Projector = projector };
}
}
The only thing that changed is my Translate method. It recognizes when it is handed a ProjectionExpression and chooses not do the work to turn an users query expression into a ProjectionExpression. Instead, it just skips down to the step that builds the command text and projection.
Did I forget to mention I added a 'Log' feature just like LINQ to SQL has. That will help us see what's going on. I added it here in my Context class too.
public class Northwind {
public Query<Customers> Customers;
public Query<Orders> Orders;
private DbQueryProvider provider;
public Northwind(DbConnection connection) {
this.provider = new DbQueryProvider(connection);
this.Customers = new Query<Customers>(this.provider);
this.Orders = new Query<Orders>(this.provider);
}
public TextWriter Log {
get { return this.provider.Log; }
set { this.provider.Log = value; }
}
}
Now let's give this new magic mojo a spin.
string city = "London";
var query = from c in db.Customers
where c.City == city
select new {
Name = c.ContactName,
Orders = from o in db.Orders
where o.CustomerID == c.CustomerID
select o
};
foreach (var item in query) {
Console.WriteLine("{0}", item.Name);
foreach (var ord in item.Orders) {
Console.WriteLine("\tOrder: {0}", ord.OrderID);
}
}
Run this and it outputs the following:
Here are the queries it executed: (I used the new .Log property to capture these)
SELECT t2.ContactName, t2.CustomerIDFROM ( SELECT t1.CustomerID, t1.ContactName, t1.Phone, t1.City, t1.Country FROM ( SELECT t0.CustomerID, t0.ContactName, t0.Phone, t0.City, t0.Country FROM Customers AS t0 ) AS t1 WHERE (t1.City = 'London')) AS t2
SELECT t4.OrderID, t4.CustomerID, t4.OrderDateFROM ( SELECT t3.OrderID, t3.CustomerID, t3.OrderDate FROM Orders AS t3) AS t4WHERE (t4.CustomerID = 'AROUT')
SELECT t4.OrderID, t4.CustomerID, t4.OrderDateFROM ( SELECT t3.OrderID, t3.CustomerID, t3.OrderDate FROM Orders AS t3) AS t4WHERE (t4.CustomerID = 'BSBEV')
SELECT t4.OrderID, t4.CustomerID, t4.OrderDateFROM ( SELECT t3.OrderID, t3.CustomerID, t3.OrderDate FROM Orders AS t3) AS t4WHERE (t4.CustomerID = 'CONSH')
SELECT t4.OrderID, t4.CustomerID, t4.OrderDateFROM ( SELECT t3.OrderID, t3.CustomerID, t3.OrderDate FROM Orders AS t3) AS t4WHERE (t4.CustomerID = 'EASTC')
SELECT t4.OrderID, t4.CustomerID, t4.OrderDateFROM ( SELECT t3.OrderID, t3.CustomerID, t3.OrderDate FROM Orders AS t3) AS t4WHERE (t4.CustomerID = 'NORTS')
SELECT t4.OrderID, t4.CustomerID, t4.OrderDateFROM ( SELECT t3.OrderID, t3.CustomerID, t3.OrderDate FROM Orders AS t3) AS t4WHERE (t4.CustomerID = 'SEVES')
Okay, maybe lots of extra little queries is not ideal. Still, its better than throwing an exception!
Now, finally, Select is done. It really can handle any projection. Maybe. :-)
This is the sixth in a series of posts on how to build a LINQ IQueryable provider. If you have not read
So, again you thought I was done with this series, that I've given up and moved on to greener pastures.
Pingback from (item updated).
I noticed that the "lots of extra little queries" require MARS, which kills SQL Server 2000 compatibility. C'est la vie.
--rj
Roger, that's true. However, making it work without MARS would take a bunch of extra work that I'm not ready to explailn. :-)
It interesting though when you realize there are other yet to be released .Net ORM's that have the same limitation.
Matt,
Yup, I know of at least one that needs MARS.
I mentioned the MARS requirement for retrieving associated entities in conjunction with my first serious test of EF/EDM and the EDM Designer last October and got lots of feedback about missing SQL Server 2000 support from readers of the related "Visual Studio Magazine" article:
New Entity Data Model Graphic Designer Prototype,; a link to the article's at the end of the post.
Risorse su Linq to SQL
This is the seventh in a series of posts on how to build a LINQ IQueryable provider. If you have not.
Here are some useful links to LINQ information. Use the comments or write me if you want to add to this
I've recently updated the list of LINQ Providers found on my Links to LINQ page, accessible from the
I've recently updated the list of LINQ Providers found on my Links to LINQ page, accessible from
微软在.NET3.5中推出了LINQ，现在各种LINQProvider满天飞，刚才在老外站点上看到了一份LINQProvider列表，近30多个：LINQtoAmazonLINQto...
I mentioned in a post a little while ago about the various LINQ To projects I had seen, but Charlie Calvert
LINQ Providers LINQ to Amazon LINQ to Active Directory LINQ over C# project LINQ to CRM LINQ To Geo
This is the tenth in a series of posts on how to build a LINQ IQueryable provider. If you have not read the previous posts you'll want to find a nice shady tree, relax and meditate
Офіційні: LINQ to SQL (DLINQ) LINQ to XML (XLINQ) LINQ to XSD LINQ to Entities BLINQ PLINQ Неофіційні
This is the twelfth in a series of posts on how to build a LINQ IQueryable provider. If you have not Here's a list of all the posts in the building
This weekend I’ve built a small application, which queries the “Simpsons” seasons guide data and updates | http://blogs.msdn.com/mattwar/archive/2007/08/09/linq-building-an-iqueryable-provider-part-vi.aspx | crawl-002 | refinedweb | 2,977 | 51.44 |
Jim Hebert wrote: > > fundamental changes more easily. Have you learned about property sheets > > yet? They are the key to success with ZClasses. > > Nope, where shall I go to read the fine manual on those? I'm not sure, but let me just give you enough info to get started. ZClasses have a "property sheets" tab where you can add sheets. Group relevant properties together. Then go to the ZClass "Views" tab and add your property sheets to the list of views of your objects available through the management interface. Voila--not only do your class instances have configurable properties with defaults, but the management interface lets you set them. And the effort required was minimal. > Yeah, this is the part of acquisition that I _get_, but just shy away > from for some reason. I need to get over that, because this is probably > one of the coolest things about acquisition. I also avoid it because its behavior is not completely obvious, thus even if I understand it, someone else who has to maintain my work later may shoot him/herself in the foot with the bullet I've provided! > Thanks so much, I hope others on the list are profiting from this exchange > as much as I am! Thanks for a nice Friday discussion. Shane _______________________________________________ Zope maillist - [EMAIL PROTECTED] ** No cross posts or HTML encoding! ** (Related lists - )
Re: [Zope] Messing with the namespaces stack [Was: (no subject)]
Shane Hathaway Fri, 30 Jun 2000 14:13:21 -0700
- RE: [Zope] (no subject) Jay, Dylan
- [Zope] (no subject) Hamish Lawson
- Re: [Zope] Messing with the namespaces stack [Was:... Jim Hebert
- Re: [Zope] Messing with the namespaces stack [... Shane Hathaway
- Re: [Zope] Messing with the namespaces sta... Jim Hebert
- [Zope] Alternative Rendering (was Re: ... Chris Withers
- Re: [Zope] Messing with the namespaces... Shane Hathaway
- Re: [Zope] Messing with the names... Jeff K. Hoffman
- Re: [Zope] Messing with the n... Shane Hathaway
- Re: [Zope] Messing with the names... Jim Hebert
- [Zope] (no subject)
- [Zope] Tim's Gripe Of The Week --> was: [Zo... Tim Cook
- Re: [Zope] Tim's Gripe Of The Week --> ... Bill Anderson
- RE: [Zope] (no subject) Chris McDonough
- RE: [Zope] (no subject) Chris McDonough
- [Zope] (no subject) Uros Midic | https://www.mail-archive.com/zope@zope.org/msg03114.html | CC-MAIN-2018-51 | refinedweb | 368 | 83.36 |
Building a Download Monitor With Android and Kotlin
Building a Download Monitor With Android and Kotlin
Learn how to use Kotlin to build an Android app that shows a progress dialog with a download percentage.
Join the DZone community and get the full member experience.Join For Free
In this article, we will develop a simple app that will download a file and show a progress dialog with a percentage using Kotlin. Kotlin is getting popular nowadays and is used a lot in Android development. The objective of this article is to show how Kotlin can be used in Android development.
Environment Setup
Download Android Studio 3.0, if you have not installed it, from the Google site. This article uses Android Studio 3.0. The reason I am using Android Studio is that it supports Kotlin already so that you do not have to download it. For earlier versions of Android Studio, you will have to add Kotlin support manually.
Create an Android project with an Empty Activity. At this point, an Android project has been created but there is no support for Kotlin development. We are going to add it next.
Gradle Scripts
Go to your "build.gradle" (module:app) and add the following line in the beginning:
apply plugin: 'kotlin-android-extensions'
We have now added the Kotlin Android extension to the project.
Then, at the end of the script, in the dependencies section, add the following:
compile "org.jetbrains.kotlin:kotlin-stdlib-jre7:$kotlin_version" compile "org.jetbrains.kotlin:kotlin-reflect:$kotlin_version"
Sync with Gradle to build the project. Now you have set up the environment. Let us begin.
Download Tracker
Create a Kotlin class by right-clicking on the project and naming it DownloadActivity. It will create a file with the extension ".kt".
We are not going to spend much time on user interface design. Our user interface will have a single button called "Download" and this will trigger the download process. Please find the layout file, activity_main, below:
<?xml version="1.0" encoding="utf-8"?> <android.support.constraint.ConstraintLayout xmlns: <Button android: </android.support.constraint.ConstraintLayout>
Open the Download.kt file and add the following lines:
class TestActivity : AppCompatActivity() { override fun onCreate(savedInstanceState: Bundle?) { super.onCreate(savedInstanceState) setContentView(R.layout.activity_main); } }
We are extending AppCompatActivty class in Kotlin. The ":" symbol shows that it is extending a class.
Next, we have overridden the onCreate lifecycle method by using the "override" and "fun" keywords. The third line is displaying the user interface with a single button in it.
Next, we are going to handle some UI operations. Upon clicking on the Download button, it will initiate a download process.
But how do we handle that? There are some synthetic functions in the Android Kotlin extension that will initialize the button id as a variable and can be used in the code, as shown below:
downloadBtn.setOnClickListener { }
It is pretty cool, isn't? You do not have to call findByViewId(R.id.......). The Kotlin extension for Android will do it for you. In order to achieve this, we need to import this synthetic API in the code. Add the following line:
import kotlinx.android.synthetic.main.activity_main.*
Now we can add the code when the button is clicked. The above code snippet is trying to achieve what we talked about just now.
downloadBtn.setOnClickListener { // we are going to call Download async task to begin our download val task = DownloadTask(ProgressDialog(applicationContext)); task.execute(""); }
You do not have to worry about null pointer exceptions. If the button element is not present, then the variable name is still available, but you may not able to call any operations on it. It will throw a compile time error saying "object reference not found" when we attempt to use any of the operations.
Download Task
This, as you guessed, is an AsyncTask that will download the file from the internet. At the moment, we have hardcoded the URL. The URL is a PDF file. You can change this value of your own choice. We are going to define the class below using an inner class of Kotlin. Let us do that now:
inner class DownloadTask (var dialog: ProgressDialog) : AsyncTask<String, Sting, String> { }
And we need to override methods like onPreExecute, doInBackground, onPostExecute, onProgressUpdate, etc.
override fun onPreExecute() { super.onPreExecute() } override fun onPostExecute(result: String?) { super.onPostExecute(result) } override fun onProgressUpdate(vararg values: String?) { super.onProgressUpdate(*values) }
Next, we need to display a progress dialog when we begin the download. When we defined an Async Task earlier, we passed an instance of ProgressDialog to its constructor. This type of constructor is called Primary Constructor because it does not contain any code. In the OnPreExecute method, we will set up this class, as shown below:
dialog.setTitle("Downloading file. Please wait.."); dialog.isIndeterminate = false; dialog.max = 100; dialog.setProgressStyle(ProgressDialog.STYLE_HORIZONTAL); dialog.setCancelable(true); dialog.show();
Please observe that we have not assigned the ProgressDialog instance to a variable, but still, it can be accessed from all of the methods of the class. This is because Kotlin treats it as a property implicitly and will be used for property initialization.
Next, we will code for downloading the file using a standard URLConnection API.
Download It Anyway
val url = URL(params[0]); val connection: URLConnection = url.openConnection(); connection.connectTimeout = 30000; connection.readTimeout = 30000; connection.connect(); val length: Int = connection.contentLength; val bis: BufferedInputStream = BufferedInputStream(url.openStream(), BUFFER_SIZE); val data = ByteArray(BUFFER_SIZE); var total: Long = 0; var size: Int? = 0;
The above code opens a URL connection to the site and constructs a buffered input stream object with a default value of BUFFER_SIZE. The BUFFER_SIZE is a final static variable here with a value of 8192. We will see how we can define a final static variable in Kotlin.
In Kotlin, we can use "Companion Object" to define a final static variable, as shown below:
companion object { // download buffer size const val BUFFER_SIZE:Int = 8192; }
We will now move on to the remaining part of the code.
while (true) { size = bis.read(data) ?: break; total += size.toLong(); val percent: Int = ((total * 100) / length).toInt(); publishProgress(percent.toString()); } out.flush();
The second line in the above code snippet reads the data from the stream until it becomes null. Then, it will exit from the loop. The real beauty of Kotlin. The remaining part of the code is calculating the percent it progressed while downloading and updating the progress dialog.
override fun onProgressUpdate(vararg values: String) { super.onProgressUpdate(*values) // we always make sure that the the below operation will not throw null pointer exception // other way is use null check like this // if(percent != null ) val percent = values[0].toInt(); dialog.progress = percent; }
The above code will update the progress bar while downloading. If all goes well, you can see the progress.
Opinions expressed by DZone contributors are their own.
{{ parent.title || parent.header.title}}
{{ parent.tldr }}
{{ parent.linkDescription }}{{ parent.urlSource.name }} | https://dzone.com/articles/download-tracker-with-android-and-kotlin?fromrel=true | CC-MAIN-2020-05 | refinedweb | 1,154 | 50.84 |
Code Inspection: 'require()' is used instead of 'import'
Reports when
require() is used and helps to replace the
require() call with the
import statement.
Enable Convert require() inside inner scopes with Fix all action to convert all
require() calls inside the nested functions and statements when using the Fix all action.
Please note that converting the
require() statements inside inner scopes to the
import statements may cause changes in the semantics of the code. Import statements are static module dependencies and are hoisted, which means that they are moved to the top of the current module.
require() calls load modules dynamically. They can be executed conditionally and their scope is defined by the expression in which they are used.
Clear Convert require() inside inner scopes with Fix all action option to prevent any changes in these complex cases when using Fix all action.
Suppress an inspection in the editor
Position the caret at the highlighted line and press Alt+Enter or click
.
Click the arrow next to the inspection you want to suppress and select the necessary suppress action. | https://www.jetbrains.com/help/phpstorm/javascript-and-typescript-require-is-used-instead-of-import.html | CC-MAIN-2021-31 | refinedweb | 179 | 59.84 |
CAM::SOAPApp - SOAP application framework
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
Do NOT subclass from this module to create your SOAP methods! That would make a big security hole. Instead, write your application like this example:
use CAM::SOAPApp; SOAP::Transport::HTTP::CGI -> dispatch_to('My::Class') -> handle; package My::Class; our @ISA = qw(SOAP::Server::Parameters); sub isLeapYear { my $pkg = shift; my $app = CAM::SOAPApp->new(soapdata => \@_); if (!$app) { CAM::SOAPApp->error('Internal', 'Failed to initialize the SOAP app'); } my %data = $app->getSOAPData(); if (!defined $data{year}) { $app->error('NoYear', 'No year specified in the query'); } if ($data{year} !~ /^\d+$/) { $app->error('BadYear', 'The year must be an integer'); } my $leapyear = ($data{year} % 4 == 0 && ($data{year} % 100 != 0 || $data{year} % 400 == 0)); return $app->response(leapyear => $leapyear ? 1 : 0); }
CAM::SOAPApp is a framework to assist SOAP applications. This package abstracts away a lot of the tedious interaction with SOAP and the application configuration state. CAM::SOAPApp is a subclass of CAM::App and therefore inherits all of its handy features.
When you create a class to hold your SOAP methods, that class should be a subclass of SOAP::Server::Parameters. It should NOT be a subclass of CAM::SOAPApp. If you were to do the latter, then all of the CAM::App and CAM::SOAPApp methods would be exposed as SOAP methods, which would be a big security hole, so don't make that mistake.
When loading this module, there are a few different options that can be selected. These can be mixed and matched as desired.
This initializes SOAPApp with all of the default SOAP::Lite options.
This tweaks some SOAP::Lite and environment variables to make the server work with SOAP-challenged clients. These tweaks specifically enable HTTP::CGI and HTTP::Daemon modes for client environments which don't offer full control over their HTTP channel (like Flash and Apple Sherlock 3).
Specifically, the tweaks include the following:
Sets Content-Type to
text/xml if it is not set or is set incorrectly.
Replaces missing SOAPAction header fields with ''.
Turns off charset output for the Content-Type (i.e. 'text/xml' instead of 'text/xml; charset=utf-8').
Outputs HTTP 200 instead of HTTP 500 for faults.
Adds a trailing '>' to the XML if one is missing. This is to correct a bug in the way Safari 1.0 posts XML from Flash.
(Experimental!) Kick off the SOAP handler automatically. This runs the following code immediately:
SOAP::Transport::HTTP::CGI -> dispatch_to(PACKAGE) -> handle;
Note that you must load PACKAGE before this statement.
Create a new application instance. The arguments passed to the SOAP method should all be passed verbatim to this method as a reference, less the package reference. This should be like the following:
sub myMethod { my $pkg = shift; my $app = CAM::SOAPApp->new(soapdata => \@_); ... }
Returns a hash of data passed to the application. This is a massaged version of the
soapdata array passed to new().
Prepare data to return from a SOAP method. For example:
sub myMethod { ... return $app->response(year => 2003, month => 3, date => 26); }
yields SOAP XML that looks like this (namespaces and data types omitted for brevity):
<Envelope> <Body> <myMethodResponse> <year>2003</year> <month>3</month> <date>26</date> </myMethodResponse> </Body> </Envelope>
Emit a SOAP fault indicating a failure. The
faultcode should be a short, computer-readable string (like "Error" or "Denied" or "BadID"). The
faultstring should be a human-readable string that explains the error. Additional values are encapsulated as
detail fields for optional context for the error. The result of this method will look like this (namespaces and data types omitted for brevity).
<Envelope> <Body> <Fault> <faultcode>$faultcode</faultcode> <faultstring>$faultstring</faultstring> <detail> <data> <$key1>$value1</$key1> <$key2>$value2</$key2> ... </data> <detail> </Fault> </Body> </Envelope>
This is a helper function used by response() to encode hash data into a SOAP-friendly array of key-value pairs that are easily transformed into XML tags by SOAP::Lite. You should generally use response() instead of this function unless you have a good reason.
Clotho Advanced Media Inc., cpan@clotho.com
Primary developer: Chris Dolan | http://search.cpan.org/~clotho/CAM-SOAPApp-1.06/lib/CAM/SOAPApp.pm | CC-MAIN-2014-10 | refinedweb | 699 | 57.87 |
11 July 2012 11:37 [Source: ICIS news]
SINGAPORE (ICIS)--Asia’s spot benzene prices extended their gains, adding $15-20/tonne (€12-16/tonne) at the close of trade on Wednesday, in line with spikes in the US benzene values overnight, market sources said.
Prices were assessed at $1,175-1,190/tonne FOB (free on board) ?xml:namespace>
On Tuesday night, US benzene prices firmed up to $5.00-5.25/gal or $1,495-1,570/tonne FOB Barges.
Offers for August loading in Asia were at $1,200/tonne FOB
A deal for August loading was heard at $1,175/tonne FOB
Market sentiment was also bolstered in late afternoon trade as European benzene prices opened higher on the back of tight supply for July, market sources said.
European benzene prices were at $1,320-1,350/tonne CIF (cost, insurance and freight) ARA (Amsterdam, Rotterdam, Antwerp), up by $40-55/tonne from Tuesday’s close.
($1 = €0.82) | http://www.icis.com/Articles/2012/07/11/9577112/asia-benzene-rises-15-20tonne-on-us-market-gains.html | CC-MAIN-2014-41 | refinedweb | 163 | 66.27 |
Red Hat Bugzilla – Bug 115349
mutex hang when using pthread_cond_broadcast() under high contention
Last modified: 2007-11-30 17:07:00 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6)
Gecko/20040113
Description of problem:
The attached test program hangs when run on a dual Xenon 2.4 GHZ box.
The main thread (and some of the worker threads) blocks in futex_wait,
waiting to acquire the mutex "mtx", which is unlocked. Attaching and
detaching a debugger causes the program to continue, as does sending
the process a STOP and CONT signal.
Version-Release number of selected component (if applicable):
glibc-2.3.2-95.6
How reproducible:
Always
Steps to Reproduce:
1. Compile the attached program with
cc -o cvtest cvtest.c -lpthread
2. In one window, run a server process
./cvtest -s
3. In the other window, run the test client
./cvtest -b
Actual Results: The test client will hang within minutes. Attach a
debugger and examine the main thread--it will be in the futex syscall,
inside __lll_mutex_lock_wait. The futex for the associated mutex will
have value 0.
Expected Results: The test client continues to print '.' characters.
If you examine the worker threads, you will find some also hanging in
the futex syscall for the same unlocked futex.
Additional info:
If you instead run cvtest with no arguments, causing it to never use
pthread_cond_broadcast(), it will not hang.
Created attachment 97574 [details]
Test program
Kernel is 2.4.21-9.ELsmp
Could you please try
These packages have temporarily disabled FUTEX_REQUEUE.
The bug does not reproduce with 2.3.2-95.10.
I've seen this same bug with the Boehm-Demers-Weiser conservative
garbage collector (aka libgc):
It was fixed by the updated glibc I got from.
Here is a simplified reproducer (hangs with -b with glibc which
doesn't have FUTEX_REQUEUE (or FUTEX_CMP_REQUEUE) commented out):
#define _XOPEN_SOURCE 500
#include <unistd.h>
#include <stdlib.h>
#include <pthread.h>
pthread_mutex_t mtx;
pthread_cond_t cv;
int broadcast;
int nn;
void *
tf (void *arg)
{
for (;;)
{
pthread_mutex_lock (&mtx);
while (!nn)
pthread_cond_wait (&cv, &mtx);
--nn;
pthread_mutex_unlock (&mtx);
}
}
int
main (int argc, char **argv)
{
int i, spins = 0;
pthread_mutexattr_t mtxa;
pthread_mutexattr_init (&mtxa);
pthread_mutexattr_settype (&mtxa, PTHREAD_MUTEX_ERRORCHECK_NP);
pthread_mutex_init (&mtx, &mtxa);
pthread_cond_init (&cv, NULL);
if (argc > 1)
{
if (!strcmp (argv[1], "-b"))
broadcast = 1;
else if (!strcmp (argv[1], "-B"))
broadcast = 2;
}
for (i = 0; i < 40; i++)
{
pthread_t th;
pthread_create (&th, NULL, tf, NULL);
}
pthread_mutex_lock (&mtx);
for (;;)
{
if ((spins++ % 1000) == 0)
write (1, ".", 1);
pthread_mutex_unlock (&mtx);
pthread_mutex_lock (&mtx);
int njobs = rand () % 41;
nn = njobs;
if (broadcast && (broadcast > 1 || (rand () % 30) == 0))
pthread_cond_broadcast (&cv);
else
while (njobs--)
pthread_cond_signal (&cv);
}
}
It happens even if cond->__data.__lock is held during the futex (FUTEX_REQUEUE)
syscall and only hangs with -b option, doesn't hang without any options
or with -B, so mixing pthread_cond_broadcast with pthread_cond_signal
syscalls is essential.
*** Bug 121283 has been marked as a duplicate of this bug. ***
Was this bug accidentally linked to the wrong errata?
I fail to see how an updated shadow-utils rpm resolves a problem with
glibc/pthreads...
No, the reference is correct. shadow-utils has to be updated in
addition to glibc. | https://bugzilla.redhat.com/show_bug.cgi?id=115349 | CC-MAIN-2018-30 | refinedweb | 531 | 58.48 |
Question:
This is probably obvious, so bear with me.
YES, I KNOW THAT java.io.File has no default constructor.
The problem is that When I try to extend java.io.File, it says "Cannot find constructor File() in java.io.File" even though I am overriding the default constructor in java.lang.Object.
Here is my code:
AbsRelFile.java
import java.io.File; public class AbsRelFile extends File { File f; private AbsRelFile(){ } }
This gives me an error, even though I am overriding the constructor.
NOTE: This class is not finished. Don't make a comment about why wouldn't I need this or a comment about how this class is useless. I just started writing it before I got this Error.
Solution:1
Because you didn't make an explicit call to
super(...) in your default constructor, it is implicitly attempting to call the default constructor for the super class, which, as you point out, doesn't exist in this case (the case of super being a
File). The solution to your problem is to make a call to the super constructor in your default
AbsRelFile() constructor. If you wan't to provide a default constructor for your class, you're going to need to call
super(...) with some default values.
Solution:2
When you define a constructor, Java inserts an implicit call to the super constructor as the very first line of the constructor. So your constructor is equivalent to:
private AbsRelFile(){ super(); }
Since there is no default constructor in the super class
File, it gives an error. To fix this, you need to place an explicit call to the super class constructor as the first line:
private AbsRelFile(){ super("fileName"); }
Most probably, you'll have to define some suitable parameters for
AbsRelFile constructor too which you can pass to super call.
On another note, constructors cannot be overridden. So it is wrong to say that you're overriding the
Object class constructor. You're simply defining a constrcutor for
AbsRelFile class.
Solution:3
Constructors, by default, call the default constructor of the super-class if you don't make a super constructor call yourself.
To avoid this, make a call to an actually-defined constructor of File.
Solution:4
Java automatically puts in a call to super() in your empty constructor, which is why you get the error.
Solution:5
The problem is that your
AbsRelFile constructor is trying to call the no-args constructor of
File. What you have written is equivalent to
private AbsRelFile() { super(); }
You need to make sure that you explicitly invoke one of the
File constructors that does exist. For example:
private AbsRelFile() { super("dummy"); }
Obviously, you will need to figure out a safe / harmless / appropriate superclass constructor and arguments to use for your particular use-case. (I haven't a clue what an
AbsRefFile is really supposed to be ... so I cannot advise on that.)
Aside - you don't "override" constructors. Constructors are never inherited in Java, so overriding simply doesn't apply here. Instead you declare constructors in the subclass, and have them chain to the appropriate constructors in the immediate superclass via an explicit
super(...) call ... or the implicit one the Java inserts by default.
Solution:6
First of all, I hope your field "File f" is not related to trying to access the superclass, but something to do with 'Rel' or 'Abs'..
The other posters have correctly pointed out that your implicit default constructor (AbsRelfile()) will attempt to call super() - which does not exist. So the only solution is to make a constructor that passes down some valid arguments.
If you are attempting to 'wrap' the whole java.util.File class (like when making your own Exception) you should probably provide a wrapper for each of the original constructors. Modern IDEs like Eclipse should have this a right-click away.
Note that File does not require that the given filename exists, in particular it does not exist when you want to do operations like file.mkdir().
If you need an actual, temporary file to work with, you could always do something like:
public class AbsRelFile() { public AbsRelFile() { super(File.createTempFile("AbsRelFile", "tmp").getAbsolutePath()); } }
.. but I'm puzzled as to why you want to subclass File in the first place.
Solution:7
To explain WHY in one line: When you define a constructor with a parameter (as in the File class), Java compiler WILL NOT generate the default constructor for you.
Note:If u also have question or solution just comment us below or mail us on toontricks1994@gmail.com
EmoticonEmoticon | http://www.toontricks.com/2019/06/tutorial-cannot-find-constructor-file.html | CC-MAIN-2019-43 | refinedweb | 762 | 64.2 |
Statements and Other Constructs
//,
and continue until the end of line.
A comment may appear anywhere in a Q# source file.
Documentation Comments
///,
are treated specially by the compiler when they appear immediately before
a namespace, operation, specialization, function, or type definition.
In that case, their contents are taken as documentation for the defined
callable or user-defined type, as for other .NET languages.
Within
/// comments, text to appear as a part of API documentation is
formatted as Markdown,
with different parts of the documentation being indicated by specially-named
headers.
As an extension to Markdown, cross references to operations, functions and
user-defined types in Q# can be included using the
@"<ref target>",
where
<ref target> is replaced by the fully qualified name of the
code object being referenced.
Optionally, a documentation engine may also support additional
Markdown extensions.
For example:
/// # Summary /// Given an operation and a target for that operation, /// applies the given operation twice. /// /// # Input /// ## op /// The operation to be applied. /// ## target /// The target to which the operation is to be applied. /// /// # Type Parameters /// ## 'T /// The type expected by the given operation as its input. /// /// # Example /// ```Q# /// // Should be equivalent to the identity. /// ApplyTwice(H, qubit); /// ``` /// /// # See Also /// - Microsoft.Quantum.Intrinsic.H operation ApplyTwice<'T>(op : ('T => Unit), target : 'T) : Unit { op(target); op(target); }
The following names are recognized as documentation comment headers.
- Summary: A short summary of the behavior of a function or operation, or of the purpose of a type. The first paragraph of the summary is used for hover information. It should be plain text.
- Description: A description of the behavior of a function or operation, or of the purpose of a type. The summary and description are concatenated to form the generated documentation file for the function, operation, or type. The description may contain in-line LaTeX-formatted symbols and equations.
- Input: A description of the input tuple for an operation or function. May contain additional Markdown subsections indicating each individual element of the input tuple.
- Output: A description of the tuple returned by an operation or function.
- Type Parameters: An empty section which contains one additional subsection for each generic type parameter.
- Example: A short example of the operation, function or type in use.
- Remarks: Miscellaneous prose describing some aspect of the operation, function, or type.
- See Also: A list of fully qualified names indicating related functions, operations, or user-defined types.
- References: A list of references and citations for the item being documented.
Namespaces
Every Q# operation, function, and user-defined type is
defined within a namespace.
Q# follows the same rules for naming as other .NET languages.
However, Q# does not support relative references to namespaces.
That is, if the namespace
a.b has been opened, a reference to an operation
names
c.d will not be resolved to an operation with full name
a.b.c.d.
Within a namespace block, the
open directive may be used to
import all types and callables declared in a certain namespace and refer to them by their unqualified name. Optionally, a short name for the opened namespace may be defined such that all elements from that namespace can (and need) to be qualified by the defined short name.
The
open directive applies to the entire namespace block within a file.
A type or callable defined in another namespace that is not
opened in the current namespace must be referenced by its fully-qualified name.
For example, an operation named
Op in the
X.Y namespace must be
referenced by its fully-qualified name
X.Y.Op, unless the
X.Y
namespace has been opened earlier in the current block.
As mentioned above, even if the
X namespace has been opened, it is not possible to reference the operation as
Y.Op.
If a short name
Z for
X.Y has been defined in that namespace and file, then
Op needs to be referred to as
Y.Op.
namespace NS { open Microsoft.Quantum.Intrinsic; // opens the namespace open Microsoft.Quantum.Math as Math; // defines a short name for the namespace }
It is usually better to include a namespace by using an
open directive.
Using a fully-qualified name is required if two namespaces define constructs
with the same name, and the current source uses constructs from both.
Formatting
Most Q# statements and directives end with a terminating semicolon,
;.
Statements and declarations such as
for and
operation that end with
a statement block do not require a terminating semicolon.
Each statement description notes whether the terminating semicolon
is required.
Statements, like expressions, declarations, and directives, may be broken out across multiple lines. Having multiple statements on a single line should be avoided.
Statement Blocks
Q# statements are grouped into statement blocks.
A statement block starts with an opening
{ and ends with a
closing
}.
A statement block that is lexically enclosed within another block is considered to be a sub-block of the containing block; containing and sub-blocks are also called outer and inner blocks.
Symbol Binding and Assignment
Q# distinguishes between mutable and immutable symbols. In general, the use of immutable symbols is encouraged because it allows the compiler to perform more optimizations.
The left-hand-side of the binding consists of a symbol tuple, and the right hand side of an expression.
Since all Q# types are value types - with the qubits taking a somewhat special role - formally a "copy" is created when a value is bound to a symbol, or when a symbol is rebound. That is to say, the behavior of Q# is the same as if a copy were created on assignment. This in particular also includes arrays. Of course in practice only the relevant pieces are actually recreated as needed.
Tuple Deconstruction
If the right-hand side of the binding is a tuple,
then that tuple may be deconstructed upon assignment.
Such deconstructions may involve nested tuples, and any full or partial deconstruction is valid as long as the shape of the tuple on the right hand side is compatible with the shape of the symbol tuple.
Tuple deconstruction can in particular also be used when the right-hand side of the
= is a tuple-valued expression.
let (i, f) = (5, 0.1); // i is bound to 5 and f to 0.1 mutable (a, (_, b)) = (1, (2, 3)); // a is bound to 1, b is bound to 3 mutable (x, y) = ((1, 2), [3, 4]); // x is bound to (1,2), y is bound to [3,4] set (x, _, y) = ((5, 6), 7, [8]); // x is rebound to (5,6), y is rebound to [8] let (r1, r2) = MeasureTwice(q1, PauliX, q2, PauliY);
Immutable Symbols
Immutable symbols are bound using the
let statement.
This is roughly equivalent to variable declaration and initialization in
languages such as C#, except that Q# symbols, once bound, may not be changed;
let bindings are immutable.
An immutable binding consists of the keyword
let, followed by
a symbol or symbol tuple, an equals sign
=, an expression to bind the symbol(s) to, and a terminating semicolon.
The type of the bound symbol(s) is inferred based on the expression on the right hand side.
Mutable Symbols
Mutable symbols are defined and initialized using the
mutable statement.
Symbols declared and bound as part of a
mutable statement may be rebound to a different value later in the code.
A mutable binding statement consists of the keyword
mutable, followed by
a symbol or symbol tuple, an equals sign
=, an expression to bind the symbol(s) to, and a terminating semicolon.
The type of the bound symbol(s) is inferred based on the expression on the right hand side. If a symbol is rebound later in the code, its type does not change, and the bound value needs to be compatible with that type.
Rebinding of Mutable Symbols
A mutable variable may be rebound using a
set statement.
Such a rebinding consists of the keyword
set, followed by
a symbol or symbol tuple, an equals sign
=, an expression to rebind the symbol(s) to, and a terminating semicolon.
The value must be compatible with the type(s) of the symbol(s) it is bound to.
A particular kind of set-statement we refer to as apply-and-reassign statement provides a convenient way of concatenation if the right hand side consists of the application of a binary operator an is to be rebound to the left argument to the operator. For example,
mutable counter = 0; for (i in 1 .. 2 .. 10) { set counter += 1; // ... }
increments the value of the counter
counter in each iteration of the
for loop. The code above is equivalent to
mutable counter = 0; for (i in 1 .. 2 .. 10) { set counter = counter + 1; // ... }
Similar statements are available for all binary operators in which the type of the left-hand-side matches the expression type. This provides for example a convenient way to accumulate values:
mutable results = new Result[0]; for (q in qubits) { set results += [M(q)]; // ... }
A similar concatenation exists for copy-and-update expressions on the right hand side. While our standard libraries contain the necessary tools for many common array initialization and manipulation needs, and thus help to avoid having update array items in the first place, such update-and-reassign statements provide an alternative if needed:
operation RandomInts(maxInt : Int, nrSamples : Int) : Int[] { mutable samples = new Double[0]; for (i in 1 .. nrSamples) { set samples += [RandomInt(maxInt)]; } return samples; } operation SampleUniformDistr(nrSamples : Int, prec : Int) : Double[] { let normalization = 1. / IntAsDouble(prec); mutable samples = RandomInts(prec, nrSamples); for (i in IndexRange(samples) { set samples w/= i <- normalization * IntAsDouble(samples[i]); } }
Tip
Avoid unnecessary use of update-and-reassign statements by leveraging the tools provided in <xref:microsoft.quantum.arrays>.
The function
function EmbedPauli (pauli : Pauli, location : Int, n : Int) : Pauli[] { mutable pauliArray = new Pauli[n]; for (index in 0 .. n - 1) { set pauliArray w/= index <- index == location ? pauli | PauliI; } return pauliArray; }
for example can simply be expressed using the function
ConstantArray in
Microsoft.Quantum.Arrays:
function EmbedPauli (pauli : Pauli, i : Int, n : Int) : Pauli[] { return ConstantArray(n, PauliI) w/ i <- pauli; }
Binding Scopes
In general, symbol bindings go out of scope and become inoperative at the end of the statement block they occur in. There are two exceptions to this rule:
- The binding of the loop variable of a
forloop is in scope for the body of the for loop, but not after the end of the loop.
- All three portions of a
repeat/
untilloop (the body, the test, and the fixup) are treated as a single scope, so symbols that are bound in the body are available in the test and in the fixup.
For both types of loops, each pass through the loop executes in its own scope, so bindings from an earlier pass are not available in a later pass.
Symbol bindings from outer blocks are inherited by inner blocks. A symbol may only be bound once per block; it is illegal to define a symbol with the same name as another symbol that is within scope (no "shadowing"). The following sequences would be legal:
if (a == b) { ... let n = 5; ... // n is 5 } let n = 8; ... // n is 8
and
if (a == b) { ... let n = 5; ... // n is 5 } else { ... let n = 8; ... // n is 8 }
But this would be illegal:
let n = 5; ... // n is 5 let n = 8; // Error!! ...
as would:
let n = 8; if (a == b) { ... // n is 8 let n = 5; // Error! ... } ...
Control Flow
For-Loop
The
for statement supports iteration over an integer range or over an array.
The statement consists of the keyword
for, an open parenthesis
(,
followed by a symbol or symbol tuple, the keyword
in, an expression of type
Range or array, a close parenthesis
), and a statement block.
The statement block (the body of the loop) is executed repeatedly, with the defined symbol(s) (the loop variable(s)) bound to each value in the range or array. Note that if the range expression evaluates to an empty range or array, the body will not be executed at all. The expression is fully evaluated before entering the loop, and will not change while the loop is executing.
The binding of the declared symbol(s) is immutable and follows the same rules as other variable bindings. In particular, it is possible to destruct e.g. array items for an iteration over an array upon assignment to the loop variable(s).
For example,
// ... for (qb in qubits) { // qubits contains a Qubit[] H(qb); } mutable results = new (Int, Results)[Length(qubits)]; for (index in 0 .. Length(qubits) - 1) { let measured = set results w/= index <- (index, M(qubits[i])); } mutable accumulated = 0; for ((index, measured) in results) { if (measured == One) { set accumulated += 1 <<< index; } }
The loop variable is bound at each entrance to the loop body, and unbound at the end of the body. In particular, the loop variable is not bound after the for loop is completed.
Repeat-Until-Success Loop
The
repeat statement supports the quantum “repeat until success” pattern.
It consists of the keyword
repeat, followed by a statement block
(the loop body), the keyword
until, a Boolean expression,
the keyword
fixup, and another statement block (the fixup).
The loop body, condition, and fixup are all considered to be a single scope,
so symbols bound in the body are available in the condition and fixup.
Note that the fixup block is required, even if there is no fixup to be done.
The loop body is executed, and then the condition is evaluated. If the condition is true, then the statement is completed; otherwise, the fixup is executed, and the statement is re-executed starting with the loop body. Note that completing the execution of the fixup ends the scope for the statement, so that symbol bindings made during the body or fixup are not available in subsequent repetitions.
For example, the following code is a probabilistic circuit that implements an important rotation gate $V_3 = (\boldone + 2 i Z) / \sqrt{5}$ using the Hadamard and T gates. The loop terminates in 8/5 repetitions on average. See Repeat-Until-Success: Non-deterministic decomposition of single-qubit unitaries (Paetznick and Svore, 2014) for details.) fixup { } }
Conditional Statement
The
if statement supports conditional execution.
It consists of the keyword
if, an open parenthesis
(, a Boolean
expression, a close parenthesis
), and a statement block (the then block).
This may be followed by any number of else-if clauses, each of which consists
of the keyword
elif, an open parenthesis
(, a Boolean expression,
a close parenthesis
), and a statement block (the else-if block).
Finally, the statement may optionally finish with an else clause, which
consists of the keyword
else followed by another statement block
(the else block).
The condition is evaluated, and if it is true, the then block is executed.
If the condition is false, then the first else-if condition is evaluated;
if it is true, that else-if block is executed.
Otherwise, the second else-if block is tested, and then the third, and so on
until either a clause with a true condition is encountered or there are no
more else-if clauses.
If the original if condition and all else-if clauses evaluate to false,
the else block is executed if one was provided.
Note that whichever block is executed is executed in its own scope. Bindings made inside of a then, else-if, or else block are not visible after the end of the if statement.
For example,
if (result == One) { X(target); }
or
if (i == 1) { X(target); } elif (i == 2) { Y(target); } else { Z(target); }
Return
The return statement ends execution of an operation or function
and returns a value to the caller.
It consists of the keyword
return, followed by an expression of the
appropriate type, and a terminating semicolon.
A callable that returns an empty tuple,
(), does not require a
return statement.
If an early exit is desired,
return () may be used in this case.
Callables that return any other type require a final return statement.
There is no maximum number of return statements within an operation. The compiler may emit a warning if statements follow a return statement within a block.
For example,
return 1;
or
return ();
or
return (results, qubits);
Fail
The fail statement ends execution of an operation and returns an error value
to the caller.
It consists of the keyword
fail, followed by a string
and a terminating semicolon.
The string is returned to the classical driver as the error message.
There is no restriction on the number of fail statements within an operation. The compiler may emit a warning if statements follow a fail statement within a block.
For example,
fail $"Impossible state reached";
or
fail $"Syndrome {syn} is incorrect";
Qubit Management
Note that none of these statements are allowed within the body of a function. They are only valid within operations.
Clean Qubits
The
using statement is used to acquire new qubits for use during a statement block.
The qubits are guaranteed to be initialized to the computational
Zero state.
The qubits should be in the computational
Zero state at the
end of the statement block; simulators are encouraged to enforce this.
The statement consists of the keyword
using, followed by an open
parenthesis
(, a binding, a close parenthesis
), and
the statement block within which the qubits will be available.
The binding follows the same pattern as
let statements: either a single
symbol or a tuple of symbols, followed by an equals sign
=, and either a
single value or a matching tuple of initializers.
Initializers are available either for a single qubit, indicated as
Qubit(), or
an array of qubits, indicated by
Qubit[, an
Int expression, and
].
For example,
using (q = Qubit()) { // ... } using ((ancilla, qubits) = (Qubit(), Qubit[bits * 2 + 3])) { // ... }
Dirty Qubits
The
borrowing statement is used to obtain qubits for temporary use.
The statement consists of the keyword
borrowing, followed by an open
parenthesis
(, a binding, a close parenthesis
), and
the statement block within which the qubits will be available.
The binding follows the same pattern and rules as the one in a
using statement.
For example,
borrowing (q = Qubit()) { // ... } borrowing ((ancilla, qubits) = (Qubit(), Qubit[bits * 2 + 3])) { // ... }
The borrowed qubits are in an unknown state and go out of scope at the end of the statement block. The borrower commits to leaving the qubits in the same state they were in when they were borrowed, i.e. their state at the beginning and at the end of the statement block is expected to be the same. This state in particular is not necessarily a classical state, such that in most cases, borrowing scopes should not contain measurements.
Such qubits are often known as “dirty ancilla”. See Factoring using 2n+2 qubits with Toffoli based modular multiplication (Haner, Roetteler, and Svore 2017) for an example of dirty ancilla use.
When borrowing qubits, the system will first try to fill the request
from qubits that are in use but that are not accessed during the body of the
borrowing statement.
If there aren't enough such qubits, then it will allocate new qubits
to complete the request.
Expression Evaluation Statements
Any call expression of type
Unit may be used as a statement.
This is primarily of use when calling operations on qubits that return
Unit
because the purpose of the statement is to modify the implicit quantum state.
Expression evaluation statements require a terminating semicolon.
For example,
X(q); CNOT(control, target); Adjoint T(q); | https://docs.microsoft.com/en-us/quantum/language/statements?view=qsharp-preview | CC-MAIN-2019-22 | refinedweb | 3,253 | 54.12 |
24 March 2005 15:41 [Source: ICIS news]
LONDON (CNI)--US coatings and chemicals group PPG faces a bill of about $150m (around Euro113m) after losing a legal battle with a customer which had claimed one of its wood preservatives was defective.
PPG lost its appeal against a decision by the District Court in Minnesota in a breach-of-warranty case brought by Marvin Windows and Doors,
The total judgement, including interest, could cost PPG between $145m and $150m (about Euro109m-113m), PPG has estimated. PPG is evaluating its options for further appeals.
Marvin said that problems started with the durability of its windows and doors after it changed, around 1985, from preservatives containing pentachlorophenol (Penta) to PPG’s Preservative In Line Treatment (PILT). The case was originally heard before the District Court in ?xml:namespace>
The United States Court of Appeals for the Eighth Circuit in its judgement said that the time lag between changing preservatives and bringing the action was reasonable partially because PPG told Marvin’s employees “wood products treated with PILT would last as long or longer than Penta-treated products”.
The district court initially instructed the jury to find against PPG.
PPG appealed against this ruling on a number of grounds. Most of which were dismissed by the Appeals court which said in its latest judgement: “Frankly, we think the District Court on balance did a fine job.”
However, the three judge appeals panel was not completely in agreement. There were two dissenting verdicts on the level of compensation that PPG should have to pay.
To see the full verdict: | http://www.icis.com/Articles/2005/03/24/663313/ppg-faces-150m-bill-after-losing-wood-preservative.html | CC-MAIN-2014-35 | refinedweb | 265 | 58.32 |
useabil-
ity
PL
an operator name (opname)
Operator names are typically small lowercase words like enter-
loop, leaveloop, last, next, redo etc. Sometimes they are
rather cryptic like gv2cv, i_ncmp and ftsvtx.
an operator tag name (optag)
Operator tags can be used to refer to groups (or sets) of oper-.
opcodes).)
opset (OP, ...)
Returns an opset containing the listed operators.
opset_to_ops (OPSET)
Returns a list of operator names corresponding to those opera-
tors in the set.
opset_to_hex (OPSET)
Returns a string representation of an opset. Can be handy for
debugging.
full_opset
Returns an opset which includes all operators.
empty_opset
Returns an opset which contains no operators.
invert_opset (OPSET)
Returns an opset which is the inverse set of the one supplied.
verify_opset (OPSET, ...) automati-
cally and will croak if given an invalid opset.
define_optag (OPTAG,
opmask Returns an opset corresponding to the current opmask.
opdesc (OP, ...)
This takes a list of operator names and returns the correspond-
ing list of operator descriptions.
opdump (PAT)
Dumps to STDOUT a two column list of op names and op descrip-
tions.', ...)
:base_core
:base_mem
These memory related ops are not included in :base_core because
they can easily be used to implement a resource attack (e.g., con-
sume all available memory).
concat repeat join range
anonlist anonhash
Note that despite the existence of this optag a memory resource
attack may still be possible using only :base_core ops.
Disabling these ops is a very heavy handed way to attempt to pre-
vent a memory resource attack. It's probable that a specific mem-
ory limit mechanism will be added to perl in the near future.
:base_loop
:base_io
custom -- where should this go
:base_math
:base_thread
These ops are related to multi-threading.
lock threadsv
:default
A handy tag name for a reasonable default set of ops. (The cur-
rent ops allowed are unstable while development continues. It will
change.)
:base_core :base_mem :base_loop :base_io :base_orig :base_thread
If safety matters to you (and why else would you be using the
Opcode module?) then you should not rely on the definition of
this, or indeed any other, optag!
:filesys_read
:filesys_open
sysopen open close
umask binmode
open_dir closedir -- other dir ops are in :base_io
:filesys_write
link unlink rename symlink truncate
mkdir rmdir
utime chmod chown
fcntl -- not strictly filesys related, but possibly as dangerous?
:subprocess
backtick system
fork
wait waitpid
glob -- access to Cshell via <`rm *`>
:ownprocess
exec exit kill
time tms -- could be used for timing attacks (paranoid?)
:others
require dofile
caller -- get info about calling environment and args
reset
dbstate -- perl -d version of nextstate(ment) opcode
:dangerous
This tag is simply a bucket for opcodes that are unlikely to be
used via a tag name but need to be tagged for completeness and
documentation.
syscall dump chroot
ops(3) -- perl pragma interface to Opcode module.
Safe(3) -- Opcode and namespace limited execution compartments
Originally designed and implemented by Malcolm Beattie, mbeat-
tie@sable.ox.ac.uk as part of Safe version 1.
Split out from Safe module version 1, named opcode tags and other
changes added by Tim Bunce.
perl v5.8.8 2001-09-21 Opcode(3) | http://www.syzdek.net/~syzdek/docs/man/.shtml/man3/Opcode.3.html | crawl-003 | refinedweb | 520 | 64.41 |
Plan a Scalable Architecture Through Fault Injection
Plan a Scalable Architecture Through Fault Injection
Join the DZone community and get the full member experience.Join For Free
Get the Edge with a Professional Java IDE. 30-day free trial.
I have been doing a lot of testing on scalable architectures since last year and at some point, Steve Harris pointed out the principles of fault injection in a discussion (Thanks Steve for making me think about this).
When I started to look at it more closely, I realized a couple of things:
- Unit testing is good, there are some great principles, but sometimes they are not explained clearly enough
- Testing an architecture for scaling is a different piece of work. You need way more than unit testing.
I didn’t really understand unit testing
You surely have heard that when writing unit tests, you should think of testing failing cases, not working ones.
Then you started writing tests and didn’t really think of a good failing case so you wrote a test for a case that works.
Then after some thinking you found a failing case (usually putting a NPE somewhere).
Then you got used to that and when writing unit tests, you cover both working and failing cases.
I used to do that, and I changed and now I see I write better applications.
Let me explain.
It all started when I wanted to add fault injections in my tests.
What is fault injection?
From Wikipedia:
“In software testing, fault injection is a technique for improving the coverage of a test by introducing faults to test code paths”
Put it more clearly, let’s take an example:
public class TextApp { public static void main(String[] arg) { try { String filename = "someFile.txt"; String content = "Let's write this in the file"; WriterService writerService = new WriterServiceImpl(); writerService.writeOnDisk(filename, content); } catch (Exception e) { System.out.println("We couldn't write on the disk. " + e.getMessage()); } } }
import java.io.IOException; public interface WriterService { void writeOnDisk(String filename, String content) throws Exception; }
import java.io.BufferedWriter; import java.io.FileWriter; import java.io.IOException; public class WriterServiceImpl implements WriterService { public void writeOnDisk(final String filename, final String content) throws Exception { try { FileWriter file = new FileWriter(filename); BufferedWriter out = new BufferedWriter(file); out.write(content); out.close(); } catch (IOException e) { throw new Exception("file writing error", e); } } }
And I want to write a unit test for the WriterService:
import org.junit.Assert; import org.junit.Test; public class WriterServiceTest { @Test public void testSuccessfulWriting() { WriterService writerService = new WriterServiceImpl(); try { writerService.writeOnDisk("test.txt", "testContent"); } catch (Exception e) { Assert.fail("We should be able to write to the disk"); } } @Test public void testFailingWriting() { WriterService writerService = new WriterServiceImpl(); try { writerService.writeOnDisk("impossiblefilename^&:/.txt", null); Assert.fail("We should not be able to write to the disk"); } catch (Exception e) { Assert.assertEquals("file writing error", e.getMessage()); } } }
And I’ve done exactly what I described : Write a unit test with a working case and a failing case.
Let’s say I’m satisfied enough (obviously the code here is minimal and should be improved but for the demonstration, we will say that the code in the WriterService is satisfying).
Now I could improve the coverage of my test through fault-injection.
So I decide to inject errors at runtime in the service while executing my unit test.
For instance, I can think of the case where the disk is full, I want to test that.
But how can I write a test that will execute when the disk is full? Well, I am going to simulate(inject) that fault.
There is a pretty good API for that called Byteman, which is a bytecode manipulation api, with a JUnit/TestNG runner, so basically, when using this library you could then write a test like this:
... @BMRule(name="throw sync failed exception", targetClass="FileInputStream", targetMethod="(File)", condition="$1.getName().equals\"diskfullname.txt\")" action="throw new SyncFailedException(\"can't sync to disk!\")") @Test public void testDiskfull() { WriterService writerService = new WriterServiceImpl(); try { writerService.writeOnDisk("diskfullname.txt", null); Assert.fail("We should not be able to write to the disk"); } catch (Exception e) { Assert.assertEquals("file writing error", e.getMessage()); } } ...
And that’s very nice, I can write more tests, have more robust code.
But I can do better.
Let’s forget about Byteman for a second, how would you do if you have to do it by yourself?
I have an idea.
I am going to break the WriterService into two classes : WriterService and DiskWriterService
import java.io.BufferedWriter; import java.io.FileWriter; import java.io.IOException; public class WriterServiceImpl implements WriterService { DiskWriterService diskWriterService = new DiskWriterServiceimpl(); public void writeOnDisk(final String filename, final String content) throws Exception { try { diskWriterService.write(filename, content); } catch (IOException e) { throw new Exception("file writing error", e); } } public void setDiskWriterService(final DiskWriterService diskWriterService) { this.diskWriterService = diskWriterService; } }
import java.io.IOException; public interface DiskWriterService { void write(String filename, String content) throws IOException; }
import java.io.BufferedWriter; import java.io.FileWriter; import java.io.IOException; public class DiskWriterServiceimpl implements DiskWriterService { public void write(final String filename, final String content) throws IOException { FileWriter file = new FileWriter(filename); BufferedWriter out = new BufferedWriter(file); out.write(content); out.close(); } }
Breaking the service in two adds an essential advantage to my application: It makes it more loosely coupled. I probably don’t need to explain in details what is loose coupling, basically it’s the idea of breaking an app into components that do not need to know the state of other components to work.
With this I can write the following class:
import java.io.IOException; public class FaultyDiskWriterServiceImpl implements DiskWriterService { private IOException exception; public void write(final String filename, final String content) throws IOException { throw exception; } public void setException(final IOException exception) { this.exception = exception; } }
and then I can replace my test case by this:
@Test public void testDiskfull() { WriterServiceImpl writerService = new WriterServiceImpl(); FaultyDiskWriterServiceImpl faultyDiskWriterService = new FaultyDiskWriterServiceImpl(); faultyDiskWriterService.setException(new SyncFailedException("disk full")); writerService.setDiskWriterService(faultyDiskWriterService); try { writerService.writeOnDisk("diskfullname.txt", null); Assert.fail("We should not be able to write to the disk"); } catch (Exception e) { Assert.assertEquals("file writing error", e.getMessage()); } }
So let’s think for a minute, what did I do?
– I wrote a failing test case (disk full)
– It forced me to break my WriterService in two, which is good because it makes my application more loosely coupled. If I have another service that needs to write on disk, I could reuse that DiskWriterService
- I can now unit test my DiskService and make sure that writing on the disk is robust.
It separates the testing of writing on the disk and the testing of my WriterService (now it is just a simple case where it calls the DiskWriterService but it could do more things like logging, writing to multiple places…).
– With this new DiskWriterService, it implicitly makes my architecture more robust for scalability, if someday I need to write to more than one disk or different kind of storage, (CloudDiskWriter, QueueDiskWriter…), I could write other DiskWriterService implementation without touching the rest of the code.
So now I see more clearly the point of testing failing cases. In case you have no idea when doing such thing, my advice would be “Think of the errors you could inject in your code when writing unit tests”
That’s it for point 1, next time for point 2 – Testing for scaling!
Get the Java IDE that understands code & makes developing enjoyable. Level up your code with IntelliJ IDEA. Download the free trial. }} | https://dzone.com/articles/plan-scalable-architecture | CC-MAIN-2019-04 | refinedweb | 1,254 | 56.96 |
This is a discussion on how is it possible to write/read in language other than english in a console app? within the C++ Programming forums, part of the General Programming Boards category; Originally Posted by tabstop This is a library issue, not a gcc issue. wcout doesn't exist on my MinGW port, ...
I think jwenting's coment refers to the fact that most european languages can be used with A-Za-z and a few others that are available in the ANSI character set (same as ISO-8859-1, I think?). However, I doubt that Arabic, Hebrew, Hindu, Urdu, Thai, Chinese, Japanese, Greek, and several other languages do not even use A-Z as representations of their native language, and in this case you NEED a wide character set. As to what you do to achieve that, my guess would be that if you want to make your life easy, you skip console applications, as support for multilingual console apps is very limited.
--
Mats
Compilers can produce warnings - make the compiler programmers happy: Use them!
Please don't PM me for help - and no, I don't do help over instant messengers.
but the catch here is , i chose C++ because its chalenging! , and i chose console since it is again more chalenging!and it wants you to understand and make what you want in a hard wayconsidered to other programmming langs(full of experience and excitment for newbies like me).
i want the job done! and im not gonna leave it alone before i test the current sulotions!and see the outcome.
to tell you the truth, im working on the GUI version too, using wxwidgets, but first ive got to accomplish this task and then go for the other one.
again many tanx .
It seems like the MS libraries do have a wcout, however, gcc-mingw uses a fairly old version of the MS C runtime, so it doesn't have that.
Not sure what options you have here - using MS Visual Studio 2008 may be one choice.
--
Mats
Compilers can produce warnings - make the compiler programmers happy: Use them!
Please don't PM me for help - and no, I don't do help over instant messengers.
there is an issue here! so to achive this, using a windows API in a expected cross platform application; i need a function which makes it clear that the code is being compiled under windows and thus would run in windows Os , then after that ,it will execute bunch of codes that does the job here! (manipulating unicode strings )and
if not the function will skip that part and accroding to the platform( here usullay *nix platforms)in its compilition time,it would normaly act. and compiles!
as far as i know ,unicode is farily implemented such platforms (rather than windows).and i will not face any problems in linux or Mac Os .
how to write that function to determing whether is windows or not! in complition time!
using size of() will do the trick?(ive heard wchar_t is 2 byte in windows and 4 bytes on *nix platforms! is it true ?)
tanx
wcout is not a windows API. My point about using the Visual studio is that it HAS wcout, which is a standard function.
As to OTHER poratibility problems, yes, they may well exist now and in the future.
--
Mats
Compilers can produce warnings - make the compiler programmers happy: Use them!
Please don't PM me for help - and no, I don't do help over instant messengers.
so first ive gotta give it a try with VS 2008 compiler ( i think!?) and then continue on! solving the *nix specific problems !
tanx .
I use STLport with MinGW - which does have wide iostreams.
It's been explained already that wide streams do not support Unicode. Trying to do so in a cross-platform manner is even further complicated in that wchar_t and literal wide-string encodings are implementation defined.
Having said all that - you can prevent the implicit wchar_t to char conversion in wide streams by using your own codecvt facet. This works ok for wide fstreams under most implementations, but still doesn't handle screen output for any windows compiler I have access to. However, starting with the VC++ 2008 CRT, you can use a new MS extension to enable Unicode (UTF16LE) output with wcout. Here's some code to play with:
I would interested in knowing if this works for anyone running *nix and a Unicode-enabled terminal.I would interested in knowing if this works for anyone running *nix and a Unicode-enabled terminal.Code:#include <iostream> #include <iomanip> #include <fstream> #include <locale> #include <string> //------------------------------------------------------------------------------ // null_wcodecvt is a codecvt facet that prevents the implicit conversion of // whcar_t to char in wide streams typedef std::codecvt<wchar_t , char , mbstate_t> null_wcodecvt_base; class null_wcodecvt : public null_wcodecvt_base { public: explicit null_wcodecvt(size_t refs = 0) : null_wcodecvt_base(refs) {} protected: virtual result do_out(mbstate_t&, const wchar_t* from, const wchar_t* from_end, const wchar_t*& from_next, char* to, char* to_end, char*& to_next) const { size_t len = (from_end - from) * sizeof(wchar_t); memcpy(to, from, len); from_next = from_end; to_next = to + len; return ok; }//do_out virtual result do_in(mbstate_t&, const char* from, const char* from_end, const char*& from_next, wchar_t* to, wchar_t* to_end, wchar_t*& to_next) const { size_t len = (from_end - from); memcpy(to, from, len); from_next = from_end; to_next = to + (len / sizeof(wchar_t)); return ok; }//do_in virtual result do_unshift(mbstate_t&, char* to, char*, char*& to_next) const { to_next = to; return noconv; }//do_unshift virtual int do_length(mbstate_t&, const char* from, const char* end, size_t max) const { return (int)((max < (size_t)(end - from)) ? max : (end - from)); }//do_length virtual bool do_always_noconv() const throw() { return true; }//do_always_noconv virtual int do_encoding() const throw() { return sizeof(wchar_t); }//do_encoding virtual int do_max_length() const throw() { return sizeof(wchar_t); }//do_max_length };//null_wcodecvt //------------------------------------------------------------------------------ std::wostream& wendl(std::wostream& out) { // this is needed for files opened in binary mode under Windows in order // to retain Windows-style newline #if defined(_WIN32) out.put(L'\r'); #endif out.put(L'\n'); out.flush(); return out; }//wendl //------------------------------------------------------------------------------ const wchar_t UTF_BOM = 0xfeff; const wchar_t CHECK_SYM = L'\u221a'; //------------------------------------------------------------------------------ void file_test() { // create UTF16LE text file for windows, UTF32[BE|LE] for *nix std::wfstream file; null_wcodecvt wcodec(1); std::locale wloc(std::locale::classic(), &wcodec); file.imbue(wloc); file.open("data.txt", std::ios::out | std::ios::binary); if (!file) { std::cerr << "Failed to open data.txt for writting" << std::endl; return; }//if file << UTF_BOM << L"data = " << 42 << CHECK_SYM << wendl; file.close(); }//file_test //------------------------------------------------------------------------------ #if defined(_MSC_VER) && (_MSC_VER >= 1500) # define HAVE_O_U16TEXT # include <fcntl.h> # include <io.h> #endif void wcout_test() { #if defined(HAVE_O_U16TEXT) // newest MS CRT supports UTF16 output using special mode int mode = _setmode(_fileno(stdout), _O_U16TEXT); #else // for everything else, use our codecvt facet - I haven't seen this work // for any windows compiler...may work for *nix but untested null_wcodecvt wcodec(1); std::locale wloc(std::locale::classic(), &wcodec); std::wcout.imbue(wloc); #endif std::wcout << L"U+221a - [" << wchar_t(0x221a) << L"]" << std::endl; #if defined(HAVE_O_U16TEXT) // revert to original mode _setmode(_fileno(stdout), mode); #endif std::cout << "Testing 1, 2, 3" << std::endl; }//wcout_test //------------------------------------------------------------------------------ int main() { file_test(); wcout_test(); return 0; }//main
There's still the underlying problem that the wchar_t encoding is implementation defined - which means that using wchar_t as a cross-platform Unicode character is useless.
"Forcing" wchar_t to do something useful with Unicode is really a platform and implementation dependent exercise.
gg
many many tanx dear Codeplug, i will test it and i will tell what happens
again tanx alot .
i really appreciate your helps , all of you .
tanx
At the risk of being assaulted by some of the others here, why not just use C standard functions instead of iostreams?
Because they're really no better in this department.
All the buzzt!
CornedBeeCornedBee
"There is not now, nor has there ever been, nor will there ever be, any programming language in which it is the least bit difficult to write bad code."
- Flon's Law
I have some insite that is only going to apply to OS specific situations. I googled and found a few libraries floating around that are more apt for doing localization. | http://cboard.cprogramming.com/cplusplus-programming/107908-how-possible-write-read-language-other-than-english-console-app-2.html | CC-MAIN-2015-35 | refinedweb | 1,350 | 61.26 |
Jeff Turner wrote:
>.
Honestly, I don't know either, and both work for me.
> I just didn't want to potentially exclude
> one possibility.
ACK
> ...
>>?
Probably yes. But this AFAIS does not preclude what I propose, it would
simply be a further evolutionary step.
>>.
All xml should have namespaces and schema definitions if we want to out
arbitrary content. With other sources, we would stick with one
processing mode.
>>Jeff, as I have already proposed, and now I ask you, please revert my
>>commits as I feel they were not appropriate.
>
> Thanks for doing so.
It was simply the right thing to do.
>>I would do so myself but cannot ATM. And again sorry if you felt
>>pushed, but I really am in good faith.
>
> You take rants far too personally ;)
Probably ;-)
But in fact it's not your rant (actually, did you rant at all?) but me
being a bit worried of having done another communication mess.
Oh well...
--
Nicola Ken Barozzi nicolaken@apache.org
- verba volant, scripta manent -
(discussions get forgotten, just code remains)
--------------------------------------------------------------------- | http://mail-archives.apache.org/mod_mbox/forrest-dev/200309.mbox/%3Cbjn31l$nov$1@sea.gmane.org%3E | CC-MAIN-2017-04 | refinedweb | 176 | 68.36 |
Table of Contents
How to import EPrints with VuFind
Tips and configurations courtesy of Ranju Upadhyay, National University of Ireland Maynooth.
As our repo is EPrints, I have needed to do the configuration in VuFind for data harvest from EPrints. It is not whole lot different than DSpace.
On the EPrints side, the data is already exposed for harvesting in oai_dc, so there was no configuration needed there. On VuFind, I did the following:
1. Set up OAI Harvester
I created a section on oai.ini which looks something like this:
[NUIMEprints] url = metadataPrefix = oai_dc idSearch[] = "/^oai:generic.eprints.org:/" idReplace[] = "nuimeprn-" ;idSearch[] = "/\//" ;idReplace[] = "-" injectId = "identifier" ;injectDate = "datestamp"
The big thing that I discovered in this stage was that VuFind does not like full colon i.e. “:” so I had to use “-” as the delimeter between my namespace and eprints uuid i.e. nuimeprn- and not nuimeprn:
2. Set up import properties
I then copied the dspace.properities file and renamed it to eprints.properties.In this file I made the following changes:
institution = "National University of Ireland Maynooth " collection = "ePrints"
3. Set up import XSLT
I then copied dspace.xsl and renamed it eprints.xsl. I made three changes to this file:
First, set the record type to eprints:
<!-- RECORDTYPE --> <field name="recordtype">eprints</field>
Next, add support for URLs from the dc:relation field:
<xsl:for-each <field name="url"> <xsl:value-of </field> </xsl:for-each>
Finally, remove the check for hdl.handle.net in the dc:identifier URL processing:
<xsl:for-each <xsl:if <field name="url"> <xsl:value-of </field> </xsl:if> </xsl:for-each>
4. Set Up Change Tracking (optional)
If you need to track record change dates (see Tracking Record Changes for details), you need to do a couple of extra things:
- Uncomment the injectDate line in the oai.ini file section above.
- Add these lines to eprints.properties:
track_changes = 1 solr_core = "biblio"
- Add these lines to eprints.xsl:
First, after the other parameter declarations:
<xsl:param1</xsl:param> <xsl:parambiblio</xsl:param>
Further down, among the other field population code:
<xsl:if <field name="first_indexed"> <xsl:value-of </field> <field name="last_indexed"> <xsl:value-of </field> </xsl:if> | https://vufind.org/wiki/indexing:eprints | CC-MAIN-2022-05 | refinedweb | 366 | 66.03 |
I am recently new to C++ and I wondered if someone could show me how to modify my program below. I want the program to test if there is more than 1 day that has the highest rainfall and susequently to output those corresponding days. Than you :)
#include <iostream> #include <conio.h> #include <iomanip> using namespace std; double rain[] = {18.0, 18.0, 3.2, 12.1, 14.6, 17.8, 3.2}; int main () { double rainiestValue (0.0); int rainyDayNo (-1); for (int dayNo ( 0); dayNo < 7; ++dayNo) if (rainiestValue < rain [dayNo]) { rainiestValue = rain [dayNo]; rainyDayNo = dayNo; } cout << fixed << setprecision (2); cout << "\nThe highest rainfall is: " << rainiestValue << "mm" << " on day " << rainyDayNo; while (!_kbhit()); return 0; } | https://www.daniweb.com/programming/software-development/threads/273287/c-problem-please-help | CC-MAIN-2017-17 | refinedweb | 116 | 72.87 |
Agenda
See also: IRC log
-> WG home page
Guus: in "W3C speak", the term "member" means an organization and the term "participant" means individuals like you and me
Guus: Bernard has already expressed
difficulties with this teleconference time
... intention is that UTC time is not fixed; we will shift by an hour when the US changes its clock
... this year both US and Europe change on 29 October
... at present we have no Asian participants
... W3C tradition is that scribe duties rotate around the WG
... we scribe in irc
... if you don't have an irc client, you can use a Web browser with the W3C cgi-irc interface; I've been using it successfully for several years
<TomB> I use ChatZilla - as a plug-in for Mozilla
Guus: this WG operates in the public view
... so Member Access is not necessary for any WG materials, only for other protected things such as Coordination Group records
Tom: I live in Berlin
Ralph: I live an hour outside of Boston
Sean: I live in Manchester, UK
... participated in WebOnt
... interested in delivery of SKOS
... have a connection to KnowledgWeb
Fabien: I live in Cannes
... am a member of INRIA
... team has background in knowledge representation and knowledge management
... also participate in the GRDDL WG
Bernard: work for Sun Microsystems Labs and
reside in the UK though I report to Sun in Burlington MA
... working on COHSE project with Sean in Manchester
<Bern> COHSE project --
Daniel: I work at Stanford Univ as part of
Center for Biomedical Ontology
... also participate in HCLSIG
Diego: I work in the Semantic Web project at CTIC in Spain
Antoine: work at Vrije Universitat in Amsterdam and also on a national archives project
-> Antoine's introduction
Guus: I also work at Vrije Universitat
Guus: was co-chair of WebOnt and SemWeb Best Practices WG
Guus: hope that there can be sharing between KnowledgeWeb and this WG
Guus: We are a Recommendation-Track Working
Group
... we have a number of deliverables, one of which [SKOS] is already identified as a Recommendation
... the Recommendation process requires the documents to move along a process from Working Draft to Last Call to testing via Candidate Recommendation
<Antoine> isn't SKOS a working draft?
Guus: see Art of Consensus
... SKOS is already a Working Draft, yes
... to start a Recommendation track we like to have a document already available
Tom: Alistair is traveling today after Dublin Core meeting, sends regrets
Guus: we have 5 deliverables mentioned in the
WG charter
... one of the 5 is explicitly expected to be a Recommendation
... the other 4 we have the ability to decide whether the end state is a Note or a Recommendation
Ralph: As one is working on document, publishing for public consumption, expectation that Editors' Drafts are visible. When published, they show up on Reports page. If our intended goal is Note, not a Recommendation, we are encouraged not to call the documents "Working Drafts". We publish versions not in final form - generally important to inform community about level of stability of documents (paragraphs about status). If we decide to go to Recommendation or to go to Note makes a difference in how we proceed.
Ralph: there's some consideration about how we name interim versions of a published technical report if we're intending the end state to be Note or Recommendation
Guus: the WG charter contains a schedule but it's up to the WG to fill in the details
Guus: Charter is there to give us clear scoping - changes can be proposed and are discussed in Semantic Web Coordination Group.
Guus: as chairs, part of our responsibility is to determine whether a proposal is in or out of scope for our charter
Guus: see -> charter
deliverables
... 5 deliverables mentioned: SKOS, ...
... we have the most input on this deliverable
... we need to draft a use cases and requirements document for SKOS
... this UCR document will be the basis for judging whether the proposed SKOS Recommendation is complete
... hope for use cases from Stanford and other groups
Daniel: absolutely
<TomB> Alistair made a presentation on requirements for SKOS last week in Mexico
Guus: I expect the WG to work on SKOS UCR for the rest of this year
Tom: see Alistair's presentation last week at DC2006 on SKOS requirements
Guus: W3C Notes have a slightly lesser status; they give guidelines but are not necessarily standards
Guus: second deliverable is a Note or
Recommendation on best practices for publishing RDF and OWL ontologies
... third deliverable is guidelines for namespaces and versioning
... we do not yet have a draft for this
Ralph: there are some collected informal notes from the Best Practices WG though not a published document
Guus: fourth deliverable is a document on using
OWL for cross-application integration
... Mike Uschold did some draft work on this
... fifth deliverable is joint between SWD and the HTML WG on incorporating RDF into HTML documents
... there are existing Working Drafts on this work as well
Ralph: use cases and requirements for RDFa are important to do soon also
Guus: for next telecon, please take a look at
Alistair's presentation
... we'll talk about approach to gathering UCR documents next week
Guus: Next week, start with SKOS requirements - how to approach the task of getting together a use-case and requirements document.
Guus: we encourage all communication within the
WG to go via the mailing list
... if we decide to break up the work into subgroups we may decide on certain Subject tags
... we discourage bilateral communication off-list
... W3C tries to keep the process as open as possible; people outside the group should be able to see the discussion, the rationale, and so on
... I would like to have input on any missing expertise within the WG
... for example, to move SKOS to Recommendation we want to show there are tools that handle SKOS and we need to think about testing
Daniel: what do you mean by "handle"; for example, you can bring SKOS into Protege
Guus: in WebOnt we defined some features and
test suites
... for every feature we wanted two implementations
... in SKOS I would assume that a successful implementation would read in SKOS and interpret it in the intended way
... SKOS may be the first non-representational SemWeb language that the W3C is doing; it is more of a pattern
... our test suite should be easily usable by tool developers
... and help measure whether our specification fulfills the requirements we initially laid out
... perhaps in a future telecon Sean can describe how this worked in WebOnt?
Sean: sure
Guus: Normally, f2f meetings are for making key decisions, such as SKOS requirements.
Guus: Tom and I will do some initial planning
on face-to-face scheduling and venue
... f2f usually are two days and try to take advantage of colocation with other events
Ralph: Next Technical Plenary - week where all
WGs encouraged to meet - will be Nov 2007
...smaller version of that being arranged for Jan 2007 - week of Jan 22
...we could have space there. Will be at MIT. Smaller than Tech Plenary.
...If we want that space, need to say so very soon.
<TomB> Jan 22+ would be inconvenient
[straw poll for January meeting]
Ralph: in favor of meeting that week
Tom: neutral but inconvenient for me
Sean: in favor but timing not great
Fabien: I would be able to attend
Bernard: fine for me
Daniel: unsure
Guus: not sure
Diego: ok for me
Antoine: unsure, depends on whether there is another meeting next to it
Ralph: will take an option in order to reserve the possibility
ACTION: Ralph communicate to January meeting planners our desire to keep the option for f2f open for a couple more weeks [recorded in]
next meeting: 17 Oct, 1500 UTC
[adjourned] | http://www.w3.org/2006/10/10-swd-minutes.html | CC-MAIN-2018-05 | refinedweb | 1,299 | 58.62 |
So you might be aware of CSS Custom Properties that let you set a variable, such as a theme color, and then apply it to multiple classes like this:
:root { --theme: #777; } .alert { background: var(—-theme); } .button { background: var(—-theme); }
Well, I had seen this pattern so often that I thought Custom Properties could only be used for color values like rgba or hex – although that’s certainly not the case! After a little bit of experimentation and sleuthing around, I realized that it’s possible to use Custom Properties to set image paths, too.
Here’s an example of that in action:
:root { --errorIcon: url(error.svg) } .alert { background-image: var(--errorIcon); } .form-error { background-image: var(--errorIcon); }
Kinda neat, huh? I think this could be used to make an icon system where you could define a whole list of images in the
:root and call it whenever you needed to. Or you could make it easier to theme certain classes based on their state or perhaps in a media query, as well. Remember, too, that custom properties can be overridden within an element:
:root { --alertIcon: url(alert-icon.svg) } .alert { background-image: var(--alertIcon); } .form-error { --alertIcon: url(alert-icon-error.svg) background-image: var(--alertIcon); }
And, considering that custom properties are selectable in JavaScript, think about the possibilities of swapping out images as well. I reckon this might useful to know!
Yep, they can store URL references alright.
And also much more. They were considered for CSS “mixins” too (the now abandoned @apply rule proposal), because they can store whole CSS rules, so you could write
And you can do this right now and that CSS would be valid. Not that it’s of any use, alas… Not without JavaScript, at least.
And by the way, custom properties can also store strings or even JavaScript snippets…
You reckon correctly! Thanks for the tip
One gotcha is that the URL will be resolved relative to the file where you use the custom property, not relative to the file where you’ve defined them… That makes it a little bit if you want to override custom properties in another file for e.g. theming
You can also do this:
Why bother?
Saves an extra download, enables you to keep all of your icons in one easy to read file.
Note that this only works if your SVGs are clean and hand crafted. In this example the search icon is just a circle and a line, if it was regurgitated from a 1990’s style desktop publishing program as an SVG it would have all kinds of cruft in there as well as the circle defined as scores of points, all specified to nine decimal places which is a bit over the top when you can just type circle with radius and stroke width.
You can also do a filter on your icons, so if the above was set with white as the stroke then in the CSS you could push the brightness down to 0.1 to make it black.
The namespace is required for in CSS SVG whereas it is not needed for inline CSS.
The backslash on each line keeps it legible and enables line breaks.
Any references to colours with a hex code need the hash character escaped as %23. Hence in this example there are named colours (black).
The title attribute can be used too, this helps identify your icons.
Can you keep the variables in a separate file for easy theming?
Absolutely — if you’re using a preprocessor like Sass, then you can keep them in a partial and attach them to the
:rootor some other top-level element or parent component where they’ll be used. | https://css-tricks.com/did-you-know-that-css-custom-properties-can-handle-images-too/ | CC-MAIN-2021-10 | refinedweb | 623 | 70.23 |
Sample 2: a first interactive application¶
Here is a simple interactive application.
Content outline: Pepper asks to be tickled, laughs if tickled, and exits after fifteen seconds without being tickled.
Guided Tour¶
Let’s discover how it works.
The main behavior of this application contains 7 boxes:
5 Standard Boxes: 3 Animated Say boxes (tickle me, another time and that was fun), Show Image box and a Wait box.
2 custom boxes: React to Tickles and Countdown.
Global overview
As a general overview, the behavior does the following:
The robot says “Tickle me!” and displays an image in his tablet that will stay during the whole behavior.
The robot reacts to the “tickle” every time his tablet is touched.
The behavior will exit if the robot’s tablet is not touched for 15 seconds.
Before exiting, the robot will say “That was fun!” if the tablet was touched at least once, otherwise he will say “Okay, another time then”.
React to Tickles¶
Aim
Ensure that Pepper:
- listens to the tablet’s On touch down event,
- reacts by playing a sound and/or animation every time the previous event is raised,
- does not play several sounds at the same time,
- does not play several animations at the same time.
How it works
This box will be triggered after the first Animated Say box, named tickle me. This box has a onTickle output that will get triggered once the robot detects and reacts to the tablet touch. When this output is triggered, the Wait box that was started in parallel is stopped. Note that the Wait box is used to set a timer that will allow to exit automatically in case the tablet is not touched during the behavior’s first fifteen seconds.
Double-click the React to Tickles box to open it.
This box contains a diagram with 3 boxes: a Touch Detection box, and 2 custom boxes: Play Laugh Sound and Play Laugh Animation.
The Touch Detection box has been parametrized to listen to the On touch down tablet event. Both custom boxes are directly linked to the onTouched output of the Touch Detection box. This means that both will be triggered after the previous event has been detected.
The Play Laugh Animation box is in charge of playing randomly one of the animations contained in the animations folder and that are tagged as tickle. This box also prevents playing an animation if another one is already being played.
Double-click the Play Laugh Animation box to open it.
class MyClass(GeneratedClass): def onLoad(self): self.finished = True self.animPlayer = ALProxy("ALAnimationPlayer") def onInput_onStart(self): if self.finished: self.finished = False self.animPlayer.runTag(self.packageUid() + "/animations/", "tickle") self.finished = True self.onStopped()
The Play Laugh Sound is in charge of playing randomly one of the chosen Sounds contained in Aldebaran’s Soundset. This box also prevents playing a sound if another one is already being played.
Double-click the Play Laugh Sound box to open it.
import random class MyClass(GeneratedClass): def onLoad(self): self.finished = True self.audioPlayer = ALProxy("ALAudioPlayer") self.soundSet = self.getParameter("SoundSet") self.sounds = self.getParameter("Sounds") def onInput_onStart(self): if self.finished: self.finished = False # Verify if the Soundset is installed and loaded if self.soundset in self.audioPlayer.getLoadedSoundSetsList(): sound = random.choice(self.sounds.split()) self.audioPlayer.playSoundSetFile(self.soundset, sound) else: self.log("Soundset not installed: " + str(self.soundset)) self.finished = True self.onStopped()
Notice that the SoundSet and the Sounds are defined as parameters.
Each time a sound has been played, the React to Tickles output is triggered since this means that the robot has reacted to a tablet touch.
Countdown¶
Aim
Ensure that Pepper:
- does not stay blocked in the behavior when the tablet is not touched anymore.
How it works
This box functions as a timer that is reset when the React to Tickles box output is triggered. If the timer reaches to its end, it means that the tablet has not been touched for fifteen seconds and the behavior must be stopped.
Double-click the Countdown box to open it.
class MyClass(GeneratedClass): def onLoad(self): import threading self.running = False self.lock = threading.Lock() def onUnload(self): self.running = False def onInput_onStart(self): import time with self.lock: # Regardless of whether the box is already running or not, reset it self.seconds_left = self.getParameter("Seconds") if self.running: # the box is already running return self.running = True # This part of the code will only be reached by the first call. while self.running and self.seconds_left > 0: time.sleep(1.0) self.seconds_left -= 1 # timeout is finished, trigger output. with self.lock: if self.running: self.running = False self.onStopped() def onInput_onStop(self): self.onUnload() # clean up
Some notes on this box:
- Its time is approximate (when reset, it will not take exactly fifteen seconds before its output is triggered).
- This box blocks one thread while running.
Try it!¶
Try the application.
You may also try the behavior only, by clicking on the
Play button.
Note that this example only works correctly on a real Pepper, since Soundset Aldebaran and ALTabletService are not present on a virtual robot.
Make it yours!¶
Try changing the reaction of the robot or adding another Touch box, like the Tactile Head box. | http://doc.aldebaran.com/2-4/getting_started/samples/sample_interactive.html | CC-MAIN-2019-13 | refinedweb | 881 | 59.9 |
Mine is a hack job to make the Servicemix mapper understand a
SoapMessage that has a Body consisting of an ever changing
XML Document. I am unsure if my change is going
to be an acceptable alteration, although it is working thus
far. I am still experiencing problems when I try to alter the
contents of the Document in the Body via a XSL transformation.
I get so much exception dump I dunno if its the processing of
my document or my code changes to the mapper... :)
Im truely hackin it as I go. Its my opinion that what IM doing
is NOT the correct solution. However, my trials and tribulations
may be of use to an authoritive ODE developer. I have some
BPEL that attems to do a "doXSLtransfrom on the SOAPMEssage
and its not working yet.... However, I am past the point where
the ODE engine dies because it cant find a MAPPER. Any help
is greatly appreciated. I have also noted that the "getUserData" gets
stomped on somewhere in the mapper process, but I have not
found where. SO I have chosen to avoid using that XML feature.
TIA
Alex Boisvert wrote:
>
> I'm all for extending the usefulness of the mappers... if you want to
> share your extensions/customizations we can bring them back into Ode
> (assuming they can apply to others).
>
> Simply open a jira issue with your code (even if experimental) and a
> description of what it does and what you're trying to achieve.
>
> alex
>
> On 8/31/07, jbi joe <joe@daggerpoint.net> wrote:
>>
>> I have done the same thing.
>> WHy isnt there more flexibility with mappers?
>> I need a mapper to be able to process soapmessages
>> that contain Documents with attachments and user data.
>>
>>
>> eburgos wrote:
>> >
>> > Tried in several ways and I got the same error, I ended up temporarily
>> > modifying ServiceMixMapper class cause I'm in a hurry :(
>> >
>> >
>> >
>> > On 8/28/07, Alex Boisvert <boisvert@intalio.com> wrote:
>> >>
>> >> Then you most likely have a classloading problem: different versions
>> of
>> >> the
>> >> Mapper interface in different classloaders. The easiest solution is
>> to
>> >> place your Mapper class in the ode-jbi.jar.
>> >>
>> >> alex
>> >>
>> >>
>> >> On 8/28/07, Eduardo Burgos <eburgos@gmail.com> wrote:
>> >> >
>> >> > I see this in OdeLifeCycle.java line 148
>> >> >
>> >> > try {
>> >> > _ode.registerMapper((Mapper) mapperClass.newInstance());
>> >> > } catch (Throwable t) {
>> >> > String errmsg =
>> >> >
>> >> >
>> >>
>> __msgs.msgOdeInitMapperInstantiationFailed(_ode._config.getMessageMapper());
>> >> > __log.error(errmsg);
>> >> > throw new JBIException(errmsg, t);
>> >> > }
>> >> >
>> >> > it catches an exception there, in the debugger I saw that Throwable
>> t
>> >> was
>> >> > actually a ClassCastException. I tried the class with a public no
>> >> argument
>> >> > constructor, the errors differ a little but it ends up being a
>> >> > ClassCastException at the same point too. I know for sure that my
>> class
>> >> is
>> >> > a
>> >> > Mapper because it inherits the ServiceMixMapper.
>> >> >
>> >> >
>> >> >
>> >> > On 8/28/07, Alex Boisvert <boisvert@intalio.com> wrote:
>> >> > >
>> >> > > Does your class have a public no-argument constructor? Ode uses
>> >> > > Class.newInstance() to create instances of Mapper classes. Also,
>> >> check
>> >> > > your log file again, I think there must be a more specific error
>> >> > > just above/below the error statement you sent.
>> >> > >
>> >> > > alex
>> >> > >
>> >> > >
>> >> > > On 8/28/07, Eduardo Burgos <eburgos@gmail.com> wrote:
>> >> > > >
>> >> > > > Is there a way to change the default ServicemixMapper to
another
>> >> class
>> >> > > of
>> >> > > > your own? It seems that editing ode-jbi.properties wont do
>> >> much.Iedited
>> >> > > > with my own class name and I got this at startup:
>> >> > > >
>> >> > > > ERROR - ComponentMBeanImpl - Could not start
>> component
>> >> > > > javax.jbi.JBIException: Message mapper class "
>> >> > > > org.test.ode.CustomOdeServicemixMapper" could not be
>> instantiated!
>> >> > > > at org.apache.ode.jbi.OdeLifeCycle.initMappers(
>> >> > OdeLifeCycle.java
>> >> > > > :153)
>> >> > > > at
>> org.apache.ode.jbi.OdeLifeCycle.init(OdeLifeCycle.java
>> >> :104)
>> >> > > > at
>> >> org.apache.servicemix.jbi.framework.ComponentMBeanImpl.init
>> >> > (
>> >> > > > ComponentMBeanImpl.java:200)
>> >> > > > at
>> >> > > org.apache.servicemix.jbi.framework.ComponentMBeanImpl.doStart(
>> >> > > > ComponentMBeanImpl.java:286)
>> >> > > > at
>> >> > org.apache.servicemix.jbi.framework.ComponentMBeanImpl.start(
>> >> > > > ComponentMBeanImpl.java.beanutils.MethodUtils.invokeMethod(
>> >> > > > MethodUtils.java:216)
>> >> > > >
>> >> > > >
>> >> > > > My class goes as follows:
>> >> > > >
>> >> > > > public class CustomOdeServicemixMapper extends ServiceMixMapper
>> {
>> >> > > >
>> >> > > >
>> >> > > > @Override
>> >> > > > public void toNMS(NormalizedMessage nms, Message odeMsg,
>> >> > > > javax.wsdl.Message msgdef, QName fault) throws
>> MessagingException,
>> >> > > > MessageTranslationException {
>> >> > > >
>> >> > > > //Some code here
>> >> > > >
>> >> > > > super.toNMS(nms, odeMsg, msgdef, fault);
>> >> > > > }
>> >> > > > }
>> >> > > >
>> >> > > >
>> >> > > >
>> >> > > >
>> >> > > > Any ideas?
>> >> > > >
>> >> > > > Thanks in advance,
>> >> > > >
>> >> > >
>> >> >
>> >>
>> >
>> >
>>
>> --
>> View this message in context:
>>
>> Sent from the Apache Ode User mailing list archive at Nabble.com.
>>
>>
>
>
--
View this message in context:
Sent from the Apache Ode User mailing list archive at Nabble.com. | http://mail-archives.eu.apache.org/mod_mbox/ode-user/200709.mbox/%3C12438619.post@talk.nabble.com%3E | CC-MAIN-2019-26 | refinedweb | 714 | 50.43 |
Create BizTalk Services using the Azure portal
Tip
To sign in to the Azure portal, you need an Azure account and Azure subscription. If you don't have an account, you can create a free trial account within a few minutes. See Azure Free Trial.
Create a BizTalk Service
Depending on the Edition you choose, not all BizTalk Service settings may be available.
- In the bottom navigation pane, select NEW:
- Select APP SERVICES > BIZTALK SERVICE > CUSTOM CREATE:
Enter the BizTalk Service settings:
Enter the Storage and Database Settings:
Enter the Database settings:
Select the check mark to complete the wizard. The progress icon appears:
When complete, the Azure BizTalk Service is created and ready for your applications. The default settings are sufficient. If you want to change the default settings, select BIZTALK SERVICES in the left navigation pane, and then select your BizTalk Service. Additional settings are displayed in the Dashboard, Monitor, and Scale tabs at the top.
Depending on the state of the BizTalk Service, there are some operations that cannot be completed. For a list of these operations, go to BizTalk Services State Chart.
Post-provisioning steps
- Install the certificate on a local computer
- Add a production-ready certificate
- Get the Access Control namespace
Install the certificate on a local computer
As part of BizTalk Service provisioning, a self-signed certificate is created and associated with your BizTalk Service subscription. You must download this certificate and install it on computers from where you either deploy BizTalk Service applications or send messages to a BizTalk Service endpoint.
- Select BIZTALK SERVICES in the left navigation pane, and then select your BizTalk Service subscription.
- Select the Dashboard tab.
- Select Download SSL Certificate:
- Double-click the certificate and run through the wizard to install the certificate. Make sure you install the certificate under the Trusted Root Certificate Authorities store.
Add a production-ready certificate
The self-signed certificate that is automatically created when creating BizTalk Services is intended for use in development environments only. For production scenarios, replace it with a production-ready certificate.
- On the Dashboard tab, select Update SSL Certificate.
- Browse to your private SSL certificate (CertificateName.pfx) that includes your BizTalk Service name, enter the password, and then click the check mark.
Get the Access Control namespace
- Select BIZTALK SERVICES in the left navigation pane, and then select your BizTalk Service.
- In the task bar, select Connection Information:
- Copy the Access Control values.
When you deploy a BizTalk Service project from Visual Studio, you enter this Access Control namespace. The Access Control namespace is automatically created for your BizTalk Service.
The Access Control values can be used with any application. When Azure BizTalk Services is created, this Access Control namespace controls the authentication with your BizTalk Service deployment. If you want to change the subscription or manage the namespace, select ACTIVE DIRECTORY in the left navigation pane and then select your namespace. The task bar lists your options.
Clicking Manage opens the Access Control Management Portal. In the Access Control Management Portal, the BizTalk Service uses Service identities:
The Access Control service identity is a set of credentials that allow applications or clients to authenticate directly with Access Control and receive a token.
Important
The BizTalk Service uses Owner for the default service identity and the Password value. If you use the Symmetric Key value instead of the Password value, the following error may occur.
Could not connect to the Access Control Management Service account with the specified credentials
Managing Your ACS Namespace lists some guidelines and recommendations.
Requirements explained
These requirements do not apply to the Free Edition.
Hybrid Connections
When you create an Azure BizTalk Service, the Hybrid Connections tab is available:
Hybrid Connections are used to connect an Azure website or Azure mobile service to any on-premises resource that uses a static TCP port, such as SQL Server, MySQL, HTTP Web APIs, Mobile Services, and most custom Web Services. Hybrid Connections and the BizTalk Adapter Service are different. The BizTalk Adapter Service is used to connect Azure BizTalk Services to an on-premises Line of Business (LOB) system.
See Hybrid Connections to learn more, including creating and managing Hybrid Connections.
Next steps
Now that a BizTalk Service is created, familiarize yourself with the different BizTalk Services: Dashboard, Monitor and Scale tabs. Your BizTalk Service is ready for your applications. To start creating applications, go to Azure BizTalk Services. | https://docs.microsoft.com/en-us/azure/biztalk-services/biztalk-provision-services | CC-MAIN-2017-09 | refinedweb | 733 | 54.42 |
This is a functional watch that I made for my nephew based on the Minecraft clock. If you or someone you know is really into Minecraft, this watch is sure to be a welcomed novelty!
The idea is based on the clock item in Minecraft. Essentially it just shows day and night so that if the player is underground, they can know when it is safe to emerge. The watch we will be making is just a replica of the clock that you can set to tic out seconds, minutes, or hours. The code can be customized to choose how you want it to function.
Step 1: Gathering the Parts
On the hardware side, there are a few parts that you will need to build the watch:
- Adafruit 3.3v Trinket
- 3.7v LiPo Battery
- USB LiPo Battery Charger
- Mini Stepper Motor (Alternate Option)
- Mini Switch
- Sun/Moon graphic
- 3D printed case
- Soldering/Hot Glue equipment
Total Cost: Approx $20
On the software side, we will be using the Arduino IDE to load our program to the Trinket, so make sure you have it downloaded and installed. Also, in order to make it communicate with the trinket, there are a few modifications you will need to make, which are covered on the Adafruit website.
Step 2: Getcher' Motor Runnin'
Now that we have are parts, the first thing to do is connect the stepper motor to the Adafruit Trinket. In pretty much every other situation when you are dealing with stepper motors, you will need a stepper motor driver. However, since our stepper motor is extremely tiny and requires very little power, in this rare instance, it can be driven directly by the Trinket.
Your stepper motor should have four output terminals, and depending on the motor you've purchased, those terminals may be different than mine. As you can see in the diagram above, the two top terminals connect one coil, and the two bottom terminals connect the other one. If you are unsure how your stepper is connected, you can use a multimeter to see which ones pass current.
Once you have your terminals established, you can connect it to the Trinket as layed out in the diagram above. The first coil set connects to pins #3 and #1 on the Trinket, while the remaining coil set connects to pins #0 and #4.
With the motor connected, we can now load up the software. The Arduino code is extremely simple. Essentially, you set up a class for the stepper motor, and then tell it how man times you want it to move per second. As a demonstration, I just set mine to delay for 1,000ms or 1 second so that each tic was easily visible. If you want to to be more accurate to day/night revolutions, you can set it to 300,000ms, which is 50 minutes (I didn't use 60 minutes, because the stepper only has 20 steps to make a full revolution). Here is a copy of my code (DOWNLOAD IT HERE):
#include <Stepper.h>
#define STEPS 720 // steps per revolution (limited to 315°)
#define COIL1 1
#define COIL2 3
#define COIL3 4
#define COIL4 0
// create an instance of the stepper class:
Stepper stepper(STEPS, COIL1, COIL2, COIL3, COIL4);
void setup(){
stepper.setSpeed(30); // set the motor speed to 30 RPM (360 PPS aprox.).
stepper.step(630); //Reset Position(630 steps counter-clockwise).
}
int pos=0; //Position in steps(0-630)= (0°-315°)
void loop(){
stepper.step(-1); // move one step to the left (change to 1 to move to right).
delay(1000); //1,000ms = 1 sec | 300,000ms will give an accurate day/night tic) pos++;
}
Step 3: Give It Some Juice
At this point, you should have a tiny stepper that is run by the Trinket at tics at your specified intervals. Now we need to make it mobile. To do that, we need to replace the Trinket's USB power with a battery. We'll be using a small 3.7v rechargable LiPo battery, which means we'll also need a way to charge it. Enter the USB LiPo Battery charger. This charger should allow you to connect the battery and the Trinket to it on one end, and a micro-usb cable on the other end to charge the battery. It also has the added benefit of protecting the battery and the Trinket from power issues.
Since it's not really the best idea to charge the battery while it's being used, you could also add a switch to the circuit to turn the Trinket off and on. Now the trick is to solder all those components together using a minimal amount of wire. You can use the images for this step as guides.
Step 4: Big Power, Little Case
This final step is to take all of the electronics that we've created and cram it into the 3D printed case for the watch. The USB charger and Trinket should snuggly fit perpendicular to each other. then you can just hot glue the rest of the components into place. Print out the day/night image, cut it out, and affix it to the motor head. Place the cover on top, and paint it a nice golden color to match the game version. Now you can add a regular watch band and show it off to your friends!
5 Discussions
2 years ago
Socool
2 years ago
Apple watch parody *AMAZING!* "We do not want to make just a watch..."
3 years ago
Question: Isn't that the wrong charging circuit for the LiPo battery as that charging circuit you linked is for a Li Ion battery?
Reply 3 years ago
The one I ordered would work fine with LiPo or Li Ion, but to avoid confusion, I changed the link to a LiPo charger. Thanks for the feedback!
3 years ago
I laughed out loud at your intro video!! Hilarious!!! Nice job! | https://www.instructables.com/id/Minecraft-Watch/ | CC-MAIN-2019-26 | refinedweb | 996 | 78.18 |
i above code, i used the different codepages, but my output from the printer is just question mark ""?"" instead of the persian words
Code: Select all
#!/usr/bin/env python # -*- coding: utf-8 -*- #" #设置解码的类型 # Some variables #fontPath = "/usr/share/fonts/opentype/fonts-hosny-thabit/Thabit.ttf" textUtf8 = u"بعض النصوص من جوجل ترجمة"()
i tried the following code,too:
but the printer prints the obscure letters...
Code: Select all
from escpos.printer import Usb """ Seiko Epson Corp. Receipt Printer M129 Definitions (EPSON TM-T88IV) """ p = Usb(0x04b8,0x0e15,0) p.codepage="iso8859_6" # Print text p.text(u"سلام\n") p.cut()
i tried the different codepages..but it wasn't usefull | https://www.raspberrypi.org/forums/viewtopic.php?f=32&p=1198009&sid=138213f17befa5c3528da0ca0903a8ef | CC-MAIN-2018-09 | refinedweb | 111 | 62.95 |
When I was learning about closures in JavaScript for the first time I saw a lot of examples like this:
const counter = (function() {let count = 0return function() {count++return count}})()console.log(counter()) //1console.log(counter()) //2console.log(counter()) //3
Now, it was pretty clear to me what was happening; the outer function (which is an Immediately Invoked Function Expression or IIFE for short) declares the
count variable, assigns the number zero to it and then returns the inner function. This inner function is executed each time we call
count() and 'knows' about the
count variable because it has closure over it.
So all well and good. The bit that was not clear to me was why this would be useful! As such, I would like to demonstrate a couple of ways in which I use closures in my daily coding.
The first technique is encapsulation. This is a way of creating a module where the inner workings are private and only certain methods are exposed for other modules to interact with them. It is exactly what is happening in the example above but it is (hopefully!) easier to see why it is useful when we look at some code which has a bit more going on.
const library = (function() {const books = ['Island', '1984', 'Dracula', 'Papillon']return {getBooks() {return [...books]},deleteBook(title) {books.splice(books.indexOf(title), 1)},addBook(title) {books.push(title)},}})()library.addBook('catch 22')library.deleteBook('Dracula')console.log(library.getBooks()) // [ Island, 1984, Papillon, catch 22 ]const libraryBooks = library.getBooks()libraryBooks.push('Beowulf')console.log(libraryBooks) // [ Island, 1984, Papillon, catch 22, Beowulf ]console.log(library.getBooks()) // [ Island, 1984, Papillon, catch 22 ]
As you can see, our
books array is protected from the outside world by the inner function's closure over it. The only way to interact with it is through the public methods;
getBooks() returns a copy of the array instead of a reference to it and books can only be added or deleted one at a time.
The second example is a pattern I use frequently. I will use a React component to demonstrate, don't worry if you don't know React, it is pretty easy to see what is happening.
class StartEndDate extends React.Component {state = {start: null,end: null}// This is our closure, I have used es6 arrow functions this timeupdateTime = startOrEnd => newTime => (this.setState({[startOrEnd]: newTime}))render() {return (<DateInput label="From" updateTime={this.updateTime('start')} /><DateInput label="To" updateTime={this.updateTime('end')} />)}}function DateInput({ label, updateTime }) {return (<label for="timeInput">From</label><input id="timeInput" type="time" onChange={updateTime}>)}
In this code we have one stateful React component, which is declared as an es6 class and one functional component. You should be able to see where the functional component is used, it is in the return of the
render method (the code that looks like HTML is actually JSX; it calls the
DateInput function, passing the props
label and
updateTime to it).
The
DateInput function returns the markup for a
<label> and
<input>. When a change event is detected, the value is passed through to the
updateTime method on the
StartEndDate class.
If you look at the
render method where this function is passed to
DateInput, it is first called with either the value "start" or "end". Our two inputs now have their own version of the
updateTime method which has closure over the
startOrEnd variable. Using computed property keys we can dynamically assign the new value to the component's state using the variable in closure.
These are two of my most often used closure patterns. I hope that I have been able to show why they are such a useful feature of the language and an important tool to have in your tool-set.
If you've found this helpful then let me know with a clap or two! | https://blog.matt-thorning.dev/using-closure | CC-MAIN-2021-04 | refinedweb | 640 | 53.21 |
Copyright © 2011 W3C® (MIT, ERCIM, Keio), All Rights Reserved. W3C liability, trademark and document use rules apply.
This CSS3 module describes the various values and units that CSS properties accept. Also, it describes how values are computed from "specified" through "computed" and "used" into "actual" values. The main purpose of this module is to define common values and units in one specification which can be referred to by other modules. As such, it does not make sense to claim conformance with this module features described in this specification that also exist in CSS 2.1 [CSS21] are intended to be backwards compatible. In case of conflict between this draft and CSS 2.1 [CSS21], CSS 2.1 probably represents the intention of the CSS WG better than this draft (other than on values and units that are new to CSS3).
The following features are at-risk and may be dropped during the CR
period: ‘
vh’,
‘
vw’, ‘
vm’, ‘
fr’, ‘
gr’, ‘
cycle()’, ‘
attr()’.
initial’ and ‘
inherit’
<identifier>’ type
<string>’ type
<url>’ type
<integer>’ type
<number>’ type
<percentage>’ type
<length>’ type
em’, ‘
ex’, ‘
ch’, ‘
rem’ units
vw’, ‘
vh’, ‘
vm’
<color>’ type
<image>’ type
<fraction>’ type and ‘
fr’ unit
<grid>’ type and ‘
gr’ unit
calc()’, ‘
min()’ and ‘
max()’ the value]
Should these keywords affect the specified or computed value? See various issues.
Would it be useful to have a value defined to be equivalent
to ‘
inherit’ for properties that are inherited
by default and equivalent to ‘
initial’ for properties that are not
inherited by default? This might be easier for authors to use than ‘
initial’ and ‘
inherit’
since it wouldn't require thinking about whether a property is inherited
by default or not (which isn't obvious for some properties, such as
text-decoration and visibility).
It's been requested to have a value that rolls back to the
bottom of that level of the cascade, e.g. for an author rule it would roll
back to the end of the user cascade, for a user rule it would roll back to
the end of the UA cascade, and for the UA it would roll back to
‘
initial’/‘
inherit’. Is that something we should add?
‘
\"’ or as ‘
\22’). Analogously for single quotes
("\‘
" or "\27").
content: "this is a ’string'."; content: "this is a \"string\"."; content: ‘
this is a "string".’; content: ‘
this is a \’string\‘
.’; (‘
) and double quotes (") appearing in a URL must be escaped with
a backslash so that the resulting value is a valid ’url(open\(parens)‘
URL token, e.g.
,
’url(close\)parens)‘
. Depending on the
type of URL, it might also be possible to write these characters as
URI-escapes (where ’‘
( =
%28,
) =
%29, etc.) as described in [URI]. Alternatively a URL containing
such characters may be represented as a quoted string within the
url()’‘
vm’ units
The viewport-relative lengths are relative to the size of the initial containing block. When the height or width of the viewport is changed, they are scaled proportionally.
In the example below, if the width of the viewport is 200mm, the font
size of
h1 elements will be 16mm (i.e.
(8×200mm)/100).
h1 { font-size: 8vw }
vw’ or ‘
vh’.
Do we need this now that we have the min() function?.
Some data types are defined in their own modules. The two common ones
are
<color> and
<image>.
<color>’ type
The
<color> data type is defined in
[CSS21] and extended in [CSS3COLOR].
UAs that support CSS Color Level 3 must interpret
<color> as defined therein.
<image>’ type
The
<image> data type is
defined herein as equivalent to
<url>. It is extended in [CSS3-IMAGES]: UAs that support
CSS Image Values Level 3 must interpret
<image> as defined therein.
<fraction>’ type and ‘
fr’ unit
The fr unit is used to represent proportions, such as the proportions used to distribute remaining space in a flex layout computation. [CSS3-FLEXBOX] When multiple fractions participate in a calculation, the remainder is distributed proportionally to their numeric value.
border-parts: 10px 1fr 10px; border-parts: 10px 1fr 10px 1fr 10px; border-parts: 10px 2fr 10px 2fr 10px;
<grid>’ type and ‘
gr’ unit
A grid is a set of invisible vertical and horizontal lines that can be
used to align content. In CSS, grid lines can be established implicitly
(as in [CSS3COL])
or explicitly (as in [CSS3GRID]). In either case, the
distance between grid lines can be referred to by the ‘
gr’ unit.
img { float: top left multicol; float-offset: 2gr; width: 1gr; }
Grid lines can be laid out in uneven patterns. Therefore, the ‘
gr’ unit is not linear.
For example, "2gr" is not necessarily twice as long as "1gr".
Some values use a functional notation
to type values and to and lump values together. The syntax starts with the
name of the function immediately followed by a left parenthesis followed
by optional whitespace followed by the argument(s) to the notation
followed by optional whitespace followed by a right parenthesis. If a
function takes more than one argument, the arguments are separated by a
comma (‘
,’) with optional whitespace
before and after the comma.
background: url(); color: rgb(100, 200, 50 ); content: counter(list-item) ". "; width: calc(50% - 2em);
calc()’, ‘
min()’ and ‘
max()’
The calc(), min(), and max() functions can be used wherever
<length>,
<frequency>,
<angle>,
<time>, or
<number> values are allowed.
section { float: left; margin: 1em; border: solid 1px; width: calc(100%/3 - 2*1em - 2*1px); }
p { margin: calc(1rem - 2px) calc(1rem - 1px); }
p { font-size: min(10px, 3em) } blockquote { font-size: max(30px, 3em) }
.box { width: min(10% + 20px, 300px) }
The expression language of these functions is described by the grammar and prose below.
S : calc | min | max; calc : "calc(" S* sum ")" S*; min : "min(" S* sum [ "," S* sum ]* ")" S*; max : "max(" S* sum [ "," S* sum ]* ")" S*; sum : product [ [ "+" | "-" ] S* product ]*; product : unit [ [ "*" | "/" | "mod" ] S* unit ]*; unit : ["+"|"-"]? [ NUMBER S* | DIMENSION S* | PERCENTAGE S* | min | max | "(" S* sum ")" S* ];
The context of the expression imposes a target type, which is one of
length, frequency, angle, time, or number. NUMBER tokens are of type
number. DIMENSION tokens have types of their units (‘
cm’ is length, ‘
deg’ is angle etc.); any DIMENSION whose
type does not match the target type causes the ‘
calc()’ expression to be
invalid. If percentages are accepted in that context and convertible to
the target type, a PERCENTAGE token in the expression has the target type;
otherwise percentages are likewise invalid.
To make expressions simpler, operators have restrictions on the types they accept. At each operator, the types of the left and right side are be checked for these restrictions. If compatible, they return roughly as follows (the following ignores precedence rules on the operators for simplicity):
Division by zero is not allowed. Declarations containing such a construct are invalid and must be ignored.
The value resulting from an expression must be clamped to the range allowed in the target context.
Note this requires all contexts accepting ‘
calc()’ to define their
allowable values as a closed (not open) interval.
width: 0px’ since widths smaller than 0px are not allowed.
width: calc(5px - 10px); width: 0px;
Given the complexities of ‘
width’ and ‘
height’ on table cells and table elements,
calc() expressions for widths and heights on table columns, table column
groups, table rows, table row groups, and table cells in both auto and
fixed layout tables may be treated as if ‘
auto’ had been specified.
cycle()’
The ‘()’, ‘
calc()’, ‘
min()’, or ‘
max()’ notations. Declarations containing such
constructs are invalid and must be ignored.
attr()’
Describe the feature fully here, not just a delta from CSS 21.
When attr is set on a pseudo-element, it should apply to the originating element
In CSS2.1 [CSS21],
the ‘
attr()’ expression always returns
a string. In CSS3, the ‘
attr()’
expression can return many different types. The new syntax for the attr()
expression is:
'attr(' ident [ ',' <type> [ ',' <value> ]? ]? ')'
The first argument represents the attribute name. The value of the attribute with that name on the element whose computed values are being computed is used as the value of the expression, according to the rules given below.
The first argument accepts an optional namespace prefix to identify the
namespace of the attribute. The namespace prefix and the attribute name is
separated by ‘
|’, with no whitespace
before or after the separator [CSS3NAMESPACE].
The second argument (which is optional but must be present if the third argument is present) is a <type> and tells the UA how to interpret the attribute value. It may be one of the values from the list below.
The third argument (which is optional) is a CSS value which must be valid where the attr() expression is placed. If it is not valid, then the whole attr() expression is invalid.
If the attribute named by the first argument is missing, cannot be parsed, or is invalid for the property, then the value returned by attr() will be the third argument, or, if the third argument is absent, will be the value given as the default for the relevant type in the list below.
color’ property..)
Should there also be a "keyword" type to, e.g., support
‘
float: attr(align)’
The attr() form is only valid if the type given (or implied, if it is missing) is valid for the property. For example, all of the following.
Note that the default value need not be of the type given. For instance,
if the type required of the attribute by the author is ‘
px’, the default could still be ‘
5em’.
Examples:
); }
The attr() expression cannot currently fall back onto another attribute. Future versions of CSS may extend attr() in this direction.
Should ‘
attr()’ be
allowed on any property, in any source language? For example, do we expect
UAs to honor this rule for HTML documents?: P[COLOR] { color:
attr(COLOR, color) }.
Shouldn't this section move to [CSS3CASCADE]?
Once a user agent has parsed a document and constructed a document tree, it must assign, for every element in the tree, a value to every property that applies to the target media type.
The final value of a CSS3 property for a given element is the result of a four-step calculation:
The specified value is the output of the cascading and inheritance process. [CSS21] [CSS3CASCADE]
If the output of the cascade is ‘
inherit’ or ‘
initial’, the specified value contains the
inherited or initial value, respectively. See examples (d) and (e) in the
table below. insofar as possible without formatting the document, as defined in the "Computed value" line of the property definition tables.
The computed value is the value that is transferred from parent to child during inheritance.
The computed value exists even when the property does not apply (as
defined by the ‘.
<color>, 7.1.
cycle()’, 9.2.
<identifier>, 3.2.
<image>, 7.2.
inherit’, 3.1.1.
initial’, 3.1.1.
<integer>, 4.1.
<length>, 5.
<number>, 4.2.
<percentage>, 4.3.
<string>, 3.3.
<url>, 3.4. | https://www.w3.org/TR/2011/WD-css3-values-20110906/ | CC-MAIN-2018-09 | refinedweb | 1,819 | 62.27 |
Thanks'd be surprised to hear that the ODBC driver would need to know anything
about namespace-mapping.
Do you have an error? Steps to reproduce an issue which you see?
The reason I am surprised is that namespace mapping is an implementation
detail of the JDBC driver which lives inside of PQS -- *not* the ODBC
driver. The trivial thing you can check would be to validate that the
hbase-site.xml which PQS references is up to date and that PQS was restarted
to pick up the newest version of hbase-site.xml
On 5/22/18 4:16 AM,.
> | https://mail-archives.eu.apache.org/mod_mbox/phoenix-user/201805.mbox/%3Cbed9768a3e958dbba95c661f7ab5582a@mail.gmail.com%3E | CC-MAIN-2021-25 | refinedweb | 101 | 75.1 |
hast-util-raw
hast utility to parse the tree again, now supporting embedded
raw nodes.
One of the reasons to do this is for “malformed” syntax trees: for example, say there’s an
h1 element in a
p element, this utility will make them siblings.
Another reason to do this is if raw HTML/XML is embedded in a syntax tree, which can occur when coming from Markdown using
mdast-util-to-hast.
If you’re working with remark and/or
remark-rehype, use
rehype-raw instead.
Install
This package is ESM only: Node 12+ is needed to use it and it must be
imported instead of
required.
npm install hast-util-raw
Use
import {h} from 'hastscript' import {raw} from 'hast-util-raw' const tree = h('div', [h('h1', ['Foo ', h('h2', 'Bar'), ' Baz'])]) const reformatted = raw(tree) console.log(reformatted)
Yields:
{ type: 'element', tagName: 'div', properties: {}, children: [ { type: 'element', tagName: 'h1', properties: {}, children: [Object] }, { type: 'element', tagName: 'h2', properties: {}, children: [Object] }, { type: 'text', value: ' Baz' } ] }
API
This package exports the following identifiers:
raw. There is no default export.
raw(tree[, file][, options])
Given a hast tree and an optional vfile (for positional info), return a new parsed-again hast tree.
options.passThrough
List of custom hast node types to pass through (keep) in hast (
Array.<string>, default:
[]). If the passed through nodes have children, those children are expected to be hast and will be handled.
Security
Use of
hast-util-raw can open you up to a cross-site scripting (XSS) attack as
raw nodes are unsafe. The following example shows how a raw node is used to inject a script that runs when loaded in a browser.
raw(u('root', [u('raw', '<script>alert(1)</script>')]))
Yields:
<script>alert(1)</script>
Do not use this utility in combination with user input or use
hast-util-santize.
Related
mdast-util-to-hast— transform mdast to hast
rehype-raw— wrapper plugin for rehype
Contribute
See
contributing.md in
syntax-tree/.github for ways to get started. See
support.md for ways to get help.
This project has a code of conduct. By interacting with this repository, organization, or community you agree to abide by its terms. | https://unifiedjs.com/explore/package/hast-util-raw/ | CC-MAIN-2022-05 | refinedweb | 367 | 55.24 |
Write your own miniature Redis with Python
December 28, 2017 15:24 / gevent python redis / 3 comments
The other day the idea occurred to me that it would be neat to write a simple Redis-like database server. While I've had plenty of experience with WSGI applications, a database server presented a novel challenge and proved to be a nice practical way of learning how to work with sockets in Python. In this post I'll share what I learned along the way.
The goal of my project was to write a simple server that I could use with a task queue project of mine called huey. Huey uses Redis as the default storage engine for tracking enqueued jobs, results of finished jobs, and other things. For the purposes of this post, I've reduced the scope of the original project even further so as not to muddy the waters with code you could very easily write yourself, but if you're curious, you can check out the end result here.
The server we'll be building will be able to respond to the following commands:
- GET
<key>
- SET
<key>
<value>
- DELETE
<key>
- FLUSH
- MGET
<key1>...
<keyn>
- MSET
<key1>
<value1>...
<keyn>
<valuen>
We'll support the following data-types as well:
- Strings and Binary Data
- Numbers
- NULL
- Arrays (which may be nested)
- Dictionaries (which may be nested)
- Error messages
To handle multiple clients asynchronously, we'll be using gevent, but you could also use the standard library's SocketServer module with either the ForkingMixin or the ThreadingMixin.
Skeleton
Let's frame up a skeleton for our server. We'll need the server itself, and a callback to be executed when a new client connects. Additionally we'll need some kind of logic to process the client request and to send a response.
Here's a start:
from gevent import socket from gevent.pool import Pool from gevent.server import StreamServer from collections import namedtuple from io import BytesIO from socket import error as socket_error # We'll use exceptions to notify the connection-handling loop of problems. class CommandError(Exception): pass class Disconnect(Exception): pass Error = namedtuple('Error', ('message',)) class ProtocolHandler(object): def handle_request(self, socket_file): # Parse a request from the client into it's component parts. pass def write_response(self, socket_file, data): # Serialize the response data and send it to the client. pass class Server(object): def __init__(self, host='127.0.0.1', port=31337, max_clients=64): self._pool = Pool(max_clients) self._server = StreamServer( (host, port), self.connection_handler, spawn=self._pool) self._protocol = ProtocolHandler() self._kv = {} def connection_handler(self, conn, address): # Convert "conn" (a socket object) into a file-like object. socket_file = conn.makefile('rwb') # Process client requests until client disconnects. while True: try: data = self._protocol.handle_request(socket_file) except Disconnect: break try: resp = self.get_response(data) except CommandError as exc: resp = Error(exc.args[0]) self._protocol.write_response(socket_file, resp) def get_response(self, data): # Here we'll actually unpack the data sent by the client, execute the # command they specified, and pass back the return value. pass def run(self): self._server.serve_forever()
The above code is hopefully fairly clear. We've separated concerns so that the protocol handling is in it's own class with two public methods:
handle_request and
write_response. The server itself uses the protocol handler to unpack client requests and serialize server responses back to the client. The
get_response() method will be used to execute the command initiatied by the client.
Taking a closer look at the code of the
connection_handler() method, you can see that we obtain a file-like wrapper around the socket object. This wrapper allows us to abstract away some of the quirks one typically encounters working with raw sockets. The function enters an endless loop, reading requests from the client, sending responses, and finally exiting the loop when the client disconnects (indicated by
read() returning an empty string).
We use typed exceptions to handle client disconnects and to notify the user of errors processing commands. For example, if the user makes an improperly formatted request to the server, we will raise a
CommandError, which is serialized into an error response and sent to the client.
Before going further, let's discuss how the client and server will communicate.
Wire protocol
The first challenge I faced was how to handle sending binary data over the wire. Most examples I found online were pointless echo servers that converted the socket to a file-like object and just called
readline(). If I wanted to store some pickled data or strings with new-lines, I would need to have some kind of serialization format.
After wasting time trying to invent something suitable, I decided to read the documentation on the Redis protocol, which turned out to be very simple to implement and has the added benefit of supporting a couple different data-types.
The Redis protocol uses a request/response communication pattern with the clients. Responses from the server will use the first byte to indicate data-type, followed by the data, terminated by a carriage-return/line-feed.
Let's fill in the protocol handler's class so that it implements the Redis protocol.
class ProtocolHandler(object): def __init__(self): self.handlers = { '+': self.handle_simple_string, '-': self.handle_error, ':': self.handle_integer, '$': self.handle_string, '*': self.handle_array, '%': self.handle_dict} def handle_request(self, socket_file): first_byte = socket_file.read(1) if not first_byte: raise Disconnect() try: # Delegate to the appropriate handler based on the first byte. return self.handlers[first_byte](socket_file) except KeyError: raise CommandError('bad request') def handle_simple_string(self, socket_file): return socket_file.readline().rstrip('\r\n') def handle_error(self, socket_file): return Error(socket_file.readline().rstrip('\r\n')) def handle_integer(self, socket_file): return int(socket_file.readline().rstrip('\r\n')) def handle_string(self, socket_file): # First read the length ($<length>\r\n). length = int(socket_file.readline().rstrip('\r\n')) if length == -1: return None # Special-case for NULLs. length += 2 # Include the trailing \r\n in count. return socket_file.read(length)[:-2] def handle_array(self, socket_file): num_elements = int(socket_file.readline().rstrip('\r\n')) return [self.handle_request(socket_file) for _ in range(num_elements)] def handle_dict(self, socket_file): num_items = int(socket_file.readline().rstrip('\r\n')) elements = [self.handle_request(socket_file) for _ in range(num_items * 2)] return dict(zip(elements[::2], elements[1::2]))
For the serialization side of the protocol, we'll do the opposite of the above: turn Python objects into their serialized counterparts!
class ProtocolHandler(object): # ... above methods omitted ... def write_response(self, socket_file, data): buf = BytesIO() self._write(buf, data) buf.seek(0) socket_file.write(buf.getvalue()) socket_file.flush() def _write(self, buf, data): if isinstance(data, str): data = data.encode('utf-8') if isinstance(data, bytes): buf.write('$%s\r\n%s\r\n' % (len(data), data)) elif isinstance(data, int): buf.write(':%s\r\n' % data) elif isinstance(data, Error): buf.write('-%s\r\n' % error.message) elif isinstance(data, (list, tuple)): buf.write('*%s\r\n' % len(data)) for item in data: self._write(buf, item) elif isinstance(data, dict): buf.write('%%%s\r\n' % len(data)) for key in data: self._write(buf, key) self._write(buf, data[key]) elif data is None: buf.write('$-1\r\n') else: raise CommandError('unrecognized type: %s' % type(data))
An additional benefit of keeping the protocol handling in its own class is that we can re-use the
handle_request and
write_response methods to build a client library.
Implementing Commands
The
Server class we mocked up now needs to have it's
get_response() method implemented. Commands will be assumed to be sent by the client as either simple strings or an array of command arguments, so the
data parameter passed to
get_response() will either be bytes or a list. To simplify handling, if
data is a simple string, we'll convert it to a list by splitting on whitespace.
The first argument will be the command name, with any additional arguments belonging to the specified command. As we did with the mapping of the first byte to the handlers in the
ProtocolHandler, let's create a mapping of command to callback in the
Server:
class Server(object): def __init__(self, host='127.0.0.1', port=31337, max_clients=64): self._pool = Pool(max_clients) self._server = StreamServer( (host, port), self.connection_handler, spawn=self._pool) self._protocol = ProtocolHandler() self._kv = {} self._commands = self.get_commands() def get_commands(self): return { 'GET': self.get, 'SET': self.set, 'DELETE': self.delete, 'FLUSH': self.flush, 'MGET': self.mget, 'MSET': self.mset} def get_response(self, data): if not isinstance(data, list): try: data = data.split() except: raise CommandError('Request must be list or simple string.') if not data: raise CommandError('Missing command') command = data[0].upper() if command not in self._commands: raise CommandError('Unrecognized command: %s' % command) return self._commands[command](*data[1:])
Our server is almost finished! We just need to implement the six command methods defined in the
get_commands() method:
class Server(object): def get(self, key): return self._kv.get(key) def set(self, key, value): self._kv[key] = value return 1 def delete(self, key): if key in self._kv: del self._kv[key] return 1 return 0 def flush(self): kvlen = len(self._kv) self._kv.clear() return kvlen def mget(self, *keys): return [self._kv.get(key) for key in keys] def mset(self, *items): data = zip(items[::2], items[1::2]) for key, value in data: self._kv[key] = value return len(data)
That's it! Our server is now ready to start processing requests. In the next section we'll implement a client to interact with the server.
Client
To interact with the server, let's re-use the
ProtocolHandler class to implement a simple client. The client will connect to the server and send commands encoded as lists. We'll re-use both the
write_response() and the
handle_request() logic for encoding requests and processing server responses respectively.
class Client(object): def __init__(self, host='127.0.0.1', port=31337): self._protocol = ProtocolHandler() self._socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM) self._socket.connect((host, port)) self._fh = self._socket.makefile('rwb') def execute(self, *args): self._protocol.write_response(self._fh, args) resp = self._protocol.handle_request(self._fh) if isinstance(resp, Error): raise CommandError(resp.message) return resp
With the
execute() method, we can pass an arbitrary list of parameters which will be encoded as an array and sent to the server. The response from the server is parsed and returned as a Python object. For convenience, we can write client methods for the individual commands:
class Client(object): # ... def get(self, key): return self.execute('GET', key) def set(self, key, value): return self.execute('SET', key, value) def delete(self, key): return self.execute('DELETE', key) def flush(self): return self.execute('FLUSH') def mget(self, *keys): return self.execute('MGET', *keys) def mset(self, *items): return self.execute('MSET', *items)
To test out our client, let's configure our Python script to start up a server when executed directly from the command-line:
# Add this to bottom of module: if __name__ == '__main__': from gevent import monkey; monkey.patch_all() Server().run()
Testing the Server
To test the server, just execute the server's Python module from the command line. In another terminal, open up a Python interpreter and import the
Client class from the server's module. Instantiating the client will open a connection and you can start running commands!
>>> from server_ex import Client >>> client = Client() >>> client.mset('k1', 'v1', 'k2', ['v2-0', 1, 'v2-2'], 'k3', 'v3') 3 >>> client.get('k2') ['v2-0', 1, 'v2-2'] >>> client.mget('k3', 'k1') ['v3', 'v1'] >>> client.delete('k1') 1 >>> client.get('k1') >>> client.delete('k1') 0 >>> client.set('kx', {'vx': {'vy': 0, 'vz': [1, 2, 3]}}) 1 >>> client.get('kx') {'vx': {'vy': 0, 'vz': [1, 2, 3]}} >>> client.flush() 2
The code presented in this post is absolutely for demonstration purposes only. I hope you enjoyed reading about this project as much as I enjoyed writing about it. You can find a complete copy of the code here.
To extend the project, you might consider:
- Add more commands!
- Use the protocol handler to implement an append-only command log
- More robust error handling
- Allow client to close connection and re-connect
- Logging
- Re-write to use the standard library's
SocketServerand
ThreadingMixin
Commenting has been closed, but please feel free to contact me | https://pythondigest.ru/view/31992/ | CC-MAIN-2018-43 | refinedweb | 2,054 | 50.84 |
Build your own multi-user photo album app with React, GraphQL, and AWS Amplify — Part 1 of 3
Bootstrap our app, added authentication, and integrated a GraphQL API
Part 1 | Part 2 | Part 3
This is the first post in a three-part series that shows you how to build a scalable and highly available serverless web app on AWS that lets users upload photos to albums and share those albums privately with others.
An app like this requires a few moving parts:
- Allowing user sign up and authentication, so we know who owns which photo albums, and so users can invite other users to their CloudFront, to cache and quickly serve our app’s static assets to users from locations around the world
If any or all of these services are new to you, don’t worry. These series will cover everything you need to know to get started using these services. And there’s no better way to learn than to build, so let’s get started!
What we’ll cover in this post
In this first post in the series, we’ll lay the groundwork for our application, and get ourselves to a place where we can create and view photo album records once a user is logged into the app. Here’s a list of the steps we’ll cover below:
- Bootstrapping our web app with React
- Adding user authentication
- Deploying a GraphQL API to manage photo album data
- Connecting our web app to the API, letting users create and list albums, with real-time updates
Pre-requisites
Before we begin, there are a few things you’ll want to make sure you have set up so you can follow along.
An AWS account - If you don’t have one yet, you’ll need an account with AWS. Creating an AWS account is free and gives you immediate access to the AWS Free Tier. For more information, or to sign up for an account, see
Node.js installed - We’ll be using some JavaScript tools and packages which require Node.js (and NPM, which comes with Node.js) to be installed. You can download an installer or find instructions for installing Node.js at
The AWS Amplify CLI installed and configured - The AWS Amplify CLI helps you quickly configure serverless backends and includes support for authentication, analytics, functions, REST/GraphQL APIs, and more. Follow the installation instructions at. Windows users please note: The AWS Amplify CLI is currently only supported under the Windows Subsystem for Linux.
Bootstrapping our app
We this tool at. Running the tool uses a command installed alongside Node.js called npx which makes it easy to run a binary included inside an NPM package.
From your command line, navigate to a directory where you’d like to create a new directory for this project.
Run npx create-react-app photo-albums (you can use any name you want if you don’t like ‘photo-albums’, but I’ll assume you call it ‘photo-albums’ for these instructions).
That command will have created a new photo-albums directory with a bare bones React web app and a helpful script to run a local development web server. Before we start writing our UI, we’ll also include Semantic UI components for React to give us components that will help make our interface look a bit nicer.
Run npm install --save semantic-ui-react and integrate the default stylesheet by editing public/index.html and adding this stylesheet link to the <head> section: <link rel="stylesheet" href="//cdnjs.cloudflare.com/ajax/libs/semantic-ui/2.3.3/semantic.min.css"></link>
Now let’s start our development server so we can make changes and see them refreshed live in the browser.
From within the photo-albums directory, run npm start. In a moment, you should see a browser window open to with a React logo and some other content.
Next, we’ll want to start with a clean slate, so let’s edit src/App.js and change it to display a simple ‘Hello World’ message. Replace the existing content of the file with the following:
// src/App.js
import React, { Component } from ‘react’; so we’ll understand everything we add.
Adding user authentication with four lines of code
Now.
On your command line, make sure you’re inside the photo-albums directory and:
- Run amplify init
- Select your preferred editor
- Choose JavaScript and React when prompted
- Accept the default values when prompted for paths and build commands
This will create a new local configuration for us which we can use to set up an Amazon Cognito User Pool to act as the backend for letting users sign up and sign in (I’ll explain more about Amazon Cognito and what a User Pool is two paragraphs down). If you want to read more about this step, take a look at the ‘Installation and Configuration’ steps from the AWS Amplify Authentication guide.
When enabling user sign-in with the Amplify CLI, the default setup will require users to sign up with complex passwords and to confirm ownership of their email address before signing in. These are great settings for a production app and, while it’s completely possible to customize this configuration, let’s stick with these default settings.
Run amplify add auth to add authentication to the app. Select ‘Yes’ when asked if you’d like to use the default authentication and security configuration. Then, as the output text indicates, run amplify push..
Now run npm install --save aws-amplify aws-amplify-react and then we can use them in our app. Let’s update src/App.js:
// src/App.js
// 1. NEW: Add some new imports
import Amplify from 'aws-amplify';
import aws_exports from './aws-exports';
import { withAuthenticator } from 'aws-amplify-react';
Amplify.configure(aws_exports);
// 2. EDIT: Wrap our exported App component with WithAuthenticator
export default withAuthenticator(App, {includeGreetings: true});
Looking back at our browser (after re-starting our server), we’ve now got a simple sign-up / sign-in form. Try signing up in the app by providing a username, password, and an email address you can use if you need to reset your password. You’ll be taken to a screen asking you to confirm a code. This is because Amazon Cognito wants to verify a user’s email address before it lets them sign in. Go check your email and you should see a confirmation code message. Copy and paste that.
Creating a serverless GraphQL API backend
Now, I suggest you take a few minutes and check out before continuing, or use the site to refer back to when you have questions as you read along.
AWS AppSync also makes adding real-time and offline capabilities to your apps pretty easy, and we’ll add a real-time subscription at the end of this post, but let’s start with the basics: creating an AWS AppSync API, adding a simple GraphQL schema, and connecting it with DynamoDB for storing our data in a NoSQL database. Again, the Amplify CLI streamlines this process quite a bit.
From your terminal,
? Do you want to edit the schema now? Yes
Please edit the file in your editor: photo-albums/amplify/backend/api/photo-albums/schema.graphql
At this point, your code editor should show you a sample one-to-many GraphQL schema, annotated with some @model and @connection bits. The Amplify CLI will transform this schema into a set of data sources and resolvers to run on AWS AppSync. You can learn more about Amplify’s GraphQL transforms here.
Below is a schema that will suit our needs for storing and querying Albums and Photos. Copy and paste this into your editor, replacing the example schema content, and remembering to save the file:
# amplify/backend/api/photo-albums/schema.graphql
type Album @model {
id: ID!
name: String!
owner: String!
photos: [Photo] @connection(name: "AlbumPhotos")
}
type Photo @model {
id: ID!
album: Album @connection(name: "AlbumPhotos")
bucket: String!
fullsize: PhotoS3Info!
thumbnail: PhotoS3Info!
}
type PhotoS3Info {
key: String!
width: Int!
height: Int!
}
Now return to your command line and press Enter once to continue.
Finally, run amplify push and confirm you’d like to continue with the updates. Then wait a few moments while the new resources are provisioned in the cloud.
Open the AWS AppSync web console, find the API matching the name you entered above, and click it. Now we can start poking around with the API.
The really cool thing is,.
Click over to the ‘Queries’ section in the sidebar..
- In the Queries section of the AppSync web console, notice there is a ‘Login with User Pools’ button at the top of the query editor. Click the button.
- In the ‘ClientId’ field, we’ll need to paste in a value from our Cognito User Pool configuration. One place to find the right value is to look in src/aws-exports.js in our React web app. This file was generated by the Amplify CLI and contains lots of configuration values for our app. The one we want to copy and use for this field is aws_user_pools_web_client_id.
- In the ‘Username’ and ‘Password’ fields, use the values you specified when you created a username in our React app above.
- Click ‘Login’. You should now be able to try out the following mutations and queries.
Add a new album by copy/pasting the following and running the query:
mutation {
createAlbum(input:{name:"First Album", owner:"Someone"}) {
id
name
}
}
Add another album by editing and re-running your createAlbum mutation with another album name:
mutation {
createAlbum(input:{name:"Second Album", owner:"Someone"}) {
id
name
}
}
List our albums by running this query:
query {
listAlbums {
items {
id
name
owner
}
}
}
As you can see, we’re able to read and write data through GraphQL queries and mutations. But, things aren’t exactly as we’d like. In the code that was generated, we’re required to specify an owner string for each Album we create. Instead, it would be great if the system would automatically set the owner field to the username of whichever user sent the createAlbum mutation. It’s not hard to make this change, so let’s get it out of the way now before we connect our web app to our AWS AppSync API endpoint.
Customizing our first GraphQL resolver in AWS AppSync
AWS AppSync supports multiple ways of resolving data for our fields. For fetching data from DynamoDB and Amazon ElasticSearch, we can resolve our requests and responses using the Apache Velocity Template Language, or VTL. We can also resolve fields using an AWS Lambda function, but since our current needs are so simple, we’ll stick to resolving from DynamoDB using VTL for now.
The resolvers that AWS AppSync generated for us are just a way to translate our GraphQL queries and mutations into DynamoDB operations.
At first glance, interacting with DynamoDB via these templates can seem a bit weird, but there’s only a few concepts you need to get in order to work with them effectively. AWS AppSync resolver mapping templates are evaluated by a VTL engine and end up as a JSON string that’s used to make API calls to DynamoDB. Each DynamoDB operation has a corresponding set of values it expects in the JSON, based on whatever that API operation is about. See the Resolver Mapping Template Reference for DynamoDB for more information. VTL templates can use variables and logic to control the output that gets rendered. There are lots of examples in the Resolver Mapping Template Programming Guide. Any arguments passed in from our GraphQL queries or mutations are available in the $ctx.args variable, and user identity information is available in the $ctx.identity variable. You can read more about these, and other variables, in the Resolver Mapping Template Context Reference.
Since we want to change how our createAlbum mutation interfaces with DynamoDB, let’s open that up resolver make make an edit. It would be great if we could edit a resolver file locally on our development machine, then do another push with the Amplify CLI, but at this time, we can’t do this via the CLI yet, so we’ll have to make the change manually in the AWS AppSync web console, and take care not to run any further API pushes from the Amplify CLI, which would overwrite the changes we want to make. Here’s how to make the edit we want:
- Click ‘Schema’ in the AWS AppSync sidebar
- On the right side of the screen in the Resolvers column, scroll down to the Mutation section (or use the search box, typing ‘mutation’), find the createAlbum(…):Album listing and click the AlbumTable link to view and edit this resolver
- Near the lines at the top that look like $util.qr(…) add another similar line to inject the current user’s username in as another argument to the mutation’s input: $util.qr($context.args.input.put("owner", $context.identity.username))
- Click ‘Save Resolver’
Also, since we’re no longer passing a value for owner when we create an Album, we’ll update the Schema to only require a ‘name’ for albums.
Edit the Schema, replacing the CreateAlbumInput type with the following definition, and click ‘Save Schema’
input CreateAlbumInput {
name: String!
}
Finally, we can test out creating another Album in the Queries tool to see if our modified resolver works correctly:
mutation {
createAlbum(input: {name:"Album with username from context"}) {
id
name
owner
}
}
That mutation should have returned successfully. Notice the album’s owner field contains the username we logged in with.
If you’re curious to see more examples of how AWS AppSync translates GraphQL queries and mutations into commands for DynamoDB, take a few minutes to examine the other auto-generated resolvers as well. Just click on any of the Resolver links you see in the Data Types sidebar of the AWS AppSync Schema view.
Connecting our app to the GraphQL API
At this point, we have a web app that authenticates users and a secure GraphQL API endpoint that lets us create and read Album data. It’s time to connect the two together! The Apollo client is a popular choice for working with GraphQL endpoints, but there’s a bit of boilerplate code involved when using Apollo, so let’s start with something simpler..
We’ll create some new components to handle data fetching and rendering of our Albums. In a real app, we’d create separate files for each of our components, but here we’ll just keep everything together so we can see all of the code in one place.
Make the following changes to src/App.js:
// src/App.js
// 1. NEW: Add some new imports
import { Connect } from 'aws-amplify-react';
import { graphqlOperation } from 'aws-amplify';
import { Grid, Header, Input, List, Segment } from 'semantic-ui-react';
// 2. NEW: Create a function we can use to
// sort an array of objects by a common property
function makeComparator(key, order='asc') {
return (a, b) => {
if(!a.hasOwnProperty(key) || !b.hasOwnProperty(key)) return 0;
const aVal = (typeof a[key] === 'string') ? a[key].toUpperCase() : a[key];
const bVal = (typeof b[key] === 'string') ? b[key].toUpperCase() : b[key];
let comparison = 0;
if (aVal > bVal) comparison = 1;
if (aVal < bVal) comparison = -1;
return order === 'desc' ? (comparison * -1) : comparison
};
}
// 3. NEW: Add an AlbumsList component for rendering
// a sorted list of album names
class AlbumsList extends React.Component {
albumItems() {
return this.props.albums.sort(makeComparator('name')).map(album =>
<li key={album.id}>
{album.name}
</li>);
}
render() {
return (
<Segment>
<Header as='h3'>My Albums</Header>
<List divided relaxed>
{this.albumItems()}
</List>
</Segment>
);
}
}
// 4. NEW: Add a new string to query all albums
const ListAlbums = `query ListAlbums {
listAlbums(limit: 9999) {
items {
id
name
}
}
}`;
// 5. NEW: Add an AlbumsListLoader component that will use the
// Connect component from Amplify to provide data to AlbumsList
class AlbumsListLoader extends React.Component {
render() {
return (
<Connect query={graphqlOperation(ListAlbums)}>
{({ data, loading, errors }) => {
if (loading) { return <div>Loading...</div>; }
if (!data.listAlbums) return;
return <AlbumsList albums={data.listAlbums.items} />;
}}
</Connect>
);
}
}
// 6. EDIT: Change the App component to look nicer and
// use the AlbumsListLoader component
class App extends Component {
render() {
return (
<Grid padded>
<Grid.Column>
<AlbumsListLoader />
</Grid.Column>
</Grid>
);
}
}
If you check back in the browser after making these changes, you should see a list of the albums you’ve created so far.
The real, dumping an error message to help us debug, or passing the successfully fetched data to our AlbumsList component.
The listAlbums query we’re above using passes in a very high limit argument. This is because I’ve decided to load all of the albums in one request and sort the albums alphabetically on the client-side (instead of dealing with paginated DynamoDB responses). This keeps the AlbumsList code pretty simple, so I think it’s worth the trade off in terms of performance or network cost.
That takes care of fetching and rendering our list of Albums. Now let’s make a component to add new Albums. Make the following changes to src/App.js:
// src/App.js
// 1. EDIT: Update our import for aws-amplify to
// include API and graphqlOperation
// and remove the other line importing graphqlOperation
import Amplify, { API, graphqlOperation } from 'aws-amplify';
// 2. NEW: Create a NewAlbum component to give us
// a way of saving new albums
class NewAlbum extends Component {
constructor(props) {
super(props);
this.state = {
albumName: ''
};
}
handleChange = (event) => {
let change = {};
change[event.target.name] = event.target.value;
this.setState(change);
}
handleSubmit = async (event) => {
event.preventDefault();
const NewAlbum = `mutation NewAlbum($name: String!) {
createAlbum(input: {name: $name}) {
id
name
}
}`;
const result = await API.graphql(graphqlOperation(NewAlbum, { name: this.state.albumName }));
console.info(`Created album with id ${result.data.createAlbum.id}`);
}
render() {
return (
<Segment>
<Header as='h3'>Add a new album</Header>
<Input
type='text'
placeholder='New Album Name'
icon='plus'
iconPosition='left'
action={{ content: 'Create', onClick: this.handleSubmit }}
name='albumName'
value={this.state.albumName}
onChange={this.handleChange}
/>
</Segment>
)
}
}
// 3. EDIT: Add NewAlbum to our App component's render
class App extends Component {
render() {
return (
<Grid padded>
<Grid.Column>
<NewAlbum />
<AlbumsListLoader />
</Grid.Column>
</Grid>
);
}
}
Check the browser again and you should be able to create new albums with a form.
Finally, we can have our list of albums automatically refresh by taking advantage of AppSync’s real-time data subscriptions which trigger when mutations are made (by anyone). The easiest way to do this is to define a subscription in our schema so AWS AppSync will let us subscribe to createAlbum mutations, and then add the subscription to our Connect component that provides data for the AlbumList.
Our GraphQL schema already contains a Subscription type with a bunch of subscriptions that were auto-generated back when we had AWS AppSync create the resources (DynamoDB table and AWS AppSync resolvers) for our Album type. One of these is already perfect for our needs: the onCreateAlbum subscription.
Add subscription and onSubscriptionMsg properties to our Connect component.
Here’s an updated version of our AlbumsListLoader component (and a new string containing our subscription) with these changes incorporated. Make the following changes to src/App.js:
// src/App.js
// 1. NEW: Add a string to store a subcription query for new albums
const SubscribeToNewAlbums = `
subscription OnCreateAlbum {
onCreateAlbum {
id
name
}
}
`;
// 2. EDIT: Update AlbumsListLoader to work with subscriptions
class AlbumsListLoader extends React.Component {
// 2a. NEW: add a onNewAlbum() function
// for handling subscription events)}
// 2b. NEW: Listen to our
subscription={graphqlOperation(SubscribeToNewAlbums)}
// 2c. NEW: Handle new subscription messages
onSubscriptionMsg={this.onNewAlbum}
>
{({ data, loading, errors }) => {
if (loading) { return <div>Loading...</div>; }
if (errors.length > 0) { return <div>{JSON.stringify(errors)}</div>; }
if (!data.listAlbums) return;
return <AlbumsList albums={data.listAlbums.items} />;
}}
</Connect>
);
}
}
With these changes in place, if you try creating a new album, you should see it appear at the bottom of our list of album names. What’s even cooler, is that if you add a new album in another browser window, or in the AWS AppSync Query console, you’ll see it appear automatically in all of your open browser windows. That’s the real-time power of subscriptions at work!
There’s more to come
Whew! We’ve come a long way since the beginning of this post. We created a new React app with user sign up and authentication, created a GraphQL API to handle persisting and querying data, and we wired these two parts together to create a real-time data-driven UI. Give yourself a pat on the back!
In future posts, we’ll continue on and add photo album uploads, thumbnail generation, querying for all the photos in an album, paginating photo results, fine-grained album security settings, and deploying our app to a CDN.
If you’d like to be notified when new posts come out, please follow me on Twitter: Gabe Hollombe. That’s also the best way to reach me if you have any questions or feedback about this post.
Part 1 | Part 2 | Part 3
This is the first post in a three-part series that shows you how to build a scalable and highly available serverless web app on AWS that lets users upload photos to albums and share those albums privately with others.
Bootstrapping what we’ve built so far
If you’d like to just check out a repo and launch the app we’ve built so far, check out this repo on GitHub and use the blog-post-part-one tag, linked here:. Follow the steps in the README to configure and launch the app.
| https://laptrinhx.com/build-your-own-multi-user-photo-album-app-with-react-graphql-and-aws-amplify-747882698/ | CC-MAIN-2020-50 | refinedweb | 3,588 | 53.41 |
Kubernetes and Helm¶
It is easy to launch a Dask cluster and a Jupyter notebook server on cloud resources using Kubernetes and Helm.
This is particularly useful when you want to deploy a fresh Python environment on Cloud services like Amazon Web Services, Google Compute Engine, or Microsoft Azure.
If you already have Python environments running in a pre-existing Kubernetes cluster, then you may prefer the Kubernetes native documentation, which is a bit lighter weight.
Launch Kubernetes Cluster¶
This document assumes that you have a Kubernetes cluster and Helm installed.
If this is not the case, then you might consider setting up a Kubernetes cluster on one of the common cloud providers like Google, Amazon, or Microsoft. We recommend the first part of the documentation in the guide Zero to JupyterHub that focuses on Kubernetes and Helm (you do not need to follow all of these instructions). Also, JupyterHub is not necessary to deploy Dask:
Alternatively, you may want to experiment with Kubernetes locally using Minikube.
Helm Install Dask¶
Dask maintains a Helm chart in the default stable channel at . This should be added to your helm installation by default. You can update the known channels to make sure you have up-to-date charts as follows:
helm repo update
Now, you can launch Dask on your Kubernetes cluster using the Dask Helm chart:
helm install stable/dask
This deploys a
dask-scheduler, several
dask-worker processes, and
also an optional Jupyter server.
Verify Deployment¶
This might take a minute to deploy. You can check its status with
kubectl:
kubectl get pods kubectl get services $ kubectl get pods NAME READY STATUS RESTARTS AGE bald-eel-jupyter-924045334-twtxd 0/1 ContainerCreating 0 1m bald-eel-scheduler-3074430035-cn1dt 1/1 Running 0 1m bald-eel-worker-3032746726-202jt 1/1 Running 0 1m bald-eel-worker-3032746726-b8nqq 1/1 Running 0 1m bald-eel-worker-3032746726-d0chx 0/1 ContainerCreating 0 1m $
You can use the addresses under
EXTERNAL-IP to connect to your now-running
Jupyter and Dask systems.
Notice the name
bald-eel. This is the name that Helm has given to your
particular deployment of Dask. You could, for example, have multiple
Dask-and-Jupyter clusters running at once, and each would be given a different
name. Note that you will need to use this name to refer to your deployment in the future.
Additionally, you can list all active helm deployments with:
helm list NAME REVISION UPDATED STATUS CHART NAMESPACE bald-eel 1 Wed Dec 6 11:19:54 2017 DEPLOYED dask-0.1.0 default
Connect to Dask and Jupyter¶
When we ran
kubectl get services, we saw some externally visible IPs:
[email protected]:~$
We can navigate to these services from any web browser. Here, one is the Dask diagnostic
dashboard, and the other is the Jupyter server. You can log into the Jupyter
notebook server with the password,
dask.
You can create a notebook and create a Dask client from there. The
DASK_SCHEDULER_ADDRESS environment variable has been populated with the
address of the Dask scheduler. This is available in Python in the
config dictionary.
>>> from dask.distributed import Client, config >>> config['scheduler-address'] 'bald-eel-scheduler:8786'
Although you don’t need to use this address, the Dask client will find this variable automatically.
from dask.distributed import Client, config client = Client()
Configure Environment¶
By default, the Helm deployment launches three workers using two cores each and a standard conda environment. We can customize this environment by creating a small yaml file that implements a subset of the values in the dask helm chart values.yaml file.
For example, we can increase the number of workers, and include extra conda and pip packages to install on the both the workers and Jupyter server (these two environments should be matched).
# config.yaml worker: replicas: 8 resources: limits: cpu: 2 memory: 7.5G requests: cpu: 2 memory: 7.5G env: - name: EXTRA_CONDA_PACKAGES value: numba xarray -c conda-forge - name: EXTRA_PIP_PACKAGES value: s3fs dask-ml --upgrade # We want to keep the same packages on the worker and jupyter environments jupyter: enabled: true env: - name: EXTRA_CONDA_PACKAGES value: numba xarray matplotlib -c conda-forge - name: EXTRA_PIP_PACKAGES value: s3fs dask-ml --upgrade
This config file overrides the configuration for the number and size of workers and the conda and pip packages installed on the worker and Jupyter containers. In general, we will want to make sure that these two software environments match.
Update your deployment to use this configuration file. Note that you will not use helm install for this stage: that would create a new deployment on the same Kubernetes cluster. Instead, you will upgrade your existing deployment by using the current name:
helm upgrade bald-eel stable/dask -f config.yaml
This will update those containers that need to be updated. It may take a minute or so.
As a reminder, you can list the names of deployments you have using
helm
list
Check status and logs¶
For standard issues, you should be able to see the worker status and logs using the
Dask dashboard (in particular, you can see the worker links from the
info/ page).
However, if your workers aren’t starting, you can check the status of pods and
their logs with the following commands:
kubectl get pods kubectl logs <PODNAME>
[email protected]:~$ kubectl get pods NAME READY STATUS RESTARTS AGE bald-eel-jupyter-3805078281-n1qk2 1/1 Running 0 18m bald-eel-scheduler-3074430035-cn1dt 1/1 Running 0 58m bald-eel-worker-1931881914-1q09p 1/1 Running 0 18m bald-eel-worker-1931881914-856mm 1/1 Running 0 18m bald-eel-worker-1931881914-9lgzb 1/1 Running 0 18m bald-eel-worker-1931881914-bdn2c 1/1 Running 0 16m bald-eel-worker-1931881914-jq70m 1/1 Running 0 17m bald-eel-worker-1931881914-qsgj7 1/1 Running 0 18m bald-eel-worker-1931881914-s2phd 1/1 Running 0 17m bald-eel-worker-1931881914-srmmg 1/1 Running 0 17m [email protected]:~$ kubectl logs bald-eel-worker-1931881914-856mm EXTRA_CONDA_PACKAGES environment variable found. Installing. Fetching package metadata ........... Solving package specifications: . Package plan for installation in environment /opt/conda/envs/dask: The following NEW packages will be INSTALLED: fasteners: 0.14.1-py36_2 conda-forge monotonic: 1.3-py36_0 conda-forge zarr: 2.1.4-py36_0 conda-forge Proceed ([y]/n)? monotonic-1.3- 100% |###############################| Time: 0:00:00 11.16 MB/s fasteners-0.14 100% |###############################| Time: 0:00:00 576.56 kB/s ...
Delete a Helm deployment¶
You can always delete a helm deployment using its name:
helm delete bald-eel --purge
Note that this does not destroy any clusters that you may have allocated on a Cloud service (you will need to delete those explicitly). | https://docs.dask.org/en/latest/setup/kubernetes-helm.html | CC-MAIN-2019-09 | refinedweb | 1,125 | 59.23 |
How to Run Swift UI Tests With a Mock API Server
How to Run Swift UI Tests With a Mock API Server
A UI test must be well written to be properly efficient. Learn to use the mock API server WireMock to run Swift UI tests on your mobile applications.
Join the DZone community and get the full member experience.Join For Free
UI testing is a great tool to test the flow of your application. It can catch regression bugs and unexpected UI behaviors. Like unit testing, a UI test must be written well and efficiently, since you may run it several times a day to refactor/add a new feature in your application.
In this article, I explain how to use a mock API server for your UI tests, step by step, with a sample project.
Sample Project
The project used in this article is a plain application which shows a table of usernames:
I used the architectural pattern MVVM-C. If you are not familiar with it, you can check out my article about architectural patterns. Of course, you can use the pattern which you prefer, it’s a personal choice.
You can find the sample project here (it has both a mock server and iOS project). I won’t explain how to create the
UITableView, since it’s beyond the goal of this article. If you don’t know how to do it, you can check in the sample project.
Project Configuration
We must configure our project to use the mock API server for UI testing and a remote API for
Debug and
Release. To achieve it, we can create a new build configuration for the tests, and read the server URL from a JSON file for each build configuration.
App Transport Security Exception
Since iOS 9, your app must support App Transport Security (ATS). This feature forces your app to use just secure connections. It means that the app must communicate with remote servers using the protocol HTTPS instead of HTTP.
Since we are running the mock server locally, we don’t need a secure connection for it. To disable ATS for
localhost, we must edit the file
Info.plist of the main target.
You can open the plist file with a text editor and append:
<key>NSAppTransportSecurity</key> <dict> <key>NSExceptionDomains</key> <dict> <key>localhost</key> <dict> <key>NSTemporaryExceptionAllowsInsecureHTTPLoads</key> <true/> </dict> </dict> </dict>
Before the last two lines:
</dict> </plist>
The result in Xcode will be:
If you have problems with the Info.plist, I suggest you copy/paste the values from the sample project.
Build Configurations
We have to load the config file programmatically with Swift, therefore we need a way to understand when our application has been run by UI tests. There are several ways to achieve it. In this article, I use
Active Compilation Conditions. It’s a feature added in Xcode 8 and allows you to add conditional compilation flags to the Swift compiler. Therefore, if we add the flag
HELLO_WORD then we can check it programmatically:
func print() { #if HELLO_WORD print("Flag set") #else print("Flag not set") #endif }
You’ll see a realistic usage later.
Let’s start creating a new build configuration:
Then, we must edit the scheme to use the new build configuration for the tests:
Remember to set the scheme as shared. By default, the scheme is not shared, therefore the changes you make remain locally in your machine. By setting shared, you allow the other users of the project to use this scheme.
As the last step, we must set
Active Compilation Conditions:
- Open
Build Settingsof your project.
- Search
Active Compilation Conditions.
- Add the following flags to the build configurations (
DEBUG,
RELEASE,
TESTS):
You don’t need to do this for the build settings of the targets, since they will inherit the settings of the project.
Thanks to these flags, everywhere in your Swift code, you can check:
#if DEBUG // do something for build configuration Debug #elseif RELEASE // do something for build configuration Release #elseif TESTS // do something for build configuration Tests #endif
Config Files
For
Debug and
Release, we will use the free API at jsonplaceholder.typicode.com, whereas for our UI tests we want to use our mock server, which will be running in localhost with the port 1234. If you want to change the port number, pay attention to change it also in the mock server configuration.
We can start creating three config files:
Config-Debug.json and Config-Release.json:
{ "api_url": "" }
Config-Test.json:
{ "api_url": "" }
And add them in our Xcode project:
Then, we must read the right JSON file depending on our build configuration:
private func jsonFileContent() -> String? { var fileName: String? = nil #if DEBUG fileName = "Config-Debug" #elseif RELEASE fileName = "Config-Release" #elseif TESTS fileName = "Config-Test" #else fatalError("Config flag not found") #endif guard let file = Bundle.main.path(forResource: fileName, ofType: "json") else { return nil } return try! String(contentsOfFile: file) }
The method
jsonFileContent returns the content of the right JSON. How to parse the JSON is beyond the goal of this article. You can choose whatever way you prefer. In the sample project, I use the library ObjectMapper:
let model = Mapper<ConfigModel>().map(JSONString: jsonFileContent())
Once we have loaded the JSON values, we can use them to send an API request to get the users list:
func fetchUsers() -> [Users] { // apiUrl is the value of the config JSON file let url = apiUrl + "users" // send a API request to url and fetch the data }
In this way, when this code is running with the build configuration
TESTS,
apiUrl has the value, therefore
url is. Instead, with
Debug and
Release the value of
url is.
How to send the API request and fetch the data is beyond the goal of this article, but you can check the sample project if you want to do it using RxSwift. Otherwise, I suggest you have a look at AlamofireObjectMapper, a library to send API requests with Alamofire and parse the responses with ObjectMapper.
Mock Server Configuration
For the mock server, we use WireMock as standalone process, which will be running in localhost with the port number 1234.
WireMock is a HTTP mock server which allows you to map API requests, in this way you can simulate the API responses with WireMock instead of using the server in production.
Setup
The setup of WireMock is extremely easy; you can find the documentation here. If you want a script to run it easily, you can use the one I added in the sample project:
#!/bin/bash VERSION=2.6.0 PORT=1234 if ! [ -f "wiremock-standalone-$VERSION.jar" ]; then wget fi java -jar wiremock-standalone-$VERSION.jar --port $PORT
This script checks if you have the WireMock
jar file, and if not, downloads it. Finally, it runs the server with the port number 1234.
At the moment of writing,
2.6.0 is the latest version of WireMock.
Request Mappings
After the first run, WireMock creates two folders:
mappings and
__files.
Mappings
This folder contains all the WireMock responses. You can create a new response adding a new JSON file in this folder. The name of the file doesn’t matter, but I suggest you use meaningful names since you may have thousands of files.
For the sample project, we need to add the current mapping file:
{ "request": { "method": "GET", "url": "/users" }, "response": { "status": 200, "bodyFileName": "users_list.json" } }
The JSON values are very straightforward. We use: method (GET, POST, DELETE, ..), url, response status (200, 404, 500, …) and the response body.
bodyFileName is the file path where to read the response—the folder root of that file is
__files, we’ll see it later.
If you want to add the response body with a string instead of reading a file you can use the node
body:
{ "request": { "method": "GET", "url": "/users" }, "response": { "status": 200, "body": "Hello World!" } }
Here you can find other values to add to your mapping file like cache, header and so on.
__files
In this folder, you can place files to download or to use in your mapping files.
For our sample project, we need this folder do add our JSON response:
[ { "id": 1, "name": "Mock User #1", }, { "id": 2, "name": "Mock User #2", } ]
At the end of the setup, our WireMock folder structure will be like this:
You can ignore
index.json. It’s just a mapping for the server Homepage to show the message
WireMock is running!.
Every time you add/edit a file in these two folders, you must restart the server.
Advanced Usages
WireMock allows you to add request matching. With this feature, you can create different responses depending on the data you send in the request. You can find the documentation here.
For example, if your app sends in the request body a JSON with the user id, you can easily match it:
Body POST request for userId 1:
{ "userId": 1, "userName": "User #1" }
Response for userId 1:
{ "request": { "method": "POST", "url": "/user", "bodyPatterns": [{ "matchesJsonPath" : "$[?(@.userId == 1)]" }] }, "response": { "status": 200, "bodyFileName": "users_list.json" } }
Body POST request for userId 2:
{ "userId": 2, "userName": "User #2" }
Response for userId 2:
{ "request": { "method": "POST", "url": "/user", "bodyPatterns": [{ "matchesJsonPath" : "$[?(@.userId == 2)]" }] }, "response": { "status": 500, } }
In this way, you can log in with different users in your UI tests to test different API responses.
UI Testing
Thanks to our configuration, every time we run a UI test, we use the WireMock responses. Therefore, when we load the table of usernames, you get the data from. The final result is:
We can test our table with the following UI test:
func test_UsernamesTable_DataIsLoaded() { let areUsernamesLoaded = XCUIApplication().tables.cells.staticTexts["Mock User #1"].exists XCTAssertTrue(areUsernamesLoaded) }
In this article, I have shown just simple examples. WireMock is a powerful mock server, you should read the documentation thoroughly to learn all its features.
You may be wondering why I wanted to use JSON files to load the app configuration instead of using other tools. The power of this approach is that, instead of embedding in the Xcode project, you can store them in a remote server. Then, at the app startup you can send a request to this server to download the newest JSON config files. In this way, you can easily edit the config files in the server to change on the fly the configuration of your application. }} | https://dzone.com/articles/how-to-run-swift-ui-tests-with-a-mock-api-server-1 | CC-MAIN-2020-29 | refinedweb | 1,723 | 62.07 |
Building Applications with AppRun
AppRun is a lightweight library (3K) for developing applications using the Elm architecture, event publication and subscription, and components.
Application code using AppRun is clean, precious, declarative and without ceremony.
- The Todo app was written in 90 lines.
- The Hacker News app was written in 250 lines.
- The AppRun RealWorld Examples app was written in 1000 lines, where same functions require 2000 lines in React/Redux and Angular.
The Architecture
There are three separated parts in an AppRun application.
- Model — the state of your application
- View — a function to display the state as HTML
- Update — a set of event handlers to update the state
The 15 lines of code below is a simple counter application demonstrates the architecture using AppRun.
The Model
The model can be any data structure: a number, an array, or an object that reflects the state of the application. In the simple counter example, it is a number.
const model = 0;
The Update
The update contains event handlers that take existing state and create new state.
const update = {
'+1': (state) => state + 1,
'-1': (state) => state — 1
}
BTW, using the on decorator, the update functions look even better.
@on('+1') increase = state => state + 1;
@on('-1') decrease = state => state - 1;
The View
The view generates HTML based on the state. AppRun parses the HTML string into virtual dom. It then calculates the differences against the web page element and renders the changes.
const view = (state) => {
return `<div>
<h1>${state}</h1>
<button onclick=’app.run(“-1”)’>-1</button>
<button onclick=’app.run(“+1”)’>+1</button>
</div>`;
};
Trigger the Update
app.run function from AppRun is used to trigger the update.
app.run(‘+1’);
It can be used in HTML markup directly:
<button onclick=”app.run('-1')”>-1</button>
<button onclick=”app.run('+1')”>+1</button>
AppRun exposes a global object named app that is accessible by JavaScript and TypeScript directly to trigger updates.
Start the Application
app.start function from AppRun ties Model, View and Update together to an element and starts the application.
const element = document.getElementById(‘my-app’);
app.start(element, model, view, update);
The three functions from AppRun (app.run, app.start and app.start) are all you need to make an AppRun application.
Try it online: Simple Counter.
State Management
Internally AppRun manages application states. It triggers Update; passes the new state created by Update to View; renders the HTML/Virtual DOM created by View to the element. It also maintains a state history. The application can have the time travel / undo-redo feature by
- Make Update create immutable state
- Enable AppRun state history
Try it online: Multiple Counters
You can also check out more AppRun examples on glitch.com:
Virtual DOM
When applications get complex, we start to think performance and build system.
In the simple counter example above, the View creates HTML string out of the state. Although HTML string is easy to understand and useful for trying out ideas, it takes time to parse it into virtual dom at run time, which may cause performance issue. It also has some problems that have been documented by the Facebook React team: Why not template literals.
Using JSX, the JSX compiler compiles the JSX into functions at compile time. The functions creates virtual dom in run time directly without parsing HTML, which gives better performance. Therefore, we recommend using JSX in production. To compile and build production code, we recommend using TypeScript and webpack
AppRun CLI
Both TypeScript and webpack are advanced, feature-rich development tools that have many configuration options. But to find out the best configuration for TypeScript and webpack from scratch is difficult. Also, to configure TypeScript and webpack for every single project is a tedious task.
Like many other frameworks and libraries, AppRun comes with a CLI. AppRun CLI can initialize project, configure TypeScript and webpack, scaffold project folders and files, run tests, and run the development server.
Let’s install AppRun and use its CLI to initialize a project.
npx apprun -i
The apprun -i command installs apprun, webpack, webpack-dev-server and typescript. It also generates files: tsconfig.json, webpack.config.js, index.html and main.tsx.
After the command finishes execution, you can start the application.
npm start
Component
Another thing to consider when applications get complex is to divide and organize code into components.
A component in AppRun is a mini model-view-update architecture, which means inside a component, there are model, view and update. Let’s use AppRun CLI to generate a component.
apprun -c Counter
It generates a Counter component:
import app, {Component} from ‘apprun’;
export default class CounterComponent extends Component {
state = ‘Counter’;
view = (state) => {
return <div>
<h1>{state}</h1>
</div>
}
update = {
‘#Counter’: state => state,
}
}
To use the Counter component, create an instance of it and then mount the instance to an element.
import Counter from ‘./Counter’;
const element = document.getElementById(‘my-app’);
new Counter().mount(element);
Notice the update has a ‘#Counter’ function. It is a route.
Routing
The third thing to consider in complex applications is routing. AppRun has an unique way of handling routing. It detects the hash changes in URL and calls functions in update by matching the hash. E.g. when URL in the browser address bar becomes, The #Couter update function of the components will be executed.
Each component defines its route in an update function. Once the URL is changed to the route the component defined, the update function is triggered and executed. It can avoid a lot of code for registering and matching routes like in the other frameworks and libraries.
The AppRun demo application was built to have 8 components that are routed into one element.
Event PubSubs
At core, AppRun is an event publication and subscription system, which is also known as event emitter. It is a commonly used pattern in JavaScript programming.
AppRun has two important functions: app.run and app.on. app.run fires events. app.on handles events. E.g. :
Module A subscribes to an event print:
import app from ‘apprun’;
export () => app.on(‘print’, e => console.log(e));
Module B publishes the event print:
import app from ‘apprun’;
export () => app.run(‘print’, {});
Main Module:
import a from ‘./A’;
import b from ‘./B’;
a();
b();
Module A and Module B only have to know the event system, not other modules, so they are only dependent on the event system, not other modules. Therefore modules are decoupled. Thus makes it easier to modify, extend and swap modules.
Use AppRun connect web page events to components. Then it’s up to your application to continue executing.
document.body.addEventListener(‘click’, e => app.run(‘click’, e) );
element.addEventListener(‘click’, e => app.run(‘click’, e));
<input onclick=”e => app.run(‘click’, e)”>
The biggest benefit of such event system is decoupling. In traditional MVC architecture, the model, view and controller are coupled, which makes it difficult to test and maintain. In result, there are many architecture patterns have been developed in order to solve the coupling problem, such as Model-View-Presenter, Presentation Model and Model-View-ViewModel.
AppRun solved the coupling problem by using event publication and subscription. Model,View and Controller/Update don’t know each. They only need to know how to publish and subscribe events by calling app.run and app.on of AppRun.
Even handling routing is just to subscribe to an event.
Conclusion
AppRun itself is lightweight. It is about 3K gzipped. More important is that it gives you a great development experience and makes your applications great.
I have created following videos to demonstrate the concepts described in the post. Enjoy!
Please give it a try and send comments / pull requests. | https://medium.com/@yiyisun/building-applications-with-apprun-d103cd461bae | CC-MAIN-2018-47 | refinedweb | 1,279 | 59.4 |
»
Swing / AWT / SWT
Author
Drawing a selection box using swing
luk luka
Greenhorn
Joined: May 09, 2011
Posts: 1
posted
May 09, 2011 06:44:06
0
I have written an application with a panel and three buttons. I want to add selection this buttons using the mouse. I mean like we have in Windows on the Desktop. I press the left mouse button and with the movement of the mouse the area selection is growing.
Is there a specific interface in this or do I have it manually call the appropriate methods for event listeners and there draw transparent rectangle? Here is a picture:
So I have a problem when I paint rectangle using event mouse-dragged, button is repainting so user see blinking button. I want to this button don't disapear when I paint rectangle. I think that I need to use glassPane. This is my conception. I have a frame. In frame I add panel with button and I need another panel where I will paint transparent rectangle. I am thinking then my button will not be still repainting. What do you think about this conception. Or maybe someone have another idea. This is code:
import java.awt.AlphaComposite; import java.awt.Color; import java.awt.EventQueue; import java.awt.Graphics; import java.awt.Graphics2D; import java.awt.Point; import java.awt.event.MouseEvent; import java.awt.event.MouseListener; import java.awt.event.MouseMotionListener; import java.awt.geom.Rectangle2D; import java.util.Vector; import javax.swing.JButton; import javax.swing.JComponent; import javax.swing.JFrame; import javax.swing.JPanel; public class zaznacz extends JPanel implements MouseListener, MouseMotionListener { Rectangle newLine=null; public zaznacz() { addMouseListener(this); addMouseMotionListener(this); } static class Rectangle { int x1,y1,x2,y2; Rectangle(int _x1, int _y1,int _x2, int _y2){ x1=_x1;y1=_y1;x2=_x2;y2=_y2; } void paint(Graphics g) { g.drawRect(x1, y1, x2, y2); } } public void mouseClicked(MouseEvent e) { } @Override public void mouseEntered(MouseEvent arg0) { // TODO Auto-generated method stub } @Override public void mouseExited(MouseEvent arg0) { // TODO Auto-generated method stub } @Override public void mousePressed(MouseEvent e) { startPoint=e.getPoint();); } @Override public void mouseReleased(MouseEvent e) { repaint(); } Point startPoint; @Override public void mouseDragged(MouseEvent e) {); paintComponent(g2); } @Override public void mouseMoved(MouseEvent e) { } int rule = AlphaComposite.SRC_OVER; float alpha = 0.85F; public static void main(String[] args) { EventQueue.invokeLater(new Runnable() { public void run() { zaznacz rys = new zaznacz(); JFrame frame = new JFrame(); JButton Button = new JButton("1"); JPanel panel = new JPanel(); panel.add(Button); rys.add(panel); frame.setSize(400,300); frame.setVisible(true); frame.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE); panel.setOpaque(false); frame.add(rys); } }); } }
I know that code is no perfect but almost work. I have a little problem. When I press the mousebutton and dragging my button disapear.
I don't need advise like "your code is wrong". I know that and I want to somebody help me what I must correct. I know that I shouldn't use paintComponent() in mouseEvents but only that way I can paint transparent rectangle. Or maybe you can othet idea how I can draw transparent rectangle. I try and try and I think that i must change mouseDragged method. because when I delete code from this method and only draw rectangle over a button all is ok. But problem is when I need draw rectangle by dragging mouse. I should change paint but I don't have idea how. I need some ideas how I can solve my problem. Maybe I should use glass pane?
Darryl Burke
Bartender
Joined: May 03, 2008
Posts: 4522
5
I like...
posted
May 09, 2011 07:11:04
0
Hi luk luka and welcome to the Ranch!
You have totally ignored my advice given several hours ago in your cross post at
I see you have also cross posted this to SO
Please
BeForthrightWhenCrossPostingToOtherSites
!
edit One more cross post with the same advice
luck, db
There are no new questions, but there may be new answers.
I agree. Here's the link:
subject: Drawing a selection box using swing
Similar Threads
clickable line
how to highlight(change color) of the button and its neighbours when it is clicked
drawing arrows
How to make rectangle drawn to be vissible when a new rectangle is drawn
Using mouse clicks to draw on a JFrame, problem probably with coordinates
All times are in JavaRanch time: GMT-6 in summer, GMT-7 in winter
JForum
|
Paul Wheaton | http://www.coderanch.com/t/537222/GUI/java/Drawing-selection-box-swing | CC-MAIN-2014-10 | refinedweb | 738 | 59.19 |
Welcome to the Ars OpenForum.
struct S{ S(){} S(const S&){}};
S s(S());
S s(); // I want to invoke the default constructor here
S s;
S s(S);
S s((S()));
S s((0, S()));
fstream fin("filename");vector<int> v(istream_iterator<int>(fin), istream_iterator<int>());.
fstream fin("filename");istream_iterator<int> iiBegin(fin);istream_iterator<int> iiEnd;vector<int> v(iiBegin, iiEnd);
quote:Originally posted by Homer J Simpson:7.1 doesn't exist until it's released. Like I said, if you write that on a conforming compiler, you'll get an error and immediately realise your mistake. Remembering such things off by heart seems a waste of brain space.
struct S{};int main(){ S s(S());}
struct S{ int foo;};int main(){ S s(S()); s.foo;}
struct S{ S() {} S(const S &s) {}};typedef S func1();typedef S func2(S func());S s(func1 f1){ return (*f1)();}S retS(){ return S();}S f(func2 f2){ return (*f2)(retS);}int main(){ S s(S()); f(s); return 0;}
void f(const S &s){}
test.cpp: In function `int main()':test.cpp:33: could not convert `s(S (*)())' to `const S&'test.cpp:26: in passing argument 1 of `void f(const S&)'
quote:Furthermore, anyone who's done any programming in C++ knows the canonical form of the default constructor is "S s;" and would never write "S s(S());". And I doubt anyone would declare a function like that (they'd use typedefs along the lines of what I've done).
quote:It's a language lawyer question, that will probably only immediately ring a bell with people who happen to use function pointers extensively on a day to day basis, or those who read the spec for kicks.
quote:Originally posted by Homer J Simpson:But you've gone and writen "TypeName(TypeName())" which is not idiomatic. When I see code that is written in an abnormal way I immediately get suspicious. Either the author is trying to do something "smart" or is an idiot, sometimes both.
quote:Yes, it may appear in other contexts, but you've chosen a poor example. In fact you've deliberately chosen your example to be misleading (as evidenced by including a copy constructor).
quote:That seems to me to be a "hey aren't I cool, I know what this does" example, not a real test of anyone's | http://arstechnica.com/civis/viewtopic.php?f=20&t=767929 | CC-MAIN-2014-41 | refinedweb | 400 | 62.38 |
[
]
Robert Metzger commented on LEGAL-198:
--------------------------------------
I have another question on this issue, regarding the binary redistribution of code licensed
under the AmazonSL.
Similar to Apache Spark, Apache Flink recently added a Kinesis connector as an optional module
into our repository.
If users want to use the Kinesis connector, they have to build it from source, using a special
flag on the maven command.
Flink soon plans to release a new version, containing the Kinesis connector. I wonder if we
can distribute the maven artifacts (binaries) of the optional module to maven central, as
part of the release.
Since Kinesis is using Google protobuf V2.6.1, but Flink depends on protobuf 2.5.0, we need
to relocate ("shade") the protobuf version used by Kinesis into a different namespace. This
effectively means that one maven artifact the Flink project deploys to maven central contains
binaries licensed under the AmazonSL.
>From above discussions, I got that in Spark's case its okay to link to AmazonSL licensed
software, because the linking happens on the user side. In the case I've described, we would
redistribute modified binaries of a restricted use license as part of an Apache release.
I guess the likelihood that we can not release the artifact like this is close to 99%, but
I still wanted to ask, because this issue is causing me a lot of trouble, and this solution
would be the easiest one for me.
> | http://mail-archives.apache.org/mod_mbox/www-legal-discuss/201606.mbox/%3CJIRA.12704304.1396026625000.26207.1466611261062@Atlassian.JIRA%3E | CC-MAIN-2017-51 | refinedweb | 241 | 51.68 |
Language in C Interview Questions and Answers
Ques 26. What is environment and how do I get environment for a specific entry?
While working in DOS, it stores information in a memory region called environment. In this region we can place configuration settings such as command path, system prompt, etc. Sometimes in a program we need to access the information contained in environment. The function getenv( ) can be used when we want to access environment for a specific entry. Following program demonstrates the use of this function.
#include <stdio.h>
#include <stdlib.h>
main( )
{
char *path = NULL ;
path = getenv ( "PATH" ) ;
if ( *path != NULL )
printf ( "nPath: %s", path ) ;
else
printf ( "nPath is not set" ) ;
}
Ques 27. How do I display current date in the format given below?
Saturday October 12, 2002
Following program illustrates how we can display date in above given format.
#include
#include
main( )
{
struct tm *curtime ;
time_t dtime ;
char str[30] ;
time ( &dtime ) ;
curtime = localtime ( &dtime ) ;
strftime ( str, 30, "%A %B %d, %Y", curtime ) ;
printf ( "n%s", str ) ;
}
Here we have called time( ) function which returns current time. This time is returned in terms of seconds, elapsed since 00:00:00 GMT, January 1, 1970. To extract the week day, day of month, etc. from this value we need to break down the value to a tm structure. This is done by the function localtime( ). Then we have called strftime( ) function to format the time and store it in a string str.
Ques 28. If we have declared an array as global in one file and we are using it in another file then why doesn't the sizeof operator works on an extern array?] ;
Ques 29. How do I write printf( ) so that the width of a field can be specified at runtime?.
Ques 30. How to find the row and column dimension of a given 2-D array?nCol: %dn", r, c ) ;
for ( i = 0 ; i < r ; i++ )
{
for ( j = 0 ; j < c ; j++ )
printf ( "%d ", a[i][j] ) ;
printf ( "n" ) ;
}
}
Most helpful rated by users:
- What will be the output of the following code?
void main ()
{ int i = 0 , a[3] ;
a[i] = i++;
printf ("%d",a[i]) ;
}
int x = 3000, y = 2000 ;
long int z = x * y ;
' target='_blank'>Why doesn't the following code give the desired result?
int x = 3000, y = 2000 ;
long int z = x * y ;
char str[ ] = "Hello" ;
strcat ( str, '!' ) ;
' target='_blank'>Why doesn't the following statement work?
char str[ ] = "Hello" ;
strcat ( str, '!' ) ;
- How do I know how many elements an array can hold?
- How do I compare character data stored at two different memory locations? | https://www.withoutbook.com/Technology.php?tech=11&page=6&subject=Language%20in%20C%20Interview%20Questions%20and%20Answers | CC-MAIN-2021-04 | refinedweb | 437 | 73.27 |
On Wed, May 04, 2016 at 03:05:34PM +0200, Gerd Hoffmann wrote: > Resuming the effort to get the gpu device specs merged. > > Support for 2d mode (3d/virgl mode is not covered by this patch) has > been added to the linux kernel version 4.2 and to qemu version 2.4. > > git branch: > > > Rendered versions are available here: > > > > Signed-off-by: Gerd Hoffmann <address@hidden> > --- > content.tex | 2 + > virtio-gpu.tex | 467 > +++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 469 insertions(+) > create mode 100644 virtio-gpu.tex Please add a command execution model section that explains the command lifecycle and interactions between commands. From reading through the spec once I've gathered the fence flag can be used to force execution. I guess a non-fenced command only completes when the operation has finished, too (so that a meaningful success/error value can be produced)? Are there any interactions between the two queues? I guess the resource_id namespace includes both queues. The 64x64 cursor would be initialized on the controlq. The actual VIRTIO_GPU_CMD_UPDATE_CURSOR and VIRTIO_GPU_CMD_MOVE_CURSOR can be sent via either queue. The cursorq does not accept any commands other than VIRTIO_GPU_CMD_UPDATE_CURSOR and VIRTIO_GPU_CMD_MOVE_CURSOR?
signature.asc
Description: PGP signature | https://lists.gnu.org/archive/html/qemu-devel/2016-05/msg01227.html | CC-MAIN-2018-05 | refinedweb | 195 | 59.4 |
Welcome to the Release Preview Windows Store. In this post, we’ll describe a few of the improvements we’ve made to the user experience. We’ll also touch on a few updates that we’ve made to our app certification policy, in an effort to ensure guidelines and expectations remain clear. And, of course, we want to encourage you to dive in and explore the hundreds of preview apps currently in the catalog—including the first desktop app listings. These apps are just the beginning—we’ll continue to add apps during the Release Preview timeframe. Ted Dworkin, Partner Director of Program Management, authored this post.
–Antoine..
Improvements in navigation
Feedback from the Consumer Preview helped identify a set of navigation challenges that we needed to address. When exploring categories, customers had difficulty discovering how to return to the main home page of the Store. We provided a Home link, but folks just didn’t see it.
In addition, customers struggled to find the list of apps they had acquired, for the purposes of installing those apps to additional PCs. This obstructed the value of allowing customers to download apps onto multiple PCs, and syncing the settings between them. We address both problems—how do I get home, and where’s my app list—by taking advantage of the Metro style affordance for navigation: a nav bar. Based on early feedback from customers, the nav bar has helped to solve these two critical navigation issues.
Improving app management
Customers also asked for more control over their app download experience. To address this, we added the Metro style solution for commands, the app bar, to the download manager. This gave us a natural location to move the Pause control and a place for the new Cancel control, which was added for the Release Preview.
Support for Share contract
Previously, we’ve described how contracts are great mechanisms for app-to-app interaction and how the entire Windows 8 operating system gets more powerful and interesting as more apps use more contracts. With the Release Preview version of the Store, we keep this momentum going by supporting the Share contract, allowing you to share info about apps with friends and family right from any app listing page.
These are just a few of the more noticeable improvements to the Store experience that we’ve made based on Consumer Preview feedback. There are thousands of smaller improvements in this preview release—from performance to animations, home page layouts and visual polish.
Desktop apps
One important addition: desktop app listings will show up in the Store for the first time tomorrow, June 1st..
Soon, we’ll post a blog entry further explaining the certification requirements and listing process for desktop apps.
Policy additions and revisions
We first published our app certification policies at our Store Preview Event, where we also reiterated our commitment to transparency in our partnership with developers.
For the Release Preview, we’ve revised a number of policies to increase developer efficiency, and introduced several new policies. As we’ve updated the policies, we’ve kept a revision history to aid consideration of the content. I’ll go over some of the bigger changes now (and note that whenever we introduce significant changes, we’ll blog about them here to further highlight and explain them).
We saw a lot of questions about the relationship between Metro style apps and related websites, so we updated the wording for policy 2.4 to clarify this. The goal is to ensure rich, satisfying experiences within the app and limit context switching to accomplish tasks. Policy 2.4 now reads:
2.4 The primary experiences your app provides must take place within the app
We added clarification to policies 3.5 and 3.6 about input support and respecting system affordances. These policies are designed to ensure predictability and robustness in the experience on every Windows 8 enabled device. We’ve also added considerable design guidance since the Consumer Preview to further explain our principles and provide specific instructions to assist implementation. Policies 3.5 and 3.6 now read: Metro style apps automatically.
Your app must suspend and resume to a reasonable state.
If your app implements an app bar, that bar must show up with a bottom-up swipe.
If your app implements a navigation bar, that bar must show up with a top-down swipe.
We also expanded on polices that relate to transaction support to help ensure that customers are aware of the transaction provider and can be more confident in making purchases. These pertain only to apps that use commerce providers other than the Store. We updated policy 4.7:
4.7 If you use a.
And we added policy 4.8:
4.8 If your app doesn’t use the Windows.ApplicationModel.Store namespace for in-app purchases, your app must prompt the user for authentication to allow a transaction to be accomplished.
The app can offer the user the ability to save this authentication, but the user must have the ability to either require an authentication on every transaction or to turn off in-app transactions.
Finally, we made a number of smaller revisions to improve readability and comprehension. Our certification policies comprise a living document and we will continue to evolve them with the best interest of developers and customers in mind.
Apps
Really, though, this Release Preview is about getting more apps into the catalog and into the hands of customers. That’s the best way for developers to exercise the platform, for customers to engage with the Preview, and for us to continue to evolve and ready all aspects of the service for broad availability.
We’re excited to have lots of new apps to offer in a range of languages—all still free, and ready for immediate download. Here are a few to check out—starting with the always entertaining Fruit Ninja.
Windows 8 is a great gaming platform, but it’s a “no compromises” release and Metro-style apps are for productivity, too. We’re also excited to have the Box app for the Release Preview.
While lots of apps have universal appeal, we want to ensure we have great support for apps that target certain markets or interests. Major League Soccer has delivered a fantastic Release Preview app—rich, immersive and intuitive.
Our full-screen, chrome-less, immersive app model is ideal for viewing content. We’re pleased to have the Financial Times app in available in this Release Preview to give you access to business news and help you stay current on the global economy.
And as we expand our Store language support for both developer markets and consumer catalog offerings, we see more and more developers taking advantage of the market opportunities and language support in the platform.
These are just a few of the many new apps we have in the catalog. We look forward to your feedback. In the meantime, we’re going to take the slightest of breaks, maybe consider taking advantage of Cocktail Flow, and then get back to working hard toward the next public release.
To learn more about some of the other changes we’ve made, check out the Windows app developer blog. And, in the meantime, enjoy the apps!
— Ted
Note: We replaced the first two images in this post on 6/1/2012, as the previous images did not reflect the most recent UI.
When will the App store be open to all developers? RTM?
As a user interface designer who is considering designing apps for Windows 8 I am disappointed with the lack of effort developers have put into their apps when they had months to prepare. — however I think it is your fault in that you are enforcing devs to port their content over into generic templates which is NOT acceptable when Windows 8 launches in a couple of months.
This is Windows Phone all over again where developers were and still are designing generic apps using the same default template. iPad apps have come along way in terms of interaction, usability, branding etc. You simply cannot allow WinRT apps to stay in this bland generic state because what you are allowing to happen is apps that are merely consumption environments like a dull RSS reader instead of a highly branded, highly interactive, highly engaging web app that adds extended functionality that you may not get on a website.
I realise that you are pushing for devs to get their apps into the store as quick as possible and that they will develop over time but judging by your track record for not following through with controlling the whole experience of your products I feel that this area is going to be a major cause for concern when there are so many iPad and even iPhone apps that are so far ahead in unique design and function that I really worry not just for my sake but for the consumers who are looking for an alternative.
I think it's a good time to mention that fruit ninja is not installing properly!
Even after the install, it still says the app needs repair from the windows store
I must say that FlipToast app has got many options but It's too slow, seems to have many errors and sometimes it needs repair from the windows store. Guys, fix that or create a Facebook app please+
Ok, let's see what's wrong:
– Box: non-standard system icons. Users won't be able to determine at first glance what an icon represents.
– Financial Times: seriously?? It totally looks like a web page! Login/register, refresh? They should be in the appbar. And where is Segoe UI Light?
– Larousse: non-metro search icon (it shouldn't ever be there in the first place)
Quality, not quantity. Quantity will eventually come later.
Glad to know window 8 preview released.
What would be nice is if we could actually get a developer registration code for the store..
While I like the concept of the Metro UI for a few things like a Notes app or calculator!
I primary like desktop Apps and as a potential app developer on Windows 8 I am very annoyed that you are not allowing to host desktop apps like the Metro apps. All you do is host a link!
This is terrible! I am almost tempted to just buy an Apple developer account and host desktop apps there!!!! >:-(
You are missing the point Microsoft, not everyone wants everything Metro-ised!
@Hal Motley
I think we aren't missing the point. That is sort of the point 🙂
Most desktop ISVs already have mechanisms to sell, license, distribute their software (including terms and conditions) and for apps in the Windows Store we want to have uniform policies for customers in how they acquire software. So by just hosting the reference to the existing mechanisms, ISVs don't have to compromise their current mechanisms in an effort to be part of the Windows Store. We talked about this when we demonstrated this feature at //build/ in September and in the Store release in December.
As a college student without a summer job (and plenty of free time), I'm starting to get pretty excited about developing for the Windows 8 app store over the summer, as I have made a few apps on Windows Phone, but every time I try to learn WinRT, I continue to run into documentation pages that are completely blank. This is a huge contrast to Silverlight/XNA Development on Windows Phone where there are plenty of samples and walkthroughs available.
How long will it be until we have more than just three basic tutorials? I mean how am I supposed to learn how to take advantage of toast notifications and live tiles when you have documentation pages that are like this? msdn.microsoft.com/…/hh868259.aspx .
Don't get me wrong, I think Windows 8 is an awesome idea, but you need to provide developers with higher quality documentation that Microsoft has always been known for in the past.
I like the look and content of the Windows store, but navigating it still feels clumsy as others have pointed out, see for example news.cnet.com/…/windows-store-youre-still-clumsy-despite-windows-8-boost
While this is sometimes framed as being about "Metro style apps feel clumsy on the desktop", it could feel less clumsy (on all form factors) if they actually implemented some of Microsoft's own guidelines for Metro style apps: msdn.microsoft.com/…/hh868270 (while this is for entertainment apps, they have similar structure to the Store).
e.g.
* Problem: While the long flat spatial layout can be nice for initially exploring the store, it's laborious to navigate to a specific category
* Solution: Implement a header drop down as recommended above: "Using the section header enables users to jump laterally between section titles quickly. For example, consider a user who's browsing the tiles for movies in the comedy category and wants to navigate to the drama category. They can easily use the drop-down for the section header to do that."
(In addition, even the "exploring" use case is currently kind of broken – say you want to scroll through the front page entering each category in turn. Once you're done exploring a category and return to the front page, you're plopped back at the left side of it so have to scroll all the way back to the category you explored. The navigation stack needs to remember your scroll position so you can resume right where you left off.)
* Problem: Exploring the store feels impersonal because apps and categories you don't care about can be promoted over ones you do
* Solution: Implement category rearranging in zoomed-out view as also recommended by Microsoft's own guidelines (but AFAICT not actually implemented in any of their apps): "When using Semantic Zoom, you can enable the user to rearrange category placement on the home page by selecting the group and moving it around. Your users can now personalize the app landing page, which gives them another reason to use your app." I think people would appreciate the spatial layout much more if it were one they could personalize.
Oh, and also the mouse zoom button appears for me on top of the scrollbar instead of in the corner proper – this is horrible, as it's both inherently worse (Fitts's law noncompliance) and inconsistent with the Start screen (people should just be able to perform these gestures by muscle memory instead of having to look for the button in a different place each time). Hope this is a bug/oversight/WIP and not a design decision.
@Hal, I have a calculator App.. I can't get an app store developer registration code!
The one thing I didn't like about the store was navigating in and out of categories.. If scroll to the right 6 or 7 categories and enter the category, when I "back" out of it the list is reset and I have to scroll back to where I was.
The use case is if I'm iterating through all the categories to see "what's new", clicking back resets everything to the beginning. I already looked at categories 1 through 7, I want to be back at category 7 so I can quickly get to 8.
@bzsys
I totally agree the position of the zoom button should be the corner, hopefully they fix this.
I also wanted add a few suggestions about improving navigation: for mouse and keyboard users searches should be done within an app by simply starting to type, similarly to how it's done on the start screen. This seems more consistent and intuitive. Considering how often search is used going to a corner and then a charm just seems like to many unnecessary steps. Also I'm kind of surprised that push scrolling isn't working in the core apps like the marketplace yet, is this going to be added?
I'm really enjoying Windows 8 so far. Good Work!
@Alex Parsons
The link you provided to the MSDN page on working with tiles, badges, and toast notifications goes to the landing page for that section. I see a ton of very details information in that section, and the samples provided there are quite extensive. Perhaps you missed the list of sub-sections on the left-hand side of the page? (i.e. "Creating a tile," "Updating badges," etc?)
i have to agree again about not being able to submit apps to the store. i have been using windows 8 since september last year and have managed to develop a couple of apps but im still unable to get them into the store which is preventing me from going any further with them because i need feedback, suggestions and testing done on various devices so i can fix problems and make improvements and the only way to do that is to get them into the store so people can try them.
if windows 8 is released around october then that is another reason why its important that you open up store submissions as soon as possible because that is only 4 short months away and the only way windows 8 will be a success especially on tablets for the holidays is if the store has 100s if not 1000s of apps available to give consumers a choice and give them something to use on their fancy new tables. people will expect full completed apps and ones without bugs or incompatibilities and it also takes time for apps to go through the verification process so unless the store is opened soon there wont be anything in the store when windows 8 is offiially released and choice will be minimal. the higher the number of apps available at launch will greatly influence the sale of windows 8 devices.
dont leave it until the last minute please open the store
@Brandon Paddock
Well that's embarrassing, you are correct about how I failed to navigate to the topics on the left. It would probably be easier to see if those links were listed in the middle of my page though.
Regardless, I still feel like the quality of the WinRT framework, isn't of the same quality as Silverlight and XNA was for Windows Phone. I know I'm not the only one who thinks this, Paul Thurott expressed similar concerns, and the creator of Weave, pretty much the best news reader app on Windows phone has expressed similar concerns. Link to his twitter is here: twitter.com if you would like to send some help to him.
I'm going to give WinRT a try, that's for sure, and I really hope the documentation improves. Unfortunately I have a feeling that Windows Phone 8 will be focused on WinRT as well, so I am trying to abandon my Silverlight projects to port them to WinRT.
Another thing about how Windows 8 differs from Windows Phone is the developer communication. Maybe I just haven't been following WinRT enough, but there's no figure head for WinRT evangelism. Brandon Watson did a phenomenal job at recruiting developers for Windows Phone, but unfortunately, I don't see anyone making equivalent advances that Watson made for Windows Phone. Once again, maybe it's because I followed channel 9 two years ago, I could be wrong.
Whoa, just realised the search in Charms is app specific when in App, so you can use this to search the Store. I feel a bit slow now for my query about needing to search the Store.
@stzeer6: I definitely agree it would be great to have "just start typing to search" be adopted in apps as well. Not only is it convenient (or if you like "fast and fluid") but after doing it in Start I find I instinctively try it elsewhere and am disappointed when it doesn't work – feels like something is broken.
@Sinofsky: I can understand that and I'm sure all those companies who have invested a lot of work and money into setting up such infrastructure appreciate that they can continue to use it.
However, I don't really understand who this stopped you from ALSO providing a new application model in which we could create "metrofied" deskop applications that ran sandboxed AND were deliverably through the store (which btw, would also give us developers the opportunity to write metrofied desktop applications that run on BOTH x86 and ARM…)
I mean, serious, every time I plug my (awesome ;-)) phone into my pc and see Zune popping up… I'm thinking "why can't I develop a sandboxed desktop app that looks like this?"
Zune is a metrofied desktop app, there's no denying that. The borders, the buttons, the content, the horizontal scrolling, the animations,… It's literally a metro app running inside a desktop window.
Honestly, there isn't a single good reason I can come up with for why you did not provide a desktop accessible WinRT api that would have allowed me to write such ARM/x86 capable desktop apps using C# and XAML.
Here's to hoping that such a platform is being developed as we speak…
FYI: I also fear that because of the lack of such a platform and because of the incredible progress Intel has been making.. windows on ARM is going to be dead on arrival.
Why does window store not available in Myanmar.
Look at that Contoso File Wizard screenshot, in the Contoso File Wizard window it has Android ICS 4.0 virtual buttons bar on the bottom!!
@Woof, oh you didn't get the memo? Android apps run in Windows 8.. one OS to rule them all!
@Zinko Lin, they mispelled Myanmar.. check under 'Burma'.
Не могу войти в магазин,выдает вышло время подключения и проблемы с сервером,хотя проблем нет.Спасибо.
My biggest gripe with the store so far is the sheer number of foreign language apps there are cluttering up my view, even though the option "help me find apps in my own language" is turned on. Is it really that hard to hide non-English apps?
windows app store is only a joke, the real app store is under mac os x.
Microsoft can not do something on they're own, they can only copy things and that was it.
What a shame!!!
I cant change the country in my developer account… WHY?
@ahdr your comment (the second one right after the article) is spot on. While the project templates are a good starting point, it encourages developers to be lazy design wise and come up with boring default apps like on wp7
I need developer Account from Algeria Thanks | https://blogs.msdn.microsoft.com/windowsstore/2012/05/31/windows-store-for-release-preview/ | CC-MAIN-2016-44 | refinedweb | 3,818 | 58.72 |
Edison Arduino i2ctoadaze Nov 9, 2014 8:00 AM
Sorry for yet another post about getting i2c to work, but here goes. Since I was unable to get my Adafruit type i2c LCD display working on the Edison, I decided to try something simpler or so I thought... a Sparkfun SEN-11931 i2c tmp102 temperature sensor.
This sensor can be hard configured to use i2c addresses 0x48 or 0x49. Needless to say, it works fine on an Arduino UNO connected to A4/A5 using this sketch.
/(2000); /;
}
However, even though it compiles and uploads OK, nothing at all prints to Arduino's Serial Monitor screen when it's running on the Edison Arduino board. Using a scope, I confirmed the Edison-Arduino is sending out clock on A5 and data on A4, so I know the script is running and I have the right pins. BTW, I scoped it with the sensor disconnected.
I have to assume the issue lies in the Wire.h library as implemented in the Edison-Arduino IDE which has:
#define I2C2 0x4819c000
#define I2C1 0x00000000
#define WIRE_INTERFACES_COUNT 1
From other posts, I am led to believe 'I2C2 0x4819c000' equates to BUS ID 6. For one thing, I don't understand how this translates to ID 6?
In any case, how can I get the above script to run on an Edsion-Arduino? The sensor itself doesn't care what bus it is on, right? It only cares about the i2c address.
1. Re: Edison Arduino i2conehorse Nov 10, 2014 12:45 PM (in response to toadaze)
The Edison is an x86, 32-bit machine. Maybe instead of int, you should try byte or uint8_t for your device address. Can you verify that you are able to read the WHO_AM_I egister of your sensor. The A4/A5 on the Arduino breakout board use the I2C6 port, one of two I2C ports on the Edison. Do you have the proper pullups on the I2C lines? The Edison has internal pullups which might be interfering with you signal, but if you can see the trace, it sounds like your device address is wrong or messed up.
2. Re: Edison Arduino i2ctoadaze Nov 10, 2014 6:15 PM (in response to onehorse)
I can read the temp sensor's i2c address on my Arduino UNO using Nick Gammon's i2c Scanner sketch. It says it's hex 48 (decimal 72). But when I try to run the i2c Scanner sketch on my Edison-Arduino it doesn't do anything. Uploads OK but no output.
The only library it loads is wire.h just like the above sketch. So I think the issue is in that library, which apparently has been specially modified for the Edison-Arduino.
Standard wire.h has:
#ifndef TwoWire_h
#define TwoWire_h
#include <inttypes.h>
#include "Stream.h"
#define BUFFER_LENGTH 32
but Edison-Arduino wire.h has:
#ifndef TwoWire_h
#define TwoWire_h
#include "Stream.h"
#include "variant.h"
#define BUFFER_LENGTH 32
#define I2C2 0x4819c000
#define I2C1 0x00000000
#define WIRE_INTERFACES_COUNT 1
So, I'm still puzzled by #define I2C2 0x4819c000 and how that translates to Bus ID 6.
Frankly, I have found very few tried & true Arduino sketches that run on the Edison-Arduino, except for simple sketches that don't involve any IO.
3. Re: Edison Arduino i2cStan_Gifford Nov 11, 2014 6:20 PM (in response to toadaze)
Hi, I intend to post some code (but havn't yet)......
I converted some Uno based code that I wrote that does the following;
1. Talks to adafruit Ultimate GPS via Serial1 (Had to convert from software serial - #define ss Serial1 was enough (and don't init software serial :-)
2. Talks to an I2c 16/2 LCD display
3. Backlight control of display is controlled by a pot.
4. Reads a (nicely debounced if I say so myself) button to change what gets displayed on LCD.
5. For GPS decode uses TinyGPS - and I stole his smartdelay function from his examples. ()
On the whole it went pretty well - It was a 2-stage conversion because on my UNO I was using the adafuit libraries - so converted to Edison plus tinyGPS!
I am still validating the code - currently my edison is playing up - segfault - but ran out of diagnosis time.
It did prove the LCD via I2c is working fine - and the routines to pull data from the GPS are working fine.
So FWIW, the conversion from UNO to Edison went fairly well - took me about an hour.
Stan
4. Re: Edison Arduino i2cStan_Gifford Nov 12, 2014 2:35 AM (in response to Stan_Gifford)
Spoke too soon - just about to make new post on issues with the code - signature points to I2c
Reason edison was 'playing up' was because / was full of journals - have reconfigured journal to put journals in volatile so problem sort of gone away (Understand new yocto image coming soon to properly fix issue).
Journals were created (I believe) by the sketch which cause a lot of i2c issues - (and I moved from Due to Edison because of the issues with i2c on the Due :-( )
Incidently - having looked at the edison wire.cpp code I am puzzled - the wire.begin seems to ignore any slave address given- can anyone confirm that the current wire implementation in edison is master (no address) mode only? (or have I lost the ability to read code?) - or is the design of the wire implementation to be master with GC for out of band transmissions to the edison master?
Stan
5. Re: Edison Arduino i2conehorse Nov 12, 2014 10:07 AM (in response to toadaze)
When I hooked up an MPU9150 sensor breakout board to the Edison/Arduino breakout board using A4/A5 for SDA/SCL, 3V3 power and GND, I was able to run an Arduino sketch written and working for the 3V3 8 MHz Arduino Pro Mini without modification. The Edison/Arduino breakout board should behave just like an Arduino UNO if you are using the Edison Arduino IDE.
6. Re: Edison Arduino i2ctoadaze Nov 12, 2014 10:36 AM (in response to onehorse)
onehorse - I don't have a MPU9150 but was wondering if you could try the following Arduino sketch? It's something I use to confirm i2c addresses. It works fine on my UNO's but not on my Edison-Arduino breakout board. I would really appreciate hearing how it works for you, since you're having a lot more success than me.
// I2C Scanner
// Written by Nick Gammon
// Date: 20th April 2011
#include <Wire.h>
void setup() {
Serial.begin (115200);
// Leonardo: wait for serial port to connect
while (!Serial)
{
}
Serial.println ();
Serial.println ("I2C scanner. Scanning ...");
byte count = 0;
Wire.begin();
for (byte i = 8;
void loop() {}
7. Re: Edison Arduino i2conehorse Nov 12, 2014 10:48 AM (in response to toadaze)
I will try it this weekend when I recover my Edison and Arduino breakout board from another work site.
In the mean time, here is an I2C scanner I have been using with success, but I haven't tried it on the Edison lately. See if it gives you any different result.
8. Re: Edison Arduino i2conehorse Nov 12, 2014 10:51 AM (in response to toadaze)
Just to make sure, you are using the Edison-specific Arduino IDE with the Edison-specific Wire.h library, aren't you?
9. Re: Edison Arduino i2ctoadaze Nov 12, 2014 11:36 AM (in response to onehorse)
Hey, thanks! Yes, I am definitely using the modified Arduino IDE from Intel's downloads on the Edison and its built-in Wire.h library.
I tried two different i2c devices with your program on my Edison-Arduino breakout board and only get "scanning, no i2c devices found". So then I tried your program on my UNO and it shows the i2c addresses just fine. Of course, it's using its own unmodified Wire.h library.
One thing I haven't done is try pullup resistors on SDA and SCL, although the devices I tried are supposed to have their own pullups and do work OK on an UNO.
10. Re: Edison Arduino i2ctoadaze Nov 12, 2014 2:18 PM (in response to toadaze)
I tried adding 10K pullup resistors to SDA and SCL but it made no difference - still no i2c devices found. I went with 10K instead of say 4.7K, because my devices are supposed to already have pullups built-in to them.
I'm still puzzled over this issue of BUS ID 6. How does one do that in an Arduino sketch? Or is that more likely part of the Wire.h library? And if so, where? Has the modified Wire.h library for the Edison-Arduino magically handled the change to BUS ID 6 somewhere in not so obvious code?
11. Re: Edison Arduino i2conehorse Nov 12, 2014 2:26 PM (in response to toadaze)
Yes, it's magic! Or software...
In the Edison Arduino library they made a default choice to assign I2C6 (SDA/SCL) to A4/A5 on the Arduino breakout board. This should make it work like any other Arduino UNO. The GY-80 already has pullup resistors so that is not your problem. Of course, I don't know what is 'cause it should work!
On the adapter board we made I wired up SDA/SCL inadvertently to I2C1 (who would have though 6 would be default, 1 comes before 6!) before I had the Arduino or Mini breakout board schematics. We had to change the default to I2C1 which was straightforward to do in the Wire library. But you should not need to do this.
I will try again this weekend with the latest Edison Arduino IDE and the Arduino breakout board and see if it still works like it did before. I'll let you know what I find.
12. Re: Edison Arduino i2ctoadaze Nov 12, 2014 3:47 PM (in response to onehorse)
Since I couldn't get any of my i2c LCD displays working on the Edison-Arduino, the other day I ordered a Grove (i2c) RGB-LCD. It arrived this afternoon, and while it works just fine using JavaScript (node.js) on the Edison-Arduino, it behaves the same as all my other i2c devices trying to use it through the Arduino IDE - "no i2c devices found". In any case, this proves my connections, hardware and pullups are OK.
But here is another mystery - I keep reading about needing to use BUS ID 6 on the Edison instead of 0 like the Galileo uses. But this JavaScript is using 0 NOT 6 and it works. In fact, it works no matter what BUS ID I use. Maybe something was changed in the mraa/upm libraries after all the posting about BUS ID's.
//Load i2clcd module
var LCD = require('jsupm_i2clcd');
//Initialize Jhd1313m1 at 0x62 (RGB_ADDRESS) and 0x3E (LCD_ADDRESS)
var myLcd = new LCD.Jhd1313m1 (0, 0x3E, 0x62); // BUS ID = 0
myLcd.setCursor(0,0);
myLcd.write('Hello World');
13. Re: Edison Arduino i2cCMata_Intel Nov 28, 2014 5:56 PM (in response to toadaze)
Hi toadaze;
Have you been able to use I2C and LCD displays with your IDE?
Regards;
CMata
14. Re: Edison Arduino i2ckrish11 Jan 14, 2015 10:26 PM (in response to Stan_Gifford)
Hello
I am trying to configure Intel Edison breakout board with Adafruit Ultimate GPS to track the live location of the device.
I tried some code but it was giving an error regarding Software Serial
I am a complete beginner and hoping some help from your side
Can you please explain how to solve SoftwareSerial issue?
Connections between Edison break out board and Adafruit GPS Ultimate:
Vin J20- pin 2 System 3.3V Output
GND J19- pin 3 GND
Rx J20- pin 4 GP45
Tx J20- pin 5 GP47
Are these connections fine for tracking the live location?
Thanks in advance | https://communities.intel.com/thread/56899 | CC-MAIN-2017-39 | refinedweb | 1,978 | 71.44 |
Introduction
Have you ever wondered how you can delete a running program? Today, I will show you how to create a program and, whilst the program is running, delete its executable file. This is quite handy when you are creating an Uninstaller application or must reload your application after it was patched.
Practical
The object of this program is to demonstrate two different ways in which you can create a self-destruction application. Open Visual Studio and create either a C# or a Visual Basic.NET Windows Forms application. Design it as shown in Figure 1.
Figure 1: Design
Add the necessary Namespaces so that you will be able to do file manipulation.
C#
using System; using System.Diagnostics; using System.IO; using System.Windows.Forms;
VB.NET
Imports System.IO
Create a Sub procedure aptly named SelfDestruct.
C#
private void SelfDestruct() Process procDestruct = new Process(); string strName = "destruct.bat"; string strPath = Path.Combine(Directory .GetCurrentDirectory(), strName); string strExe = new FileInfo(Application.ExecutablePath) .Name; StreamWriter swDestruct = (Exception) { Close(); } }
VB.NET
Private Sub SelfDestruct() Dim procDestruct As Process = New Process() Dim strName As String = "destruct.bat" Dim strPath As String = Path.Combine _ (Directory.GetCurrentDirectory(), strName) Dim strExe As String = New _ FileInfo(Application.ExecutablePath).Name Dim swDestruct As StreamWriter = ex As Exception Close() End Try End Sub
The SelfDestruct sub procedure creates a batch file named destruct.bat. A batch file is a computer file containing a list of instructions to be carried out in turn. The batch file you create will delete the running executable file.
Add the following code behind Button1.
C#
private void button1_Click(object sender, EventArgs e) { SelfDestruct(); Close(); }
VB.NET
Private Sub button1_Click(ByVal sender As Object, _ ByVal e As EventArgs) SelfDestruct() Close() End Sub
The first button simply calls the SelfDestruct sub procedure and then calls the Close method to close the form.
Add the following code to your Button2.
C#
Process.Start("cmd.exe", "/C choice /C Y /N /D Y /T 3 & Del " + Application.ExecutablePath); Application.Exit();
VB.NET
Process.Start("cmd.exe", "/C choice /C Y /N /D Y /T 3 & Del " _ & Application.ExecutablePath) Application.[Exit]()
This little piece of code spawns the Command prompt and executes the command to delete the currently running EXE file; then, it exits the application. This works, but the only problem is that it displays the Command Prompt window whilst deleting the file. To circumvent this, edit the preceding code to look like the following.
C#
private void button2_Click(object sender, EventArgs e) { ProcessStartInfo piDestruct =); }
VB.NET
Private Sub button2_Click(ByVal sender As Object, _ ByVal e As EventArgs) Dim piDestruct As ProcessStartInfo =) End Sub
Here, you create a ProcessStartInfo object and supply its parameters as to what to execute as well as its Window Style.
Figure 2 shows the EXE still present while the program is running. Figure 3 doesn't show the EXE after its deletion.
Figure 2: Running
Figure 3: Deleted
Conclusion
Although this is pretty useful when it comes to uninstalling and reinstalling applications, it is tricky territory. Windows is designed to not delete files that are running; circumventing this feature may open unwanted doors. | https://mobile.codeguru.com/csharp/.net/net_general/tipstricks/creating-a-program-that-can-self-destruct-with-.net.html | CC-MAIN-2018-47 | refinedweb | 526 | 58.79 |
React-Touch is a set of wrapper components that handle touch interactions in a more declarative way, abstracting out and giving you hooks into behaviors such as dragging, holding, swiping, and custom gestures. React-Touch also works with mouse events as well.
Here's a quick example of the API.
import { Holdable } from 'react-touch';<Holdable onHoldComplete={handleHold}>({ holdProgress }) => <Button style={{opacity: holdProgress}} /></Holdable>
npm install react-touch --save
If you've ever written mobile web software, then you might've found yourself needing the ability to touch drag a component, measure a hold, or react to a swipe or gesture. This library is a set of wrapper components that abstract out the details of those things so you can wrap your components and move on.
Exports:
defineHold
defineSwipe
Holdable
Draggable
Swipeable
CustomGesture
defineHold(config?: Object)
Used in conjuction with
Holdable,
defineHold is an optional helper function that creates a configuration for your holdable component. The arguments to it are:
config: Optional. Object with the following keys:
updateEvery: Optional. Defaults to 250. Units are in milliseconds.
holdFor: Optional. Defaults to 1000. Units are in milliseconds.
const hold = defineHold({updateEvery: 50, holdFor: 500});<Holdable config={hold} onHoldComplete={handleHold}><Button /></Holdable>
defineSwipe(config?: Object)
Used in conjuction with
Swipeable,
defineSwipe is an optional helper function that creates a configuration for your swipeable component. The arguments to it are:
config: Optional. Object with the following keys:
swipeDistance: Optional. Defaults to 100. Units are in pixels.
const swipe = defineSwipe({swipeDistance: 50});<Swipeable config={swipe} onSwipeLeft={deleteEmail}><Email /></Swipeable>
<Holdable />
Used to create a component that understands holds.
Holdable will give you hooks for the progress and the completion of a hold. You can pass a component or a function as its child. Passing a function will gain you access to the hold progress.
<Holdable onHoldComplete={handleHold}>({ holdProgress }) => <Button style={{opacity: holdProgress}} /></Holdable>
onHoldProgress?: Function
When the hold makes progress, this callback is fired. Update intervals can be adjusted by the
updateEvery key in the configuration.
onHoldComplete?: Function
When the hold has completed, this callback is fired. Length of hold can be
adjusted by the
holdFor key in the configuration.
holdProgress
<Draggable />
Used to create a component that can be dragged.
Draggable requires a
style prop defining its initial position and will pass updates to the child component via a callback.
<Draggable style={{translateX: 150, translateY: 200}}>{({translateX, translateY}) => {return (<div style={{transform: `translate3d(${translateX}px, ${translateY}px, 0)`}}><Bubble /></div>);}}</Draggable>
style: Object Required. An object that defines the initial position of the draggable component. You can pass any of the following styles to it and they'll be updated and passed back out in the callback with every animation tick.
translateX
translateY
top
left
right
bottom
Any of the above keys depending on what you set as your
style. Additionally:
dx
dy
<Swipeable />
Used to create a component that understands swipes.
Swipeable gives you hooks to the swipe directions up, down, left, right, with the swipe threshold being customized using the
defineSwipe helper.
<Swipeable onSwipeLeft={deleteEmail}><Email /></Swipeable>
onSwipeLeft?: Function
When the swipe threshold has been passed in the left direction, fire this callback.
onSwipeRight?: Function
When the swipe threshold has been passed in the right direction, fire this callback.
onSwipeDown?: Function
When the swipe threshold has been passed in the down direction, fire this callback.
onSwipeUp?: Function
When the swipe threshold has been passed in the up direction, fire this callback.
<CustomGesture />
Used to create a component that understands a customized gesture. Gestures are passed through the config prop. When the gesture is recognized,
onGesture will fire.
Gestures are just a combination of discrete linear movements. For instance, a "C" gesture would be composed of a left, down-left, down, down-right, and right. The user doesn't have to do this perfectly, the library will do a distance calculation and fire or not fire the
onGesture callback based off that. This algorithm is a port of a Swift library by Didier Brun.
import { CustomGesture, moves } from 'react-touch';const CIRCLE = [moves.RIGHT,moves.DOWNRIGHT,moves.DOWN,moves.DOWNLEFT,moves.LEFT,moves.UPLEFT,moves.UP,moves.UPRIGHT,moves.RIGHT,];<CustomGesture config={CIRCLE} onGesture={unlockApp}><AppLockScreen /></CustomGesture>
onGesture?: FunctionCallback fired when the gesture is complete.
Want to be able to drag and hold a component? You can wrap react-touch components with other react-touch components to achieve this. For example:
const hold = defineHold({updateEvery: 50, holdFor: 500});<Holdable config={hold} onHoldComplete={() => console.log('held out')}><Draggable style={{translateX: 150, translateY: 200}}>{({translateX, translateY, holdProgress}) => {return (<div style={{transform: `translate3d(${translateX}px, ${translateY}px, 0)`}}><Bubble style={{opacity: holdProgress}} /></div>);}}</Draggable></Holdable>
Notice the callback argument keys are the combination of the two parent components. This feature means you don't have to do multiple nested callbacks to achieve the same effect. | https://www.npmjs.com/package/react-touch | CC-MAIN-2017-26 | refinedweb | 799 | 50.02 |
EEx v1.9.4 EEx.Engine behaviour View Source
Basic EEx engine that ships with Elixir.
An engine needs to implement all callbacks below.
An engine may also
use EEx.Engine to get the default behaviour
but this is not advised. In such cases, if any of the callbacks
are overridden, they must call
super() to delegate to the
underlying
EEx.Engine.
Link to this section Summary
Functions
Handles assigns in quoted expressions.
Callbacks
Invoked at the beginning of every nesting.
Called at the end of every template.
Invokes at the end of a nesting.
Called for the dynamic/code parts of a template.
Called for the text/static parts of a template.
Called at the beginning of every template.
Link to this section Types
state()View Source
Link to this section Functions
handle_assign(arg)View Source
Handles assigns in quoted expressions.
A warning will be printed on missing assigns. Future versions will raise.
This can be added to any custom engine by invoking
handle_assign/1 with
Macro.prewalk/2:
def handle_expr(state, token, expr) do expr = Macro.prewalk(expr, &EEx.Engine.handle_assign/1) super(state, token, expr) end
Link to this section Callbacks
handle_begin(state)View Source
Invoked at the beginning of every nesting.
It must return a new state that is used only inside the nesting.
Once the nesting terminates, the current
state is resumed.
handle_body(state)View Source
Called at the end of every template.
It must return Elixir's quoted expressions for the template.
handle_end(state)View Source
Invokes at the end of a nesting.
It must return Elixir's quoted expressions for the nesting.
handle_expr(state, marker, expr)View Source
Called for the dynamic/code parts of a template.
The marker is what follows exactly after
<%. For example,
<% foo %> has an empty marker, but
<%= foo %> has
"="
as marker. The allowed markers so far are:
""
"="
"/"
"|"
Markers
"/" and
"|" are only for use in custom EEx engines
and are not implemented by default. Using them without an
appropriate implementation raises
EEx.SyntaxError.
It must return the updated state.
handle_text(state, text)View Source
Called for the text/static parts of a template.
It must return the updated state.
init(opts)View Source
Called at the beginning of every template.
It must return the initial state. | https://hexdocs.pm/eex/EEx.Engine.html | CC-MAIN-2020-05 | refinedweb | 377 | 61.33 |
Inform!
Cheers,
Shawn
If you have been following Informix events for the last three years, you have heard us talking about the Wire Listener. This is a Java daemon process that was originally implemented in 12.10.xC2 to allow seamless compatibility with MongoDB. A later variation of the Listener provided the REST interface to Informix. The recent (June 2016) 12.10.xC7 release now brings MQTT support to the Informix party.
MQTT is a simple and lightweight PUBlish-SUBscribe messaging protocol that is widely used in Internet of Things (IoT) applications. Since it is asynchronous, one common IoT use case is for sensors to publish readings at configured intervals to a predefined topic hosted by a MQTT message broker. Applications register interest in various sensor readings by subscribing to the particular topic for that sensor.
There are many implementations of MQTT, ranging from open source (Eclipse Mosquitto and Paho) to commercial implementations (IBM Message Sight). We've integrated a MQTT message broker into the Informix Wire Listener component, and so now with 12.10.xC7, applications can use one of the various client API libraries to directly publish messages to the Informix broker. There are client libraries for most programming languages/environments. This page () gives a very complete list of MQTT client implementations.
So, why is this of interest and how might it be used? Let’s take the use case of a sensor generating a reading each second. For this simple example, we will read the amount of free memory available to the Java VM. (Ok, this is not really a sensor in strict terms, but it makes the application source code easier to show. You can swap some other “real” sensor and the MQTT publish part of the application will work the same way) .
Before we look at the implementation, we will set up the database and start the listener.
1) We need a database table to store the data. MQTT readings can be stored in relational tables, in JSON collections, and in TimeSeries storage. For this example, we will store the readings in a relational table in our "test" database. Our table schema is simple – it contains just the sensor_id, the timestamp and the sensor reading value. Here is the dbschema output for the table "sensor_tab":
DBSCHEMA Schema Utility INFORMIX-SQL Version 12.10.FC7AEE
{ TABLE "smoe".sensor_tab row size = 23 number of columns = 3 index size = 0 }
create table "smoe".sensor_tab
(
sensor_id integer,
tstamp datetime year to fraction(5),
value float
);
revoke all on "smoe".sensor_tab from "public" as "smoe";
This table becomes the “topic” to which the sensor data is published.
2) Create the MQTT properties file. The key item here is “listener.type=mqtt’. Here is my simple properties file, MQTTListener.properties:
3) Start the MQTT listener. The MQTT listener is the same JAR file as the NoSQL and REST variants, and is started in the same manner. Just remember to start it with your MQTT properties file. You’ll want to listen on a different port than you are using for the NoSQL or REST listeners to avoid conflicts. Start the listener using a command similar to this:
4) Now that we have the database created and the MQTT listener started, we can look at the application that publishes data. In this example we will use the Java Eclipse Paho MqttClient synchronous API.
The program connects to the message broker (the Informix MQTT Listener) and then loops to get 100 readings (obviously a trivial number for this demo). Each time through, it grabs the system time, reads the available memory for the JVM “sensor”, generates a JSON document with the appropriate values, and publishes the JSON to the specified topic (the topic is the database.table_name). The program sleeps for a second and then repeats. Here is the sample Java code for this:
// Very rudimentary Java example demonstrating publishing of sensor data to the
// Informix database using MQTT. This implementation uses the Eclipse Paho
// MqttClient synchronous API. Note that there is nothing particular about
// Informix here. The database and table become the "topic" to which the
// message is published.
//
// This example (as written) runs against an Informix database named "test"
// and a table named "sensor_tab". The "sensor_tab" table has three columns:
// (sensor_id : integer,
// tstamp : datetime year to fraction(5),
// value : float)
//
// smoe@us.ibm.com 9 June 2016 java.util.Calendar;
public class MqttPublishSample {
public static void main(String[] args) {
// Topic to which the MQTT message is published. To work with
// Informix, the topic represents the database and table using
// the format of <database>.<table>
String topic = "test.sensor_tab";
int qos = 0;
// The broker is the port configured for the Informix MQTT
// listener
String broker = "tcp://localhost:27019";
String clientId = "JavaSample";
MemoryPersistence persistence = new MemoryPersistence();
Calendar calendar;
java.util.Date now;
java.sql.Timestamp currentTimestamp;");
for (int i = 0; i < 100; i++) {
// get the current timestamp
calendar = Calendar.getInstance();
now = calendar.getTime();
currentTimestamp = new java.sql.Timestamp(now.getTime());
String ts = new String(currentTimestamp.toString());
// For the sensor value, we will use the total amount of
// free memory available to the JVM */
String fm = new String
(Long.toString(Runtime.getRuntime().freeMemory()));
/* create the JSON document using the various bits. The
JSON should matches the structure of the database table:
{
sensor_id : < >,
tstamp : < >,
value : < >
}
*/
// Create a string representing the JSON from the various bits
String content = new String("{sensor_id:23, tstamp:\"" + ts
+ "\", value:\"" + fm + "\" }");
System.out.println("Publishing message: "+ content);
// Publish the JSON document as a message to the database
// table topic
MqttMessage message = new MqttMessage(content.getBytes());
message.setQos(qos);
sampleClient.publish(topic, message);
System.out.println("Message published");
// Sleep until time to take the next reading. Values are
// in milliseconds.
Thread.sleep (1000);
}
/* Disconnect to clean up when finished. This Java implementation
is synchronous, and so the message queue is flushed as part of
the disconnect process. */
sampleClient.disconnect();
System.out.println("Disconnected");
System.exit(0);
}
// Catch block for the exception that could be thrown from the
// Thread.sleep() call
catch (InterruptedException swallowed) {
}
// Catch block for the MQTT exceptions - just print out some debug info
catch(MqttException me) {
System.out.println("reason " + me.getReasonCode());
System.out.println("msg " + me.getMessage());
System.out.println("loc " + me.getLocalizedMessage());
System.out.println("cause " + me.getCause());
System.out.println("excep " + me);
me.printStackTrace();
}
}
}
5) Compile and run the program. Here is some debug output as the messages are being published:
6) Examine the data in the sensor_tab table. See that the sensor data has been “published” to this table:
This is a basic example, but you can see the process to publish data using MQTT is very simple. Some benefits of this approach? You can push data to your Informix database from a wide variety of languages, not just your usual suspects of C, C++ and Java, but also Go, Haskell, Erlang, Lua, Python, Ruby, and Swift, as there are MQTT client libraries for all of these languages and more. You don’t need to use a SQLI or DRDA driver - if fact, you don’t need to use SQL at all. And the process is asynchronous, and so your application can publish and go on. This is clearly not the right paradigm for many types of OLTP applications, but depending on your application requirements, this may be a good mechanism to get data inserted into your Informix database. Give it a try and let us know what you think. the standard MongoDB create_index() function from your MongoDB compatible program connected to Informix 12.10, each part of the compound index will require 4K of storage. This is because we don’t know the data type of the data to be indexed, and so have to allocate index space for the worst case scenario, long string data. The max size of any Informix index is 32K, and so only seven parts for a MongoDB style compound key can only be specified. (Yes 4K * 7 is only 28K, but there is some overhead).
As an example of the standard createIndex() command that creates an ascending index on the state field and a descending index on the zipcode field:
db.collection.createIndex( { state: 1, zipcode: -1} )
Informix does not know which data types these two fields are, and so it uses the internal bson_get() function mechanism to create the index and then reserves 4K space for each bson.
So if you have a need to specify an index with more than 7 parts, what to do? We have developed an extension to the createIndex() function which allows specification of the data type. As an example of this Informix alternative:
db.collection.createIndex( { state: [1, “$string”], zipcode: [-1, “$int”] } )
This call uses the same MongoDB createIndex() command, but includes the data types for the two fields. (String for the values for state and int(eger) for the values for zipcode). Knowing the data types allows Informix to create the index more efficiently, using internal bson_value_<data type>() functions, and should allow many more parts to be included in the compound index.
Depending on the data types of the underlying data, this approach can allow a compound index to include up to 16 parts, which is the Informix limit for compound indexes. This is still short of the MongoDB limit of 31 parts, but much better than the 7 parts supported with the standard createIndex() syntax. The Informix extensions would not be recognized if the application is run against MongoDB, so if you are writing an app that works with both databases, put these changes in some kind of IF check to run only with the Informix server.
For. | https://www.ibm.com/developerworks/community/blogs/smoe?lang=es | CC-MAIN-2017-17 | refinedweb | 1,599 | 55.34 |
I.Making files into directories caused only two applications out of the entire OS to notice the change, and that was because of a bug in what error code we returned that we are going to fix. You think that was a disaster; I think it was a triumph.Now a cleanly architected filesystem with no attributes and just files and directories that can do everything attributes are used for exists. You don't want it to have the competitive advantage. Instead, you want it to have its clean design excised until you have something that duplicates it ready to go, and only then should it be allowed that users will use the features of your competitor's filesystem which you disdained implementing for so long.Since you never studied or understood namespace design principles (or you would not have created and supported xattrs), you want to rename it to be called VFS, rewrite what we have done, and take over as the maintainer, mangling its design in a committee clusterfuck as you go. We have just implemented very trivial semantic enhancements of the FS namespace, nothing like as ambitious as or WinFS, and you are already pissing your pants.Is that a fair summary?Eat my dust. HansPSI should of course qualify what I have said. The use of files and directories in place of attributes is not a finished work. It has bugs, sys_reiser4() does not yet work, and there are little features still missing like having files readdir ignores.Still, except for the bugs, what we have is usable, and there are a lot of happy reiser4 users right now even with the bugs. It will need a little bit more time, and then all the pieces will be in place.PPSIf you implement your filesystems as reiser4 plugins, and rename reiser4's plugin code to be called "vfs", your filesystems will go faster. Not as fast as reiser4 though, because it has a better layout and that affects performance a lot, but faster is faster.... See for details.PPPS. The gap between us is about to widen further.Christoph Hellwig wrote:>After looking trough the code and mailinglists I'm quite unhappy with>a bunch of user-visible changes that Hans sneaked in and make reiser4>incompatible with other filesystems >if we leave you in the dust, run faster.... not my problem....>Given these problems I request that these interfaces are removed from>reiser4 for the kernel merge, and if added later at the proper VFS level>after discussion on linux-kernel and linux-fsdevel, like we did for>xattrs.>>> >If you can't help fight WinFS, then get out of the way. Namesys is on the march. Read, be smart, recognize that reiser4 is faster and more flexible than your storage layers because we are older and wiser and worked harder at it, join the team, and start contributing plugins that tap into the higher performance it offers.Microsoft tried to build a storage layer that could handle small objects without losing performance, failed, and gave up at considerable cost to their architecture and pocketbook.We just broke a hole in the enemy line. You could come swarming through it with us, but it sounds like you prefer complaining to HQ that we are getting too far in front of you.-To unsubscribe from this list: send the line "unsubscribe linux-kernel" inthe body of a message to majordomo@vger.kernel.orgMore majordomo info at read the FAQ at | http://lkml.org/lkml/2004/8/25/183 | CC-MAIN-2017-51 | refinedweb | 584 | 69.01 |
#include <hallo.h> * Goswin von Brederlow [Tue, May 16 2006, 11:55:18PM]: > What do you mean with invasive? Multiarch is designed to be > implemented without disrupting the existing archive or mono-arch > systems at any time. Each package can get converted to multiarch by > itself and once a substantial number of packages have done so a > multiarch dpkg/apt/aptitude can be used. And that is why I question it. Do we need that? You demonstrated that it is quite easy to setup the depencency chain for a package... but why should we care about change the whole distribution for the sake of few 3rd party packages if we have sufficient workarounds? > But cooking the packages is not 100% successfull and involves a lot of > diversions and alternatives. Every include file gets diverted, every > binary in a library gets an alternative. All cooked packages depend on > their uncooked other architecture version for pre/postinst/rm scripts, > forcing both arch to be installed even if only the cooked one is > needed. I don't see a bad problem with that, sounds like an acceptable compromise. > And still some things won't work without the multiarch dirs being used > by any software using dlopen and similar. That includes X locales, > gtk-pixmap, pango to start with. Such things are not okay but there could be few workarounds as well. > It works for a stable release but for unstable the constant stream of > changes needed in the cooking script would be very disruptive for > users. Only if you port the whole distribution. If you port few dozens of library packages, maintaining them should be feasible. > It also is disruptive to building packages. Build-Depends will only > work for the native arch and not for the cooked packages and > building for the cooked arch will give precooked Depends (I do cook > shlibs files) so they are invalid for uploads. This problem is only implied by "porting the whole arch and using everything like a native package". Eduard. | https://lists.debian.org/debian-devel/2006/05/msg00837.html | CC-MAIN-2015-48 | refinedweb | 334 | 63.59 |
Copyright © 2004.
This visual layout and structure of the specification was adapted from the FOAF Vocabulary Specification by Dan Brickley and Libby Miller..
NOTE: This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications can be found in the W3C technical reports index at.
This specification is an evolving document. This document is generated by combining a machine-readable SIOC Core Ontology Namespace expressed in RDF/XML with a specification template and a set of per-term documents.. SIOC, is an attempt to link online community sites, to use Semantic Web technologies to describe the information that communities have about their structure and contents, and to find related information and new connections between content items and other community objects. SIOC is based around the use of machine-readable information provided by these sites. ().Is that MUST be used by implementations of this specification are: | email | email_sha1 | feed | function_of | has_administrator | has_container | has_creator | has_function | has_host | has_member | has_moderator | has_modifier | has_owner | has_parent | has_reply | has_scope | has_space | has_subscriber | has_usergroup | host_of | id | ip_address | link | links_to | member_of | moderator_of | modifier_of | name | next_by_date | next_version | note | num_replies | num_views | owner_of | parent_of | previous_by_date | previous_version | related_to | reply_of | scope_of | sibling | space_of | subscriber_of | topic | usergroup_of |.
Here is a very basic document describing a blog entry:
>
This brief example introduces the basics of SIOC. It says:
sioc:Postobject identified as has the following properties:
dc:
sioc:content. elements for simple bibliographic description, FOAF, RSS 1.0, etc.) can also be mixed in with SIOC terms, as can local extensions. SIOC is designed to be extended, and some modules have been added (see below).
SIOC modules are used to extend the available terms and to avoid making the SIOC Core Ontology too complex and unreadable. At present SIOC has two modules: Types and Services..
SIOC introduces the following classes and properties. See the SIOC Core Ontology Namespace in RDFS/OWL for more details..
Item - A content Item that can be posted to or created within a Container.
Item is a high-level concept for content.
Role - A Role is a function of a User within a scope of a particular Forum, Site, etc.
Roles are used to express functions or access control privileges that Users may have..
sioc:User describes properties of an online account, and is used in combination with a foaf:Person (using the property sioc:account_of) which describes information about the individual itself..
about - Specifies that this Item is about a particular resource, e.g., a Post describing a book, hotel, etc..
administrator_of - A Site that the User is an administrator of.
attachment - The URI of a file attached to a Post.
avatar - An image or depiction used to represent this User.
container_of - An Item that this Container contains.
Containers and Forums may contain Items or Posts, and this property is used to link to these.
content - The content of the Post in plain text format.
This property is for a plain-text rendering of the content of a Post. Rich content (e.g., HTML, wiki markup, BBCode, etc.) can be described using the Content class from AtomOwl or the content:encoded property from the RSS 1.0 Content Module.
creator_of - An Item that the User is a creator of.
Links to a User account that an Item was created by. Being a creator of an Item is not a Role.
email_sha1 - An electronic mail address of the User, encoded using SHA1.
feed - A feed (e.g., RSS, Atom, etc.) pertaining to this resource (e.g., for a Forum, Site, User, etc.).
function_of - A User who has this Role.
has_administrator - A User who is an administrator of this Site.
has_container - The Container to which this Item belongs.
has_creator - This is the User who made this Item..
has_function - A Role that this User has.
Links a User to a Role that this User has (it is a function of this User). Users can have many different Roles in different Forums.
has_host - The Site that hosts this Forum.
has_member - A User who is a member of this Usergroup.
has_moderator - A User who is a moderator of this Forum.
has_modifier - A User who modified this Item.
This property links an Item or Post to the User who modified it (e.g., contributed new content to it after its creation). Being a modifier of a Post is not a Role.
has_owner - A User that this Container is owned by.
has_parent - A Container or Forum that this Container or Forum is a child of..
has_scope - A Forum that this Role applies to.
has_space - A data Space which this resource is a part of.
has_subscriber - A User who is subscribed to this Container.
has_usergroup - Points to a Usergroup that has certain access to this Space.
host_of - A Forum that is hosted on this Site.
id - An identifier of a SIOC concept instance. For example, a user ID. Must be unique for instances of each type of SIOC concept within the same site.
ip_address - The IP address used when creating this Item. This can be associated with a creator. Some wiki articles list the IP addresses for the creator or modifiers when the usernames are absent..
links_to - Links extracted from hyperlinks within a SIOC concept, e.g., Post or Site.
member_of - A Usergroup that this User is a member of.
moderator_of - A Forum that User is a moderator of.
modifier_of - An Item that this User has modified..
next_by_date - Next Item or Post in a given Container sorted by date.
next_version - Links to the next revision of this Item or Post.
note - A note associated with this Post, for example, if it has been edited by a User.
num_replies - The number of replies that this Post has. Useful for when the reply structure is absent..
owner_of - A Container owned by a particular User, for example, a weblog or image gallery.
parent_of - A child Container or Forum that this Container or Forum is a parent of.
previous_by_date - Previous Item or Post in a given Container sorted by date.
previous_version - Links to a previous revision of this Item or Post.
reply_of - Links to an Item or Post which this Item or Post is a reply to.
scope_of - A Role that has a scope of this Forum.
sibling - A Post may have a sibling or a twin that exists in a different Forum, but the siblings may differ in some small way (for example, language, category, etc.). The sibling of this Post should be self-describing (that is, it should contain all available information).
A recent development in online discussion methods is an article or Post that appears in multiple blogs, or has been copied from one Forum to another relevant Forum. In SIOC, we can treat these copies of Posts as siblings of each other if we think of the Posts.
space_of - A resource which belongs to this data Space.
subscriber_of - A Container that a User is subscribed to..
usergroup_of - A Space that the Usergroup has access to..
May be used to represent topics or tags defined on a community site. The sioc:topic property can be used to link an Item or Post to a skos:Concept.
Can be used for keywords describing the subject of an Item or Post. See also: sioc:topic.
Specifies the title of a resource. Usually used to describe the title of an Item or Post.
Details the date and time when a resource was created. Usually used as a property of an Item or Post.
A resource that is a part of this subject. Usually used from the domain of a Post or Community.
A resource that the subject is a part of. Usually used with the range of a Post or Community.
Details the date and time when a resource was modified. Usually used as a property of an Item or Post.
Used to link a foaf:Person to a sioc:User. See also: sioc:account_of.
Used to describe the encoded content of a Post, contained in CDATA areas.
These are the ontology namespaces referenced:. | http://www.w3.org/Submission/2007/SUBM-sioc-spec-20070612/ | crawl-002 | refinedweb | 1,334 | 59.6 |
Libec uses roles to implement a hierachal set of granular permissions attached to a certificate. This is intended to allow issuing certificates with various embedded permissions, which can be verified up to a trust root without needing to query any third party.
Roles are defined as a set of period-delimited strings, where each string is considered to be a child of the string to its left. The rightmost string may optionally be '*' - this is a wildcard match.
As an example, the role of
com.example.myPond.goFishing might allow the user to go fishing in the pond, but not feed the fish, or stir the water. In contrast, the role
com.example.myPond.* would allow the user to do anything to the pond, or anything subordinate to it (i.e. the user is also considered to have
com.example.myPond.lilyPad.locateFrog).
In order for a role to be considered valid, the certificate's signer must have either the same role, or a parent / wildcard role which matches it, defined as a grant. In order for a grant to be considered valid, the certificate's signer must have the same grant, or a parent / wildcard grant which matches it, defined as a grant.
This means that a certificate may also delegate granting authority for any role it is capable of granting.
A certificate with the flag
EC_CERT_TRUSTED may define any role or grant, with no higher authority required to validate it. Such roles and grants must be explicitly defined in the certificate; by default a certificate defines nothing at all.
ec_record_t *ec_role_add(ec_cert_t *c, char *role);
Add a role to a certificate. Returns the new role record on success, NULL otherwise. The created record will be automatically freed when the certificate is destroyed.
#include <ec.h> ... ec_record_t *r = ec_role_add(c, "com.example.myPond.goFishing"); if(r == NULL) { //failed to add role } ...
ec_record_t *ec_role_grant(ec_cert_t *c, char *role);
Allow a certificate to grant a role. Returns the new grant record on success, NULL otherwise. The created record will be automatically freed when the certificate is destroyed.
#include <ec.h> ... ec_record_t *g = ec_role_grant(c, "com.example.myPond.*"); if(g == NULL) { //failed to add grant for role } ...
ec_err_t ec_role_has(ec_cert_t *c, char *role);
Check whether a certificate holds the given role. Returns EC_OK on success, or a nonzero error code otherwise.
If the certificate holds a matching wildcard role, this is considered sufficient (e.g. if a certificate holds
com.example.*, it is also considered to hold
com.example.myPond.goFishing).
#include <ec.h> ... if(ec_role_has(c, "com.example.myPond") == EC_OK) { //certificate holds the com.example.myPond permission } ...
ec_err_t ec_role_has_grant(ec_cert_t *c, char *role);
Check whether a certificate is allowed to grant the given role. This does not imply that the certificate also holds this role; grants and roles may be held independantly.
If the certificate holds a matching wildcard grant, this is considered sufficient (e.g. if a certificate defines a grant for
com.example.*, it may also grant
com.example.myPond.goFishing).
#include <ec.h> ... if(ec_role_has_grant(c, "com.example.myPond.goFishing") == EC_OK) { //certificate may grant com.example.myPond.goFishing } ... | https://libec.erayd.net/chapters/roles_&_grants.html | CC-MAIN-2019-09 | refinedweb | 520 | 58.38 |
Arguments for detaching a basic block.
After a basic block is detached from the CFG/AUM, or after a placeholder is removed from the CFG, these arguments are passed to the callback. If a basic block was detached then the bblock member will be non-null, otherwise the arguments indicate that a placeholder was removed from the CFG. This callback is invoked after the changes have been made to the CFG/AUM. The partitioner is passed as a const pointer because the callback should not modify the CFG/AUM; this callback may represent only one step in a larger sequence, and modifying the CFG/AUM could confuse things.
Definition at line 262 of file ControlFlowGraph.h.
#include <ControlFlowGraph.h>
Partitioner in which change occurred.
Definition at line 263 of file ControlFlowGraph.h.
Starting address for basic block or placeholder.
Definition at line 264 of file ControlFlowGraph.h.
Optional basic block; otherwise a placeholder operation.
Definition at line 265 of file ControlFlowGraph.h. | http://rosecompiler.org/ROSE_HTML_Reference/structRose_1_1BinaryAnalysis_1_1Partitioner2_1_1CfgAdjustmentCallback_1_1DetachedBasicBlock.html | CC-MAIN-2017-47 | refinedweb | 162 | 58.58 |
Opened 10 years ago
Closed 9 years ago
#9972 closed enhancement (fixed)
Add fan morphisms
Description (last modified by )
This ticket adds a module for fan morphisms - morphisms between lattices with specified fans in the domain and codomain, which are compatible with this morphism. Compatibility check and automatic construction of the codomain fan or refinement of the domain fan are implemented.
Patch order (applies cleanly to sage-4.6.rc0):
- trac_9972_add_cone_embedding.patch
- trac_9972_improve_element_constructors.patch
- trac_9972_remove_enhanced_cones_and_fans.patch
- trac_9972_add_fan_morphisms.patch
- trac_9972_fix_fan_warning.patch
See #9604 for dependencies.
Attachments (8)
Change History (74)
Changed 10 years ago by
comment:1 Changed 10 years ago by
- Cc vbraun added
- Status changed from new to needs_info
comment:2 Changed 10 years ago by
- Reviewers set to Volker Braun
I don't have any good suggestion for how to improve
_repr_, we can always leave that for later.
I'll rewrite the Polyhedron constructor in cython one of these days, that should fix the speed issues. Though its a good idea to minimize the number of intersections computed :)
The current version looks good for an initial shot at toric morphisms. Are you still changing things around or should I officially review it?
comment:3 Changed 10 years ago by
- Status changed from needs_info to needs_review
I don't plan on changing these functions further, so this ticket is ready for review!
comment:4 Changed 9 years ago by
I like the functionality, but I'm confused about the name. Is this supposed to be a
ToricLatticeMorphism or a
FanMorphism? I am thinking that it would be good to split those apart, and perhaps make the latter inherit from the former.
ToricLatticeMorphism.make_compatible_with(fan) doesn't make the morphism compatible with the fan, it is the other way round. So it should be either
fan.make_compatible_with(toric_morphism) or, say,
ToricLatticeMorphism.subdivide_domain(domain_fan,codomain_fan). Or see below.
Another functionality that I would like to have is to figure out the image fan from the lattice morphism and the domain. How about the following proposal:
- separate
ToricLatticeMorphismand
FanMorphism.
FanMorphism(lattice_hom, domain=Fan, codomain=Fan)constructs the fan morphism. If
lattice_homis a matrix the corresponding
ToricLatticeMorphismis constructed automatically. Raises
ValueErrorif the fans are not compatible.
FanMorphism(lattice_hom, domain=Fan)constructs the image fan and uses it as codomain. Raise
ValueErrorif not possible.
FanMorphism(lattice_hom, domain=Fan, codomain=Fan, subdivide=True)will subdivide the domain fan as necessary.
Let me know what you think & I'd be happy to help, of course!
comment:5 follow-up: ↓ 6 Changed 9 years ago by
- Status changed from needs_review to needs_info
I am now putting finishing touches on plotting, but then return back to morphisms.
After actually working on morphisms a bit, I had second thoughts about class organization. In particular, I didn't see how
FanMorphism can fit nicely into Sage and I didn't even understand what it is mathematically. A fan is a collection of cones with certain restrictions, right? Then a fan morphism should be a map between sets of cones, in our case finite. But that's not what we want, we rather want a morphism between supports of fans, which is given on points of the space by a linear map. Put it another way, we want a
ToricLatticeMorphism restricted to the support of the domain. During construction of such a morphism we have to check that everything is compatible, which can be quite expensive. During any application of this morphism to a point, we should check that this point is in the domain and this is also non-trivial (the only thing that will work for general fans is checking this point against every generating cone). So if we do have a special class for
FanMorphism, how about this definition: "it is a morphism between toric lattices with distinguished fans in the domain and codomain which are compatible with this morphism." I.e. the domain is still a whole toric lattice and there is no need for complicated checks. One can write then
sage: phi = FanMorphism(lattice_morphism, F1, F2) sage: phi.image_cone(cone_of_F1) <some cone of F2> sage: phi.preimage_cones(cone_of_F2) <a tuple of cones of F1>
Going further, I don't think anymore that we should derive special fans for domain/codomain of morphisms between toric varieties themselves, let's call this class
ToricMorphism. The problem I have is that a single variety can be a (co)domain of different morphisms. (I definitely need such functionality.) This can make it very inconvenient to work with divisors and classes because they will get confused which parent do they belong to, we will have to work with coercion, and the code may need to look something like this:
sage: fan = ... sage: X = ToricVariety(fan) sage: fan = X.fan() # This is already a bit strange... sage: phi = ToricMorphism(matrix(...), X, Y) sage: psi = ToricMorphism(matrix(...), Z, X) sage: X_fan_codomain = psi.codomain().fan() # These two lines are plain confusing sage: X_fan_domain = phi.domain().fan() sage: phi.domain().rational_divisor_group() == psi.codomain().rational_divisor_group() ???
All four fans above are mathematically the same, the only difference is what kind of extra functionality do they get. But they will be different Python objects and associated toric varieties will be also different objects for no apparent reason, i.e. how these reasons can be explained to a user rather than a developer?
So I think that either morphisms derive their own fans for domain/codomain and use them internally without actually changing the varieties they were created for (i.e.
phi.domain().fan() is the same as
X.fan(), but
phi.domain_fan() can be something specialized), or they store this information in some other way and return (pre)images taking cones of usual fans as arguments.
I recall that we already had a similar argument in the beginning, whether or not we need any kind of specialized cones, which provide clean access to new features. I just checked how many new features got added:
- TWO methods for
Cone_of_fan:
star_generatorsand
star_generator_indices. (There were actually many new methods here originally, but others migrated to plain cones.) Still, I think that this class is justified, because these are very natural operations to perform on cones that belong to a fan. Also, a cone cannot quite belong to several fans in the sense that its internal data structures are severely tied to a fan and it is very important performance-wise. (E.g. intersection of cones of the same fan is incomparably easier/faster than for arbitrary ones, and this will remain true even when polyhedra are made fast.)
- ONE method for
Cone_of_toric_variety:
cohomology_class. Here I feel less convinced that it is necessary,
cone.cohomology_class()does not feel more natural to me than
X.cohomology_class(cone). However, I think there are more methods added here by patches which are not applied in my queue and something else may come up during later development. If not, we should probably reconsider this class, because it would be nice to have
sage: ToricVariety(fan).fan() is fan True
- Hypothetical cones of (co)domain would each add one more method, but make it difficult/inconvenient to deal with multiple morphisms, while the whole point of making new classes should be making life easier...
For this ticket I propose the following:
- Rename
make_compatible_withto
subdivide_domain.
- Add to
ToricLatticeMorphisma method like
image_fan(domain_fan)to construct the "natural" fan in the codomain, as you have suggested.
-).
- Cones of fans also get
image_coneand
preimage_conesmethods that take as an argument a
FanMorphismwith appropriate fans.
A follow-up ticket will add
ToricMorphism for arbitrary scheme morphisms between toric varieties and
ToricEquivariantMorphism for those coming from
FanMorphisms.
Let me know what you think!
comment:6 in reply to: ↑ 5 ; follow-up: ↓ 58 Changed 9 years ago by
"it is a morphism between toric lattices with distinguished fans in the domain and codomain which are compatible with this morphism."
Yes, that is the usual definition. No restriction on the support of the underlying lattice map.
On an unrelated note, I would call "point" = "0-dimensional torus orbit" = full-dimensional cone. A toric morphism maps points to potentially higher-dimensional torus orbits. Though I do understand that you meant lattice points.
I understand your issue about having multiple morphisms. But if the fans don't know about the toric morphism then they shouldn't know about the toric variety either, otherwise its confusing. So in principle I don't mind getting rid of the
Cone_of_toric_variety class. At least this solves the dilemma
cone_of_variety.cohomology_class() vs.
variety.cohomology_class(cone); since the cone doesn't know about the variety only the latter can work. But instead of adding a
cohomology_class method, I'd rather have the
CohomologyRing element constructor accept cones:
sage: HH = X.cohomology_ring() sage: HH( X.fan(2)[23] )
This pattern works already for the divisor group and Chow group:
sage: Div = X.divisor_group() sage: Div( X.fan(1)[0] ) # output should have been V(z0)?!? V(1-d cone of Rational polyhedral fan in 2-d lattice N) sage: A = X.Chow_group() sage: A( X.fan(1)[0] ) The Chow cycle (1, 0, 0)
For this ticket I propose the following:
- Rename
make_compatible_withto
subdivide_domain.
- Add to
ToricLatticeMorphisma method like
image_fan(domain_fan)to construct the "natural" fan in the codomain, as you have suggested.
If we can agree on a hierarchy
ToricLatticeMorphism ->
FanMorphism ->
ToricMorphism then
ToricLatticeMorphism shouldn't know about fans at all, if only to avoid circular imports. Similar to how
ToricLattice doesn't know about
Fan. So
make_compatible_with and
image_fan become special cases of the
FanMorphism constructor.
And instead of an
is_compatible_with method, why not have
FanMorphism(matrix,fan1,fan2) raise
ValueError, "Cone <3,4,5> is not contained in any image cone".
-).
Yes, sounds good!
- Cones of fans also get
image_coneand
preimage_conesmethods that take as an argument a
FanMorphismwith appropriate fans.
I'd rather have only
morphism(cone) (or
morphism.image(cone)) and
morphism.preimages(cone).
A follow-up ticket will add
ToricMorphismfor arbitrary scheme morphisms between toric varieties and
ToricEquivariantMorphismfor those coming from
FanMorphisms.
I agree, but can we use, say,
AlgebraicMorphism and
ToricMorphism? Toric should really always be replaceable with "torus-equivariant".
comment:7 Changed 9 years ago by
- Status changed from needs_info to needs_work
- Work issues set to switch to FanMorphism
OK, I wanted to avoid "compatibility check by exception" but I can live with it ;-) How about this:
- Move
cone.cohomology_classfunctionality to the element constructor of cohomology ring (I actually think this is the best and the most clear way there can be.) Can you perhaps make a patch for this change since it involves mostly your code?
- Get rid of
Cone_of_toric_variety(that's my part so I can do this). This raises a question whether
EnhancedCone/Fanshould survive at all. One option is to put
_in front so that they disappear from the documentation.
- Make
FanMorphismwork as you have described, including informative error message.
- See if there is then any point in having
ToricLatticeMorphismat all.
- No changes to cones, all (pre)images are computed/stored by morphisms. Which is probably also the cleanest way to do it.
I am also OK with
ToricMorphism for equivariant ones. I'll see what name should fit nicely with existing classes for affine/projective morphisms for a "non-toric morphism between toric varieties."
comment:8 Changed 9 years ago by
- I'll write a patch and post it here for the
cohomology ringand
divisor_group.
- I don't see why we need the
Enhanced*versions, then.
comment:9 Changed 9 years ago by
I'll also tighten
x in fan to only return
True if
x is actually a
cone. I'll add another method
fan.support_contains(point) for the other usage. That'll make it easier to ensure that "something" is a cone of the correct fan. Otherwise we have stuff like
0 in fan return True...
comment:10 Changed 9 years ago by
Sounds good!
comment:11 Changed 9 years ago by
Patch is attached. I removed 'cohomology_class' from
Cone_of_toric_variety to make sure that I got all occurrences, but tried to make as few other changes in that area as possible. Can you try to put my patch at the bottom of this ticket's queue?
I added an overriding method
Cone_of_fan.is_equivalent (see the "TODO" comment) that you should uncomment after removing the enhanced cones.
The
CohomologyRing._element_constructor_ now accepts cones as well. For the other issue, you should have complained that
divisor_group returns the non-toric divisor group and only
toric_divisor_group should accept cones ;-) The latter already works as it should.
comment:12 Changed 9 years ago by
Regarding
Cone_of_fan.is_equivalent - cones of fan are also constructed by intersection method and during computing the cone lattice. They can certainly be non-unique now and I don't think that we should try to make them unique. So I propose removal of the overriding method.
comment:13 Changed 9 years ago by
My thought was that it can be expensive to ensure that two cones of a fan are not the same, and comparing ambient ray indices would be faster. And you always have to go through all cones of the fan to test membership, so it will be called often. I thought about whether that is a problem during the construction, but then I found it easiest to leave that question up to you ;-)
How about constructing
Cone_of_fan always though a factory function that tests (via
Cone.is_equivalent) if the cone is already in the fan. That would be simple to implement and we can then rely on uniqueness of the
Cone_of_fan.
comment:14 Changed 9 years ago by
What membership exactly do you want to test?
The assumption is that there are no invalid cones of the fan, so if you want to check if a cone is in the fan, it is enough to check if the ambient fan of the cone is the fan in question. For checking equivalence of two cones of the same fan it is enough indeed to check that their ambient ray indices are the same, as tuples, since they are always assumed to be sorted and when they are not - it is a bug.
I don't mind adding uniqueness of cones of fan or even cones themselves for that matter (on a separate ticket, perhaps?), I just don't quite understand why do you want it. The advantages that I see are
- memory saving;
- cached data sharing;
and disadvantages
- more complicated code for construction;
- longer time for construction.
Do you have something else in mind?
comment:15 Changed 9 years ago by
Some more thoughts:
- When there is an attempt to check if a point is in the fan, should we try to interpret this point as a ray, i.e. 1-d cone?
- The last line of
_containsshould be replaced with the original version:
try: return data.is_equivalent(self.cone_containing(data.rays())) except ValueError: # No cone containing all these rays return FalseFor example, if you take a fan which is a subdivision of a cone, then this cone will trickle down to this block, but
cone_containingwill raise an exception. There probably should be a test for such a situation (although I certainly don't claim that all other places test all possible exceptions...)
Fan.containsshould no longer accept many arguments after your change - let's replace
*argwith
cone.
- Typo: "or a something" should be without "a".
- Typo: "class associated to cones" should be "classes".
Otherwise looks great: the new approach allows users to create their own shortcuts and write
HH(cone) instead of
cone.cohomology_class(). While I like names exposed in Sage to be descriptive, I certainly don't want to force users do it in their code ;-)
I will base other patches of the ticket on top of this one.
comment:16 Changed 9 years ago by
A fan is a collection of cones. A point is never in a fan, as it is not a cone. I don't think we should make cones unique in general, only cones of a fan. You can have arbitrarily many cones (memory permitting), but the cones of a fan are a finite set; To me it then feels right to have the
Cone_of_fan be unique. In the current implementation they actually are unique after the fan is constructed. So it ends up being a minor change to guarantee that they are always unique, I think.
comment:17 Changed 9 years ago by
One potential danger is that
cone.ambient_ray_indices() is meaningless if
cone is only mathematically equivalent to a fan, but not a
Cone_of_fan. I've added a new method
RationalPolyhedralFan.get_cone(cone) that finds the
Cone_of_fan corresponding to the cone, and I tried to make sure it gets called wherever we accept arbitrary cones from the user.
But its a potential pitfall to watch out for. We could always insist on the user passing only
Cone_of_fan, but that seems to be too unwieldy for the user. I would suggest that we move the
ambient_ray_indices() accessor method to
Cone_of_fan and convert the code in
cone.py,
fan.py to use the data member
_ambient_ray_indices instead. All higher level functions then shall only use
cone.ambient_ray_indices(), converting a generic cone to a
Cone_of_fan via
fan.get_cone(cone) if necessary.
Changed 9 years ago by
Updated patch
comment:18 Changed 9 years ago by
In what sense
ambient_ray_indices is dangerous? If ambient structures of two cones are different, there is no point to look at these attributes at all. Otherwise they are the same if and only if cones are the same.
I am definitely against moving
ambient_ray_indices to
Cone_of_fan because the point in having it is uniform treatment of faces of cones and cones of fans, which are pretty much the same things. In fact, even
star_generators make sense for faces of a cone in the sense of containing facets, it just should not be called that name. So currently the main reason for a
Cone_of_fan to exist is that terminology for faces of cones and cones of fans is a bit different. I think that it should stay this way as much as possible, so that all cones behave the same.
I am also against new containment check - cones are equal if they have the same rays in the same order and equivalent if they define the same set of points. If the same cone happened to belong to different fans and so has two objects representing it, it does not change anything. We can check if cones belong to the same ambient structure for code optimization, but the output should be the same. Note that in this case this check actually can be done in generic cones and there is no need to override the method for
Cone_of_fan.
I still don't understand exactly what are you trying to achieve in general and in particular why do we need
get_cone method. I agree that
fan.contains should return
True only for (some) cones and not for anything else, because a fan is a collection of cones. I agree that it may be good to have uniqueness of
Cone_of_fan but I don't see any reasons for doing this except for some performance gain and it is not clear how significant it can be. It also seems to me that it makes more sense to make all cones unique based on the ordered rays and the ambient structure, because essentially this is how cones of fans can be made unique.
comment:19 follow-up: ↓ 20 Changed 9 years ago by
So for
is_equivalent optimization I propose inserting
if self == other: return True if self.ambient() == other.ambient(): return self.ambient_ray_indices() == other.ambient_ray_indices()
in the beginning of
Cone.is_equivalent. Maybe with
== replaced with
is in the
if statements - both variants will work correctly, the question is how many simple but potentially non-informative checks we want to perform before using a generic algorithm.
comment:20 in reply to: ↑ 19 Changed 9 years ago by
In what sense ambient_ray_indices is dangerous? If ambient structures of two cones are different, there is no point to look at these attributes at all.
Thats of course true, but we got that wrong in the
ToricDivisor constructor (no check that the ambient was the same) and you didn't catch it ;-)
I'm just trying to explore options that make it impossible to make this mistake in the future. One more idea would be to force the user to pass the ambient to
ambient_ray_indices(),
sage: fan1 = ... sage: fan2 = ... sage: fan1.generating_cone(2).ambient_ray_indices(fan1) # fast sage: fan1.generating_cone(2).ambient_ray_indices(fan2) # slower
This would also get rid of the need for
get_cone()
The point of the
get_cone method is to avoid this extra branch whenever the user passes a cone that is not necessarilly within the same ambient:
if is_Cone(c): if c.ambient()==fan indices = c.ambient_ray_indices() else: try: indices = tuple(sorted([ fan.rays().index(r) for r in c.rays() ])) except ValueError: ...
and instead just write
if is_Cone(c): indices = fan.get_cone(c).ambient_ray_indices()
comment:21 Changed 9 years ago by
Aha, now I see! I really don't like the idea of forced arguments to
ambiet_ray_indices and
get_cone seems to be a confusing name. How about this:
- Implement
get_conefunctionality using
__call__for fans and cones, i.e. you will have to write
fan(c).ambient_ray_indices()to make sure that indices are correct. The same goes for other relative things like face walking methonds - in all cases it is assumed that your cone knows where does it sit. This goes very well with the concept of fans being collections of cones - you "convert" a certain cone to an element of this collection. For cones it is not as transparent but still makes perfect sense, IMHO.
- In principle, I don't terribly mind adding *optional* argument so that one can write
c.ambient_ray_indices(fan)and internally it will be translated to
fan(self).ambient_ray_indices(). However, I then want to have it for all functions where it does make sense, it will add the same piece of code and documentation to all of them, and in the user code it does not lead to any significant clarification or space saving. So I'd rather not add such functionality and you don't usually like having two different ways to do the same thing ;-)
- Leave equivalence and containment checks mathematical, i.e. it is possible to get
Truefor a check that a cone of one fan is "in" another fan. Those who want to check if it is an actual element of the collection should use
cone.ambient() is fan.
If you agree with these proposals, then I can start implementing them and uniqueness of cones (and then fans too for uniformity, I guess) working on top of your patch.
comment:22 Changed 9 years ago by
Adding the functionality to
__call__ still does not enforce that
ambient_ray_indices is used correctly. In particular, it would not have caught the bug in
ToricDivisor.
How about we remove the
ambient_ray_indices() method from cones altogether and replace it with
fan.ray_indices(cone).
I agree with 3.
I don't think that the
Cone_of_fan uniqueness is particularly urgent, we can come back to that later.
comment:23 Changed 9 years ago by
It is not quite Python style to try to eliminate any possibility of user mistakes by making sure that everything is called properly in proper places. The assumption is that users know what they are doing ;-) These methods assume that your cone is in some fixed fan:
adjacent
ambient_ray_indices
face_lattice(indirectly - since elements will be cones of the same fan)
faces(indirectly)
facet_of
facets(indirectly)
star_generator_indices
star_generators
All of these methods are pretty important and convenient, in particular, in many cases
ambient_ray_indices is more useful than
rays or
ray_matrix since it allows you to see clearly how different cones are related to each other. It would be quite annoying if for using all these methods it was necessary to mention fan explicitly.
Yes, there are bugs that appear because users make wrong assumptions, but I don't think that it is a valid reason to require users to always explicitly state all of their assumptions so that each function can check that they are correct and when possible fix them.
In addition, as you pointed out in a sample code above, passing ambient explicitly will mean that it always works correctly, but sometimes fast and sometimes slow. This sometimes slow can be sometimes significantly slow and the reason for keeping track of these ray indices and ambients is precisely code optimization, otherwise all cones could store only their rays and all fans - only their cones.
So I still think that if you have a cone and it is important that things about this cone are computed relevant to a certain fan, you are responsible for making sure that this cone "sits" in this fan and if not - trying to convert it there. I see a point in making this conversion easy, but not in making it mandatory every time you need it. Compare this
sage: indices = fan.ray_indices(cone) sage: supercones = fan.cones_with_facet(cone)
and this
sage: indices = fan(cone).ambient_ray_incides() sage: supercones = fan(cone).facet_of()
I think in the second case it is quite clear that
sage: cone = fan(cone) sage: indices = cone.ambient_ray_incides() sage: supercones = cone.facet_of()
is likely to be faster, plus it is a bit easier to read and there is only one place where an exception can occur due to
cone being incompatible with
fan at all.
I also think that using cones from a wrong fan is likely to be exposed very quickly in actual work due to index-out-of range exceptions, so given several month of working with current interface, I really don't want it to change...
comment:24 Changed 9 years ago by
Well
face_lattice,
faces,
facet_of, and
facets should return mathematically equivalent results if the cone is in the wrong ambient, so they don't count. For
adjacent and
star_generators I would tend to let your argument pass that the output would be so wildly wrong that it is immediately obvious. But
ambient_ray_indices of a
Cone always returns valid index ranges (namely,
range(0,cone.nrays())).
So I do maintain that this method is particularly dangerous. Note that I don't want to get rid of the
_ambient_ray_indices data member, just
def ray_indices(self, cone): if self is cone.ambient(): return cone._ambient_ray_indices # slow fallback
comment:25 Changed 9 years ago by
Well, if you just want to add
fan.ray_indices(cone) method (which can be potentially slow), I am OK with it, although I'd rather not ;-) But I really want to keep
cone.ambient_ray_indices() as is without any arguments, especially forced ones.
And I still think that the best way is to explicitly convert input cones to cones of a particular fan and then freely use any fan-dependent methods. Given the current size of toric code it does not seem to me that we have any particularly dangerous methods (the most dangerous are
check=False and
normalize=False options, which are described as dangerous). So maybe instead of changing code we can put a warning in the documentation that this method may be dangerous and one should ensure that
cone.ambient() is what it should be before calling this method. It seems to me that it is quite difficult to start using
ambient_ray_indices without reading its documentation at least once.
comment:26 Changed 9 years ago by
I agree that the best way is to first convert the input cone to the fan and then work with that.
But 1) its easy to accidentally omit that step and 2) it is very hard to detect that you got it wrong. In my book, thats a very dangerous API design. I don't think a warning in the documentation is sufficient here, after all, we should have known better yet we fell into that trap with
ToricDivisor.
At the very least
ambient_ray_indices() should spit out a warning if
self==self.ambient() [what ARE you trying to do?]. But the best option seems to me to be the
fan.ray_indices(cone) method. If you wrote your code correctly there won't be any slowdown, and if you forgot to convert the cone to a cone of the fan then you'll still get the right answer. Ideally we'd then make
ambient_ray_indices private not use it outside of
cone.py/
fan.py (and fan morphisms).
comment:27 Changed 9 years ago by
After some more contemplating and checking
$ grep "ambient_ray_indices()" sage/geometry/* |wc 30 179 2579 $ grep "ambient_ray_indices()" sage/schemes/generic/* |wc 13 85 1383
on Sage-4.6.alpha3, I think that you may be a bit overestimating the danger. We have used this function in 43 places and so far everything is working quite well. The "mistake" you ran into got quickly caught even before the ticket got ready for a review. Also, why did you get this mistake? Because you took a code from a place where the ambient definitely was set correctly and moved it to the place where potentially other cones maybe passed in. Finally, you actually have not done any mistake, you just left a possibility for a user to make one by passing a wrong cone. (In this case, I think, the mistake would be realized quite fast even if the code did not throw up an exception.)
Adding a warning for calling
ambient_ray_indices on the ambient itself may not work because it may be called internally in some cycles. (I didn't check but it is quite likely.) Besides, nothing is wrong with such a call. I also strongly believe that this method must be open/documented/public, because if we have used it in 43 places, it is quite important for development and therefore users who build their own structures on top of stuff shipped in Sage are likely to find it convenient in their code. I personally use it all the time when working with varieties and sometimes even regret that these indices are not the default output for cones, so when I have a list of them I need to write a cycle calling this method.
A few months ago I already suggested switching to notation like
fan.ray_indices(cone) etc. but you opposed and pointed out that it is not in the spirit of OOP. Now I actually completely agree and think that it is very fortunate that we have not done so then. I even still see some benefits of having special classes for cones of toric varieties and morphisms (in particular, this problem would not occur ;-)) but for these cases the disadvantages overweight.
Have I convinced you yet or should I abandon any hope?-)
comment:28 Changed 9 years ago by
I had counted the occurrences of
ambient_ray_lattice as well, but my conclusion was that it is seldomly used outside of
/sage/geometry/cone,fan.py. Replacing the 13 occurrences wouldn't be hard. The fact that copy/pasting "correct" code turns it so easily into "incorrect" code is precisely what I find disturbing. Also, the mistake in
ToricDivisor was not caught easily. It is in the current release. And saying that its the user's fault ("You are holding it wrong" :-) is not helpful.
comment:29 Changed 9 years ago by
I meant that mistake would be realized quite fast once the code was actually used. I didn't actually use code for toric divisors yet, especially with cones from another fan. But anyway - this is all hypothetical and I may be very wrong.
My main point is not that it is difficult to replace 13 uses of
ambient_ray_indices, but that it is a very convenient and natural function which I personally use both when developing new code and when actually using Sage interactively. I would rather not pass any arguments to it, because it is annoying for interaction and in functions it may lead to code like
cone.ambient_ray_indices(cone.ambient())
or
cone.ambient().ray_indices(cone)
which is plain confusing.
The name of the function clearly indicates that there is something ambient and cone must be aware of it, since no arguments were passed. Therefore, when using this function, the user should be sure that this something ambient is what it is supposed to be. In fact, it may be more clear for new outside users than for us, since we are used to writing code inside the class where in many cases this requirement is definitely satisfied. (How can
self have a wrong
self.ambient()???) And after all this discussion I will probably remember very well for a long time to make an extra check that this function is used after making sure that ambient is what it is assumed to be.
If you want to add extra functionality like
fan.ray_indices(cone) which will be safer to use and then use this one when you want - I am fine with it. But if you insist on getting rid of
cone.ambient_ray_indices() in its present form/namespace, we need to invite more people from sage-devel for opinions...
comment:30 Changed 9 years ago by
By the way, in schemes
ambient_ray_indices is used several times in the examples and it was also used in Hal's examples for their book. It is very-very natural and should be as easy to access as possible...
comment:31 Changed 9 years ago by
ambient_ray_indices is not natural! Natural operations on cones should work without referring to a particular enumeration of the rays. Think
cone1.intesection(cone2) vs.
set(cone1.ambient_ray_indices()).intersection(set(cone2.ambient_ray_indices())). Of course sometimes you can't avoid it, definitely not in the internal implementations, but higher-level code (like the toric varieties) could easily move away from it. Case in point is that it is used only 9 times (the other 4 are in doctests).
And this looks horrible
cone.ambient().ray_indices(cone)
precisely because it is bad! It will break things! Its like a big, flashing sign: Do not write this code! You really want to use
self.fan().ray_indices(cone)!
I don't mind the availability of
cone.ambient_ray_indices() so much if you use it on the command line; In that case you know what your ambient() is. But I think we should ban its use from the toric varieties code. Then we algorithmically audit it for correctness by grepping
sage/schemes/generic for
ambient_ray_indices. How does that sound?
comment:32 Changed 9 years ago by
Well, instead of "very-very" let me say that
ambient_ray_incices are as natural as coordinates. Coordinates may be not very convenient for definitions, since you need to make sure that the choice of the coordinate system does not matter. They are in many cases not mathematically natural in the sense that there are many different choices without any preference. At the same time they are quite useful both for proofs and for working with concrete examples.
Now, choosing between
cone.ambient_ray_indices() and
fan.ray_indices(cone) is like choosing between always mentioning which coordinate system you are using whenever you use coordinates and having some coordinate system "understood from context." Of course, in the second case you are risking to encounter a situation when it was not quite clear and it leads to mistakes since coordinates in one system may have very little to do with coordinates in another one. People do such mistakes from time to time, that's just life. But it is very convenient to have this default and I still vote for it despite the mistake that you have discovered. (By the way - how did you find it?)
Leaving a function but banning it for internal use is a bit silly - the way to enforce it is to scan for occurrences of this function and then replace it with a "safe version". If you do it, then the second step can be making sure that it is used safely and correctly.
comment:33 Changed 9 years ago by
In an effort to be less stubborn, here is what I think I should do if we cannot live with the existing interface:
- Keep internal fields
_ambientand
_ambient_ray_indiceswhich can be related either to a cone or a fan (from the coding point of view these situations are similar and it is very convenient to treat them together - I know for sure since my initial version didn't do it).
- Remove
ambient()and
ambient_ray_indices()methods. In the code we still can use attributes since they always must be set in the constructor. Also remove
adjacentand
facet_ofmethods from cones and
star_generator_indicesand
star_generatorsfrom cones of fan. The class
Cone_of_fanbecomes completely unnecessary and should be discarded.
- Add methods corresponding to the above ones with respect to an explicit ambient (of course, there is no analog for
ambientitself anymore):
sage: ambient_cone.ray_indices(cone) sage: ambient_cone.adjacent_faces(cone) sage: ambient_cone.faces_with_facet(cone) sage: fan.ray_indices(cone) sage: fan.adjacent_cones(cone) sage: fan.cones_with_facet(cone) sage: fan.star_generators(cone) sage: fan.star_generator_indices(cone)Cones still will cache all this stuff for their own
_ambientin a hope that it will be useful later and that's how usually they will be called.
But I still need some voting before performing such a change.
I still think that it is convenient to have a default for "the thing before dot" above. This opinion is based on actually working with such structures, so maybe I have bad habits, but that's what I prefer. I also still think that it is natural in mathematics. E.g. Fulton on p.52 defines Star(tau), where it is assumed that tau is a cone in some fan. In a lot of papers using reflexive polytopes people introduce some notation for dual faces. I have never seen such a notation mentioning the ambient polytope, yet the notion of a dual face does not make sense without one. One can argue that a face is supposed to know its ambient, because it is a face and not just a standalone polytope. But with this line of thoughts I think toric divisors may require a cone of an appropriate fan as an input. Or we can agree that we will use any input cone and try to interpret it as a cone in this appropriate fan, in which case it is our responsibility to ensure that we do such an interpretation correctly. It is also common to fix some index set and then use it to index rays of the fan, their generators, corresponding homogeneous variables, and divisors. So I still think that these indices are relevant beyond the implementation guts of cones...
comment:34 follow-up: ↓ 35 Changed 9 years ago by
I don't want to rewrite the entire interface, and I think that having some implicit assumption about what the
ambient() is is usually fine. Also, 3.) is spectacularly ugly :-P As I said before, if the method call e.g. returns again a collection of cones then you'll immediately notice that you had the wrong
ambient(). The difference with
ambient_ray_indices is that its output will fail in much more subtle ways if the hidden assumption is wrong.
If you desperately want to keep
ambient_ray_indices(), how about we prefix any use in the toric varieties code with an assertion that makes it explicit. This would be yet another way to make the implicit assumption explicit and have it easily machine-verifiable.
On an unrelated note, I don't like
cone_of_fan = fan(cone), its too similar to
Fan([cone]). How about
Fan.cone_equivalent_to(cone), see also the already-existing similar method
Fan.cone_containing(cone).
comment:35 in reply to: ↑ 34 Changed 9 years ago by
3.) is spectacularly ugly :-P
I wholeheartedly agree ;-)
If you desperately want to keep
ambient_ray_indices(), how about we prefix any use in the toric varieties code with an assertion that makes it explicit. This would be yet another way to make the implicit assumption explicit and have it easily machine-verifiable.
I desperately want to keep it! I think that assertions are great - I don't use them much mostly because I didn't know about them until reviewing some of your code, but I am totally for them. However using them always before
ambient_ray_indices() is too harsh - I went through
cone.py, fan.py, toric_variety.py, fano_toric_variety.py and in all cases it was actually already clear that
cone._ambient is correct, e.g. because
cone was constructed in the same code. So I think that more generally we should use assertions for input parameters, which was the problem in this case.
On an unrelated note, I don't like
cone_of_fan = fan(cone), its too similar to
Fan([cone]). How about
Fan.cone_equivalent_to(cone), see also the already-existing similar method
Fan.cone_containing(cone).
fan(cone) is similar to
Fan([cone]) only if your fan is called
fan ;-) I was thinking of such format to mimic Element-Parent behaviour in Sage, but:
- currently we don't use this model for cones and fans;
- if we did, then cones must be both elements (of fans) and parents (of points), and this is not yet supported in Sage;
- I don't clearly see advantages of such a switch;
- I would like to have the same interface for "putting" cones into fans and faces into cones.
So the attached patch implements for cones and fans methods
embed(cone).
cone_equivalent_to seems unclear to me despite of being long, but I am not sure if
embed is much better - what do you think of it? Maybe
embed_cone/embed_face would be more clear?
I propose a bit different implementation of this method than in your patch. For fans it computes potential ray indices and then uses
cone_containing method. This should be faster than using
cone_containing directly on the rays and does not trigger cone lattice computation. For cones it will go through all faces of the relevant dimension, but instead of checking cone equivalence it still computes potential ray indices and then looks for a face with them. It will not work for non-strictly convex cones, where I also check for equivalence. (Although it will not work yet anyway - we cannot compute face lattice in such a case.) Documentation is mostly the same for fans and cones, but I tried to explain and illustrate clearly what the function does and why one should care.
I have also tweaked
Fan constructor a little bit - now rays of the fan generated by a singe cone will have the same order as this cone. Shuffling bugged me some time before and again now that I was writing doctests for embeddings. My general attitude to such situations is that users enter data in the way they like, so it is good to preserve it as much as possible.
comment:36 Changed 9 years ago by
P.S. I also optimized
is_equivalent for cones of the same ambient, as was suggested in.
comment:37 Changed 9 years ago by
Fixes for non-strictly convex cones:
cone -->
self for convexity check in
Cone.embed and extra convexity check in
Cone.is_equivalent for the case of common ambient (does not require extra computations for cones of fans, they are always directly set to be strictly convex).
comment:38 Changed 9 years ago by
I couldn't figure out whether your patch goes before or after mine, so I folded it into one and fixed everything.
I changed
# Optimization for fans generated by a single cone if len(cones) == 1: cone = cones[0] return RationalPolyhedralFan((tuple(range(cone.nrays())), ), cone.rays(), lattice, is_complete=lattice.dimension() > 0)
to
is_complete=(lattice.dimension()==0).
Now that we agree on this ;-) can you go ahead and remove the
Enhanced* versions of fans and cones? Then we can go on with fan morphisms...
comment:39 Changed 9 years ago by
Oops, sorry - I should have written that it was supposed to be the first. I unfolded the patches back so that it is clear who is writing/reviewing what and we don't need to seek the third person for the final review. I have updated my patch to fix the mistake that you caught. In your part I have removed
is_equivalent from
Cone_of_fan since this optimization is now performed by general cones. I have also removed extra parenthesis from
cone = fan().embed(x).
I also have one more issue with your patch which got lost above, regarding new containment check: cones are equal if they have the same rays in the same order and equivalent if they define the same set of points. If the same cone happened to sit in different fans (or cones, for that matter) and so has several different objects representing it, it does not change anything. We can check if cones belong to the same ambient structure for code optimization, but the output should be the same. Also, to me it feels perfectly natural to ask e.g. whether a cone of some fan belongs to a subdivision of this fan. So I think that
cone in fan should return
True if
cone is equivalent to any of the cones of
fan. What is the ambient structure of
cone and what is its ray order does not matter. If you really disagree with this, then I think that
cone in fan should return
True ONLY if
cone.ambient() is fan is
True. But I definitely prefer the first variant. Then one can write
sage: if cone in fan: sage: cone = fan.embed(cone) sage: Do something, say with cone.adjacent() sage: else: sage: Deal with it somehow.
comment:40 Changed 9 years ago by
Enhanced cones are removed, all tests pass. Current ticket queue: last three patches then the first one (which is going to be changed).
comment:41 Changed 9 years ago by
I agree with
cone in fan meaning that cone is equivalent to any cone of the fan. I updated my patch to reflect this.
comment:42 Changed 9 years ago by
Oops I had forgotten to refresh the patch. Correct version follows.
Changed 9 years ago by
Updated patch
comment:43 Changed 9 years ago by
I'll post an updated patch with
FanMorphism shortly.
Regarding
image_cone and
preimage_cones - I guess we want them to work for arbitrary cones of domain/codomain fans. In this case it is still clear what to return for
image_cone(cone) - the smallest cone of the codomain fan which contains the image. But what about
preimage_cones? Should we return all cones that are mapped to it, or only maximal ones? (We can also consider packing them into a fan, but in this case we loose connection with the domain fan, so I don't think that it is a good idea.)
I am also not sure what would be an efficient way to determine preimage cones. I am thinking about determining rays that are mapped into the given codomain cone and then scanning trough all domain cones to see which are generated by such rays only. An alternative approach, which is likely to be better for intensive work with such cones, is to compute full dictionary for image cones and then revert it to get preimage ones. Any suggestions?
comment:44 Changed 9 years ago by
- Description modified (diff)
- Summary changed from Add toric lattice morphisms to Add fan morphisms
- Work issues changed from switch to FanMorphism to implement preimage_cones
I am attaching a patch that does not compute
preimage_cones yet, but the rest is claimed to be ready for review/comments. I have removed classes for lattice morphisms themselves since they were not adding anything anymore. All of the old functionality is moved to
FanMorphism constructor as you have suggested above. Codomain fan can now be omitted and will be computed, if possible.
This feature has exposed a problem I mentioned (I think) when we were adding warnings to the
Fan constructor. The image fan is generated by images of cones of the original fan and these images may coincide or become non-maximal. As a result, one of the doctests fails due to the warning that some cones were discarded and users may see such a message as well, which will be quite confusing, I think. What should we do? Add a parameter to
Fan(..., warn=True) and set it to
False explicitly in the internal code?
Current queue:
- trac_9972_add_cone_embedding.patch
- trac_9972_improve_element_constructors.patch
- trac_9972_remove_enhanced_cones_and_fans.patch
- trac_9972_add_fan_morphisms.patch
comment:45 Changed 9 years ago by
- Status changed from needs_work to needs_info
- Work issues changed from implement preimage_cones to Warning from Fan constructor
OK,
preimage_cones can now be computed, everything is doctested, the last issue is with warnings.
comment:46 Changed 9 years ago by
- Reviewers changed from Volker Braun to Volker Braun, Andrey Novoseltsev
- Status changed from needs_info to needs_review
- Work issues Warning from Fan constructor deleted
After some more contemplating, I don't really see any other solution to the problem except for adding a parameter to suppress warnings. (Well, the other option is to remove the warning completely, but I have a feeling that it will not be accepted ;-)) So the last patch does exactly that, now all doctests pass smoothly and I propose merging this ticket.
For the record: I am giving positive review to the current "trac_9972_improve_element_constructors.patch" written by Volker Braun.
comment:47 Changed 9 years ago by
comment:48 Changed 9 years ago by
- Status changed from needs_review to needs_work
Looks good overall, but I think there is a bug in
image_cone():
sage: N1 = ToricLattice(1) sage: N2 = ToricLattice(2) sage: Hom21 = Hom(N2,N1) sage: pr = Hom21([N1.0,0]) sage: P1xP1=toric_varieties.P1xP1().fan() sage: f = FanMorphism(pr, P1xP1) sage: f.codomain_fan().generating_cones() (1-d cone of Rational polyhedral fan in 1-d lattice N, 1-d cone of Rational polyhedral fan in 1-d lattice N) sage: [ f.image_cone(c) for c in P1xP1.generating_cones() ] [0-d cone of Rational polyhedral fan in 1-d lattice N, 0-d cone of Rational polyhedral fan in 1-d lattice N, 0-d cone of Rational polyhedral fan in 1-d lattice N, 0-d cone of Rational polyhedral fan in 1-d lattice N]
comment:49 Changed 9 years ago by
- Status changed from needs_work to needs_review
comment:50 Changed 9 years ago by
comment:51 Changed 9 years ago by
- Status changed from needs_review to needs_work
Ok now
image_cone works but I can't compute the preimage:
sage: c = f.image_cone( Cone([(1,0),(0,1)]) ) sage: c 1-d cone of Rational polyhedral fan in 1-d lattice N sage: f.preimage_cones(c) --------------------------------------------------------------------------- TypeError Traceback (most recent call last) /home/vbraun/opt/sage-4.6/devel/sage-main/<ipython console> in <module>() /home/vbraun/Sage/sage/local/lib/python2.6/site-packages/sage/geometry/fan_morphism.pyc in preimage_cones(self, cone) 858 for dcone in dcones: 859 if (possible_rays.issuperset(dcone.ambient_ray_indices()) --> 860 and self.image_cone(dcone).is_face_of(cone)): 861 preimage_cones.append(dcone) 862 self._preimage_cones[cone] = preimage_cones /home/vbraun/Sage/sage/local/lib/python2.6/site-packages/sage/geometry/fan_morphism.pyc in image_cone(self, cone) 809 RISGIS = self._RISGIS() 810 CSGIS = set(reduce(operator.and_, --> 811 (RISGIS[i] for i in cone.ambient_ray_indices()))) 812 image_cone = codomain_fan.generating_cone(CSGIS.pop()) 813 for i in CSGIS: TypeError: reduce() of empty sequence with no initial value
comment:52 Changed 9 years ago by
- Status changed from needs_work to needs_review
Indeed, it does not work and moreover - the error was in
image_cone.
I have updated the main patch by adding a special branch for the image of the trivial cone and inserted your example into
preimage_cones docstring.
comment:53 Changed 9 years ago by
- Status changed from needs_review to needs_info
Why is
preimage_cones not the preimage of
image_cones? This confuses me:
sage: cone = f.image_cone( Cone([(1,0),(0,1)]) ) sage: [ cone is f.image_cone(c) for c in f.preimage_cones(cone) ] [True, True, True, False, False, False]
I understand that it is working as documented, but it seems like it would be more useful to actually return the preimage cones. In other words, take preimage cones in the sense of fan morphisms and not in the sense of convex sets.
If one is interested in all cones mapping geometrically into a given cone one could always construct the
Fan of the (fan morphism) preimage cones.
comment:54 Changed 9 years ago by
I am actually not quite sure I completely understand your comments, which makes me think that method names and definitions require some adjustments. How about the following:
cones_mapped_into(cone)returning all cones of the domain fan mapped into the given
coneof the codomain fan (this is what
preimage_conesdoes now). They do form a fan, but I am not sure that it is a good idea to construct it, since we will loose connection to the domain fan then.
cones_mapped_onto(cone)returning all cones of the domain fan whose
image_coneis equal to the given
coneof the codomain fan. Is this what you are talking about?
In both cases we may add extra parameter to allow returning only maximal cones with these properties. Or we can add a function to
fan module like
maximal_cones(cones) that will select the maximal ones. In the case of cones of the same fan the selection process is not terribly expensive.
As for
preimage_cones, I would say that
preimage_cone would make perfect sense if it was just the set-theoretic preimage of a given cone, which in general is not strictly-convex and is likely to be not a part of the domain fan. I actually compute these cones internally for generating cones of the codomain fan, but I am not sure they have much value to the end user.
Thoughts?
comment:55 Changed 9 years ago by
I think the geometric image/preimage of a cone is not particularly useful for toric morphisms. A fan morphism maps cones to cones, but does *not* define linear maps of the individual cones. One should think of it as a map from the poset of domain cones to the poset of codomain cones. Or, in geometric terms, maps of the poset of torus orbits to the poset of torus orbits.
The usual blow-up example
sage: c1 = Cone([(1,0),(1,1)]) sage: c2 = Cone([(1,1),(0,1)]) sage: domain_fan = Fan([c1, c2]) sage: codomain_fan = Fan([Cone([(1,0),(0,1)])]) sage: f = FanMorphism(identity_matrix(ZZ,2),domain_fan,codomain_fan) sage: ray = Cone([(1,1)]) sage: f.image_cone(ray) 2-d cone of Rational polyhedral fan in 2-d lattice N
means, geometrically, that the orbit closure
P^1 corresponding to the cone
ray (given by the relative star of
ray) is mapped to the point corresponding to the
f.image_cone(ray). Conversely, the preimage cones of
f.image_cone(ray) are
c1,
c2, and
ray. Geometrically, this means that the preimage of the point
f.image_cone(ray) consists of the torus orbits (north pole), (south pole),
C^* making up the fiber
P^1.
comment:56 Changed 9 years ago by
Why a fan morphism does not define linear maps on individual cones? A fan morphism is a linear map between lattices, so it can be restricted to any (compatible) pair of domain/codomain cones and still induce a morphism between corresponding affine varieties. Just specifying mapping of fans as finite sets of cones is not sufficient, e.g. identity morphism and multiplication by 2 are different but obviously would induce the same cone correspondence. So since the actual linear map does matter, it makes sense to use preimages under this linear map. I agree that full (potentially non-strictly convex) preimages are probably of little use, we need to restrict to the domain fan, but I don't see why something called "preimage" should ever exclude the origin. In the sense of orbit closures it does correspond to the whole variety, but in the sense of affine patches it corresponds to the torus itself which is a part of any toric variety of the given dimension and always gets mapped to itself.
So I still propose the switch to
cones_mapping_into for the current version and adding
cones_mapping_onto for the version that you want to have. No
preimage_cones at all since it is a confusing name in this context. On the level of toric varieties the corresponding methods may refer to orbits to make things clear, but here they are not present yet.
comment:57 Changed 9 years ago by
There is clearly important information in how the lattice spanned by a cone is mapped to a sublattice of the lattice spanned by the image cone. But that is just the restriction of the underlying lattice morphism, and not associated to just any cone of the domain. You never need the geometric image/preimage of a cone for toric geometry.
I would find it confusing if the
FanMorphism contained any other "map of cones" than fan morphisms. Since there is no ambiguity in "image cone", there is no question about "preimage cones" in that category. If you want
cones_mapping_into and
cones_mapping_onto for implementation purposes that is fine, but
image_cone and
preimage_cones are what they are and should be called like that.
comment:58 in reply to: ↑ 6 Changed 9 years ago by
- Status changed from needs_info to needs_work
I kind of complained before that I don't quite understand the term "fan morphism" ;-)
From the beginning of the ticket:
"it is a morphism between toric lattices with distinguished fans in the domain and codomain which are compatible with this morphism."
Yes, that is the usual definition. No restriction on the support of the underlying lattice map.
So, as I understand we have a category with fans being objects. A fan is a collection of cones, each cone is a set of points. All cones of the fan sit in the same lattice, so there is also "the lattice of the fan."
Morphisms in this category are morphisms between lattices associated to fans, which map cones into cones as sets of points. So in particular they do define linear maps between individual cones as well, and I got lost when you said that they don't.
But I guess I agree that when we talk about images/preimages of cones, we should refer to the induced map between cones as a map between two finite sets, since we do not have a nice correspondence between cones as sets of points. Let me think a little more about it and post an updated patch.
comment:59 Changed 9 years ago by
- Status changed from needs_work to needs_review
I am convinced, the updated patch does what you have suggested and now
preimage_tuple returns a tuple, as it was intended and documented. Thanks for the input!
comment:60 Changed 9 years ago by
Of course, I meant that
preimage_cones returns a tuple...
comment:61 Changed 9 years ago by
In the aforementioned blow-up example, I get now
sage: f.preimage_cones(f.image_cone(ray)) (2-d cone of Rational polyhedral fan in 2-d lattice N, 2-d cone of Rational polyhedral fan in 2-d lattice N) sage: filter(lambda c:f.image_cone(c).is_equivalent(f.image_cone(ray)), flatten(domain_fan.cones())) [1-d cone of Rational polyhedral fan in 2-d lattice N, 2-d cone of Rational polyhedral fan in 2-d lattice N, 2-d cone of Rational polyhedral fan in 2-d lattice N] sage: ray in f.preimage_cones(f.image_cone(ray)) # should be True False
The
preimage_cones misses the 1-d cone that maps to (in the fan morphism sense) the single 2-d cone of the codomain fan.
Changed 9 years ago by
preimage_cones implemented
comment:62 Changed 9 years ago by
Grrr... That was due to my optimization attempt without proper thinking. The new version uses the same cycle as the very first one (which was finding everything and even more), but requires equality of image cones. Should work now, the blow up example in the documentation is extended to include the preimage cones of the quadrant...
comment:63 Changed 9 years ago by
- Status changed from needs_review to positive_review
Works great now!
comment:64 Changed 9 years ago by
- Milestone changed from sage-4.6.1 to sage-4.6.2
comment:65 Changed 9 years ago by
For the tracbot:
Apply trac_9972_add_cone_embedding.patch, trac_9972_improve_element_constructors.patch, trac_9972_remove_enhanced_cones_and_fans.patch, trac_9972_add_fan_morphisms.patch, trac_9972_fix_fan_warning.patch
comment:66 Changed 9 years ago by
- Merged in set to sage-4.6.2.alpha0
- Resolution set to fixed
- Status changed from positive_review to closed
The patch is in principle ready, but while we are at it - do we want to make custom
_repr_for such morphisms? If yes, how should they be different from the standard?
Also, the speed is far from spectacular, but it is not easy to make it better until simple polyhedra work faster - currently most time is spend on constructing them for intersection purposes and I tried hard not to intersect more cones than necessary. | https://trac.sagemath.org/ticket/9972 | CC-MAIN-2020-16 | refinedweb | 10,179 | 62.38 |
03 August 2012 02:57 [Source: ICIS news]
MELBOURNE (ICIS)--?xml:namespace>
The company is unable to increase its etac output following a 26 July restart because of tight supply of acetic acid and ethanol, the official said.
Jiangsu Sopo operates three acetic acid plants with a combined capacity of 1.2m tonnes/year as well as a 500,000 tonne/year methanol unit at the same
While its acetic acid supply is integrated, it has to purchase co-feedstock ethanol from other producers.
“We are keeping our acetic acid output at a reduced level because of weak margins,” said the official.
“Supply of ethanol has also tightened because several plants have cut output in recent weeks [for the same reason of compressed margins],” he added.
Limited feedstock supply is expected to keep
Etac, which is mainly produced by the esterification of ethanol with acetic acid, is used in coating formulations like epoxies and vinyls, and in solvent applications for inks and | http://www.icis.com/Articles/2012/08/03/9583520/chinas-jiangsu-sopo-runs-zhenjiang-etac-plant-at-low.html | CC-MAIN-2015-22 | refinedweb | 162 | 57.61 |
11 July 2012 12:40 [Source: ICIS news]
LONDON (ICIS)--UBS has downgraded the share rating of specialty chemicals producer Croda and fellow UK-based catalyst maker Johnson Matthey to “neutral” from “buy”, the investment bank said on Wednesday.
The two ?xml:namespace>
“Other than valuation and stock outperformance we find nothing to criticise with Croda and very little with JM [Johnson Matthey],” UBS said.
The investment bank said that Croda remains a core holding, “in our view, and offers one of the most robust, if not the most robust, setup of any chemical company we follow.
“The key here is that almost 100% of Croda’s production is tailor-made and batched based rather than 24/7 continuous production and that it sells its products directly rather than selling via distributors, thereby capturing the distributor margin and gaining the ability to build long-term relationships with its customers,” UBS added.
The investment bank did revise down Johnson Matthey’s estimated earnings per share for fiscal 2013 and 2014 by 4% on the basis the group’s Precious Metal Products (PMP) segment will be hit by an expected fall in precious metal prices.
“With double digit growth projected for the catalyst business as well as Fine Chemicals, this [earnings] compression is entirely due to PMP, with reduced metal prices as well as less high margin autocatalyst scrap arriving in JM’s refining operations,” UBS said.
UBS cut Johnson Matthey’s share price target to £23.50 (€29.70) from £27, although Croda’s was lifted to £23 from £22.80.
At 11:19 GMT, Croda’s shares were trading at £22.55 on the London Stock Exchange, down 1.53% from the previous close, while Johnson Matthey’s shares at the same time were trading at £21.06, down 2.81% from the previous close.In related news, UBS has upgraded its share rating for Germany-based chemicals group LANXESS from “neutral” to “buy”, after the group’s earnings proved resilient despite a volatile market.
(€1 = £ | http://www.icis.com/Articles/2012/07/11/9577137/uk-chem-firms-share-ratings-cut-on-valuation-grounds-analyst.html | CC-MAIN-2015-06 | refinedweb | 336 | 57.5 |
Hi, I am completely new to using the Pololu servo controller and I need this for a competition, Basically I am building a robot for a search and rescue operation and I have two problems the most important one is finding the easiset way to control the 12-channel servo controller from a long distance assuming that some sort of RC method would be the easiest, the least important being getting two servo controllers to communicate with each other in a master slave configuration. any help will be very much appreciated
Hello.
The Maestro Servo Controllers have no way of measuring the RC pulses from your receiver. To control the Maestro from RC you would have to use some other piece of hardware to actually read the RC pulses and convert to something that the Maestro can use. One option would be the Pololu RC Switch with Digital Output.
Why do you want two 12-channel servo controllers in a master-slave configuration? You can accomplish that by writing an internal script on the master Maestro and using the serial_send_byte command to send commands to the other Maestro, but it might be easier to use one Mini Maestro 24-Channel USB Servo Controller.
–David
Correction, its two 18 channel controllers designed to control 28 servos and sensors, the design is rather complicated and ambitious, but if anything it’s a learning experience. The Idea behind the master slave configuration was that I could use one servo which would have the Sensors, RC receiver. I want the robot to be able to make use of that information as well as respond to remote control signals. Controller 1 would then send the relevant information to control the specified servos on controller 2, that is possible right ? I have literally. had little say in the design process otherwise I would not be struggling so much getting to grips with this equipment.
Yes, that should be possible.
–David
Hi David I was wondering if you could elaborate on the serial_send_byte command, I we have designed a custom controller using a long range am transmitter with an encoder, and a receiver with a decoder on the other end, If I send an input signal and want to move want to activate a servo on channel 8 of the slave device how would I do this and is it possible to test this command from the maestro control centre ?
To achieve that, you’ll have to learn about the Maestro scripting language and then you’ll need to figure out two things and combine them:
You’ll need to figure out how to get your Maestro script to respond to commands from the RC system. Since you will be converting the RC signals to digital outputs, the example scripts in the user’s guide having to do with switches and buttons will be relevant to you. To the Maestro, a digital output from an RC switch is indistinguishable from the output of a button.
You’ll need to figure out how to send a serial command from the master Maestro to the other one. This can be done on the master Maestro’s script using the serial_send_byte command. Here’s a simple example that sends a command to move servo 8 to a particular position:
0xFF serial_send_byte 8 serial_send_byte 140 serial_send_byte
The serial protocol of the Maestro is documented in the user’s guide.
You can easily test parts #1 or #2 from the Maestro control center by writing a simple script, and running it on the Maestro to see if it works.
Does this make sense?
–David
Yes, thanks it clears up a lot. I have managed to work it out now and I should be o.k for a while
We need a bit more help in this case I have been given a partner to program the pololu and whilst playing around with one of the controllers it was put into bootloader mode and I don’t know how to recover it and I have tried to find firmware to overwrite what was there in the first place, is there a way to remove the controller from bootloader mode that I haven’t discerned?
Hello. You can get the Maestro out of bootloader mode by power cycling it, or driving the RST line low temporarily. Please do NOT attempt to do a firmware upgrade, because there are no firmware upgrades available for you device. --David
I’m having a new problem now. Every time I try to run the send serial byte script, I get a serial protocol error with error code “0x0010”, the servos also twitch rather than move to the required position. Initially i thought it was my software causing the problem but it doesn’t seem to be.
I used the following example code and the same problem persisted.
Could you help me understand what is going on. The robot 70% complete and it all comes down to the Controllers.
I’m not sure why you are getting the serial protocol error, but it’s probably just something odd that happens on startup and shouldn’t cause any real problems. If you clear the error from the Maestro Control Center while your program is running does it come back?
The real problem is that you are not giving the servo any time to reach its target position; you are just changing the target between two different points hundreds of times per second. The servo can’t move that fast so it just twitches instead.
–David
It appears now and again and I think the problem is likely a loose connection between the respective Rx and Tx lines because it only happens when the line is disturbed.
I also have two servo twitch. when tested with one everything was fine. in the “Maestro Control Center” is the same … trying to work immediately with two servos, they immediately go out of their degrees with a crunch of gears, but no error 0x0010. Error 0x0010 when you work out with two servos from Java program.
how to get around?
Hello again, CMYK.
An error code of 0x0010 means that a Serial protocol error occurred. The most common reason for this is that the program sending serial commands to the Maestro sent an invalid sequence of bytes. I recommend checking your java program to make sure you are sending the right bytes. If you cannot find the problem, please simplify the program to the simplest possible thing that causes the error and post it here.
–David
Hello David)
import gnu.io.CommPort; import gnu.io.CommPortIdentifier; import gnu.io.SerialPort; import java.io.IOException; import java.io.OutputStream; public class PololuSerialExample { public PololuSerialExample() { } static void listPorts() { java.util.Enumeration<CommPortIdentifier> portEnum = CommPortIdentifier.getPortIdentifiers(); } static String getPortTypeName(int portType) { switch (portType) { case CommPortIdentifier.PORT_I2C: return "I2C"; case CommPortIdentifier.PORT_PARALLEL: return "Parallel"; case CommPortIdentifier.PORT_RAW: return "Raw"; case CommPortIdentifier.PORT_RS485: return "RS485"; case CommPortIdentifier.PORT_SERIAL: return "Serial"; default: return "unknown type"; } } public static void main(String[] args) { String portName = "COM5"; try { listPorts(); CommPortIdentifier portIdentifier = CommPortIdentifier.getPortIdentifier(portName); if (portIdentifier.isCurrentlyOwned()) { System.out.println("Error: Port is currently in use"); } else { CommPort commPort = portIdentifier.open("Owner", 2000); if (commPort instanceof SerialPort) { // we create a output stream for serial port: SerialPort serialPort = (SerialPort) commPort; serialPort.setSerialPortParams(9600, SerialPort.DATABITS_8, SerialPort.STOPBITS_1, SerialPort.PARITY_NONE); byte command = (byte) 0x84; int servo = 0; int target1 = 10; int target2 = 10; byte[] bytes = new byte[]{command, (byte) servo, (byte) target1, (byte) target2}; byte command1 = (byte) 0x84; int servo1 = 1; int target11 = 10; int target21 = 10; byte[] bytes1 = new byte[]{command1, (byte) servo1, (byte) target11, (byte) target21}; process(serialPort, bytes); Thread.sleep(3000); process(serialPort, bytes1); commPort.close(); } } } catch (Exception e) { e.printStackTrace(); } } public static void process(SerialPort serialPort, byte[] bytes) throws IOException { OutputStream serialStream = serialPort.getOutputStream(); serialStream.flush(); serialStream.write(bytes); serialStream.close(); } }
same thing happens and the Maestro control center if the servo to move the two. Only maesto not write - error against the wrong moves with a crunch of both. even if you move the slider one of the servo
One actuator can I manage.
Your code looks OK to me, but I have a hard time understanding your description of the problem. Is there a problem when you just control the servos through the Maestro control center? If so, we should focus on that problem and not worry about the java for now.
How are you powering your servos?
–David
Yes there are problems with the management servo of Maestro control center.
powering servo so 5v 320mA
A 320mA power supply will probably not be sufficient for multiple servos. We usually recommend that you have 1 A of capacity per servo. For example, if you have two servos try to find a power supply that can deliver 2 A or more.
–David
Thank you David. It worked. I’m not careful, read about three times the power. | https://forum.pololu.com/t/how-to-use-an-rc-controller-on-pololu-12-channel-controller/4400 | CC-MAIN-2022-21 | refinedweb | 1,487 | 62.38 |
We’ve upgraded several Rails 2.0 application to Rails 2.1 now, and we’ve compiled a list of little things to keep in mind as you upgrade. Hopefully this list will help you avoid banging your head against a wall.
Partial Updates
The updated_at and updated_on columns are NOT automatically updated on a #save
on an AR object in Rails 2.1, unless another column has also changed. In each
of the cases where we were relying on this behavior, we were using it to detect
in a general way that something had changed with the model (without introducing
a dependency on
acts_as_modified). Because Rails 2.1 has dirty attribute
checking, these methods were able to be refactored using this new functionality.
Will Paginate
Older versions of will_paginate are
broken on 2.1 (stack level too deep errors), to resolve, install the latest
version of
will_paginate like this:
sudo gem install mislav-will_paginate -s
Then put require
will_paginate in an initializer or in environment.rb (you
need to have added vendor/gems to load path already to do that).
This newer
will_paginate has changed the
#page_count method to be
#total_pages instead, so you’ll have to keep that in mind. Which means, some
of the
will_paginate view helpers changed as well – if you’ve monkey patched
them (you haven’t unless you’re me!) look for changes there too!
Finally, Squirrel’s
WillPagination module provides a
#page_count method. To get a
#total_pages
from squirrel results, we temporarily monkey patched Squirrel by adding a
lib/extensions/squirrel.rb which was just…
module Squirrel module WillPagination alias_method :total_pages, :page_count end end
We’ll need to update Squirrel soon to have this built in so that its compatible
with the latest version of
will_paginate.
Shoulda
Make sure you’ve upgraded Shoulda if you get errors about not being able to find fixture methods and/or assert, especially if these errors appear in setup blocks.
HAML
Unfortunately, haml isn’t Rails 2.1 compatible. To fix this, upgrade to haml-2.0.
Reply-to in mailers
If you have the following hack in your mailer, just remove it. Rails 2.1 does it for you:
def reply_to(str) @headers["Reply-To"] = str end
Changes in Active Record Attribute Filtering
ActiveRecord::Base#attributes does not allow filtering anymore (it does not accept :only, for example). You must do the filtering manually, with something like this:
def json_attributes_for(model, *attrs) attrs = [attrs].flatten.map(&:to_s) model.attributes.delete_if{|k,v| !attrs.include?(k) }.to_json end
Called like so:
json_attributes_for(page, :id, :keyword)
Changes in Template Rendering
Now in Rails 2.1, if you both foo.rhtml and foo.rxml exist and you aren’t explicitly specifying one or the other, Rails will render with foo.rxml. Renaming foo.rhtml to foo.html.erb fixes this, but in Rails 2.0, this was ok.
Relationship Optimized Eager Loading
In order to deal with the 1+n query problem, Active Record has changed how it does eager loading. Now, it will optimize out :includes on finders when they are not being used. When they are not being used is the key here. Active Record is supposed to noticed when there are additional conditions on the find that rely on the included table, and not leave it out.
However, on an association like this:
has_many :active_sites, :through => :clients, :source => :sites, :include => :domains, :conditions => 'domains.id IS NOT NULL'
Active Record leave out the domains table, even though it shouldn’t. We fixed
this by changing
active_sites to just a normal method, as the bug doesn’t seem
to happen in find, like this:
def active_sites sites.find :all, :include => [:domains], :conditions => 'domains.id IS NOT NULL' end
has_finder
has_finder has been integrated into Rails 2.1 as
named_scope, so you don’t
need it anymore. However, you also shouldn’t keep it around. For example,
has_finder-1.0.5 was giving us stack trace too deep exceptions when traversing
a
has_many :through association. Removing it in favor of
named_scope fixed
that issue.
In Conclusion
Those are the things that we’ve found that we feel might be helpful to other people out there. For the most part, upgrading to Rails 2.1 is a a straightforward process, and we’re quickly upgrading most of our apps.
Have you upgraded to 2.1 and run into anything that other people might hit? If so, feel free to add your gotchas to the comments. | https://robots.thoughtbot.com/gotchas-when-upgrading-to-rails-2-1 | CC-MAIN-2018-22 | refinedweb | 745 | 66.44 |
Opened 10 years ago
Closed 10 years ago
Last modified 8 years ago
#1529 closed Bug (Fixed)
TrayItemSetText() sets incorrect text
Description
Hello.
Windows 7 x64 En / XP SP 3 En
AutoIt 3.3.6.0
TrayItemSetText() sets incorrect text. Some kind of special symbol.
#include <Constants.au3> TrayItemSetText($TRAY_ITEM_PAUSE, 'Pause') TrayItemSetText($TRAY_ITEM_EXIT, 'Exit') MsgBox(0, 'Bug!', 'Check tray menu :p ')
Attachments (0)
Change History (8)
comment:1 Changed 10 years ago by MrCreatoR <mscreator@…>
comment:2 follow-up: ↓ 3 Changed 10 years ago by doudou
FYI: AutoIt 3.3.4.0 doesn't have this bug (at least not on WIN_XP/X86/Service Pack 3).
comment:3 in reply to: ↑ 2 Changed 10 years ago by MrCreatoR <mscreator@…>
comment:4 follow-up: ↓ 5 Changed 10 years ago by Valik
It is relevant. It narrows the range we have to search to find the change that caused the bug.
comment:5 in reply to: ↑ 4 Changed 10 years ago by MrCreatoR <mscreator@…>
comment:6 Changed 10 years ago by Jpm
- Resolution set to Fixed
- Status changed from new to closed
Fixed by ticket 1599
comment:7 Changed 10 years ago by TicketCleanup
- Milestone set to Future Release
Automatic ticket cleanup.
comment:8 Changed.
Same on WinXP SP2: | https://www.autoitscript.com/trac/autoit/ticket/1529 | CC-MAIN-2020-05 | refinedweb | 207 | 62.98 |
Hello. I have rolled out the Help Desk system for IT and have received such a positive response that I have been asked if there is a way to create a separate instance for our fleet maintenence guys to use and our Records Management group to receive requests and manage fullfilling those requests.
Is this possible? If so, how?
Thanks.
2 Replies
Nov 18, 2011 at 12:51 UTC
no you can not add any multiple help desk on same install, you will have to install separately for them. if you will create a help desk account for them they will only can access to the help desk portion, but can not add any multiple help desk, if they are on different location, then you might consider the remote collector feature of SW for them.
http:/
http:/
You can also take a look at this how-to for spiceworks for different departments,
http:/
Nov 20, 2011 at 10:15 UTC
I have a related question... I'm considering creating a helpdesk for internal users as well as one for external users. Is there a way to namespace the different ticket emails somehow i.e. if a support ticket is forwarded or copied to the IT helpdesk it won't wreak havoc due to ticket number collisions? | https://community.spiceworks.com/topic/170822-multiple-help-desks | CC-MAIN-2016-44 | refinedweb | 216 | 64.75 |
its ok you r right thanx anyway
its ok you r right thanx anyway
thx for the help :)
hi again to all i need to do this exercises
i am not good at java but i have to deliver them so
there they are , can someone help me
thanks
EXERCISE 1
Write a java program that receives the...
yes thanks a lot
thank for the help
import java.io.*;
public class IO_Tester {
public static int readInt() {
byte b[] = new byte[16];
String str;
hi to all i am new to java and i am doing homework
i must compile this but there is an error and i cant find were
can you help me please
i am using BlueJ and jdk
import java.io.*;
public... | http://www.javaprogrammingforums.com/search.php?s=086143fe2fb82204874b2085b0847279&searchid=1725655 | CC-MAIN-2015-35 | refinedweb | 125 | 64 |
There are some conditions where we have to convert one data type to the other data type. In programming languages these conversions are referred to as Type Casting.
Java type casting can be classified in teo types:
1. Implicit type conversion (Java’s automatic conversions)
2. Explicit type conversion
1. Java’s Automatic Conversions
When one type of data type is assigned to another data type value then java will convert the data types automatically under two conditions.
1. The two types are compatible to each other
2. The destination type is of larger range than the source type.
For example, an Byte is internally casted to int, because int has range larger than that of byte range.
but the vice versa conversion will result in an error
byte b=5; int i = b; // automatic type conversion b = i; // will give error (byte is of 1 byte whereas int if of 4 bytes)
2. Explicit or Forced Conversions
For this conversion a type is forcefully converted to another type as :
byte b=5; int i = b; // automatic type conversion b = (byte) i; // (byte) is for typecast from int to byte
have a look at another casting example
public class anotherCastingExample { public static void main(String args[]) { byte b = 42; char c = 'a'; int i = 50000; float f = 5.67f; double d = .1234; int cast1 = b; // implicit cast int cast2 = c; //implicit cast float cast3 = (float)d; //explicit casting System.out.println("cast 1st " + cast1); System.out.println("cast 2nd " + cast2); System.out.println("cast 3rd " + cast3); } } | http://www.loopandbreak.com/type-casting/ | CC-MAIN-2019-39 | refinedweb | 256 | 56.25 |
Non conforming type returned by TYPE-OF for displaced arrays.
This bug affects 1 person
Bug Description
The result of evaluating TYPE-OF on a displaced array returns a type
that involves an AND expression and a NOT expression. e.g.
(type-of (make-array 5 :displaced-to (make-array 10)))
=> (and (vector t 5) (not simple-array))
According to the CLHS page for TYPE-OF [1] it states that:
1. For any object that is an element of some built-in type:
...
b. the type returned does not involve AND, EQL, MEMBER, NOT, OR,
SATISFIES, or VALUES.
Thanks to |3b| for identifying the issue on #lisp.
[1] http://
Christophe Rhodes (csr21-cantab) on 2014-06-28
commit 7dfdf1224921ab0
696e95702c1cdf0 8203c1bf78
Author: Douglas Katzman <email address hidden>
Date: Tue May 27 20:28:28 2014 -0400
Don't return (AND ... (NOT SIMPLE-ARRAY)) from TYPE-OF. | https://bugs.launchpad.net/sbcl/+bug/1317308 | CC-MAIN-2016-36 | refinedweb | 146 | 63.7 |
The memory that (on the bottom of the stack). When we push an item onto our mailbox stack, we put it in the mailbox that is marked (which is the first empty mailbox), and move the marker up one mailbox. When we pop an item off the stack, we move the marker down one mailbox (so it’s pointed at the top non-empty stack itself technical note: on some architectures, the call stack grows away from memory address 0. On others, it grows towards memory address 0. As a consequence, newly pushed stack frames may have a higher or a lower memory address than the previous ones.
A quick and dirty call stack example
Consider the following simple application:
The call stack looks like the following at the labeled points:
a:
main()
b:
foo() (including parameter x)
main()
c:…) On modern operating systems, overflowing the stack will generally cause your OS to issue an access violation and terminate the program.
Here is an example program that will likely cause a stack overflow. You can run it on your system and watch it crash:
This program tries to allocate a huge (likely 40MB) array on the stack. Because the stack is not large enough to handle this array, the array allocation overflows into portions of memory the program is not allowed to use.
On Windows (Visual Studio), this program produces the result:
HelloWorld.exe (process 15916) exited with code -1073741571.
-1073741571 is c0000005 in hex, which is the Windows OS code for an access violation. Note that “hi” is never printed because the program is terminated prior to that point.:
hey I got a few questions:
1.All memory allocated on the stack is known at compile time. Consequently, this memory can be accessed directly through a variable.
if "int *arr" is allocated at compile time then how can it point to something that is allocated at runtime that is on heap segment ?
2.Saved copies of any registers modified by the function that need to be restored when the function returns.
3.Return values can be handled in a number of different ways, depending on the computer’s architecture. Some architectures include the return value as part of the stack frame. Others use CPU registers.
could you elaborate 2 & 3 a little pls ?
thanks.
1. The memory location and size of stack variables is known at compile-time. Their values are unknown.
2.
`fn` has a local variable `i`. When it calls `apple`, `i` has to be stored somewhere, so it gets pushed onto the stack. After `apple` finished, `i` gets restored.
3. Most commonly there's a register dedicated to storing return values. When a function is done, it writes its return value into that register and the caller reads the value from the register to know what the function returned.
thanks for your answer! and is the comment I've written in the code correct ? arr itself is on the stack and points to the first element of the array which is allocated on the heap right ?
2_you said the stack is like structs and arrays, I've also heard that on discord group,that there is a std::stack.but how must we use it I mean it doesn't seem it works like defining a vector like we say :
if it is not something we use this way then how is it used ?
Your comment is correct.
You have to specify the element type of the `std::stack`
See cppreference
Explantions on the stack in this chapters are too vague, confusing and even misleading.
Mailbox analogy is super confusing..
To actually understand on how this stuff works, watch this 2-day class on YouTube:
Taught by a brilliant Xeno Novah.
."
My understanding of this chapter is precisely that the callee is saving 'copies of registers' that will be restored at the end. I don't think the authors disagree with you.
How to find the size of code segment?
what is the work of registers ?
Assuming you mean, what is the "job" of a register:
Registers are one of the lowest-level denominations of memory. Registers are memory that sit on the CPU. The memory size of registers are incredibly small, but register I/O is also incredibly fast.
Registers can do a variety of things, depending on the CPU architecture, but commonly, registers hold vital values, such as memory addresses (e.g. stack pointers) or even literal values (for address offsets).
(Anyone feel free to correct me if I'm wrong)
If the stack has this limited memory, then where all the huge games like PES are stored? int the heap?
Yes.
Hello!
What is the meaning of the following line of code:
array is a pointer which points to an address in the heap that contains an int with value = 10?
OR
array is a pointer which is assigned 10 int (40 bytes with one int being 4 bytes) in the heap?
Thanks :)
`array` is a pointer which points to the first of 10 consecutive `int`s. Those `int`s are most likely on the heap, but the standard doesn't require it.
How can array point to first 10 consecutive ints? I mean it can let's say store the address of the first int but what about the other 9 consecutive ints?
You can calculate their positions using pointer arithmetic, or access them using array syntax
The pointer doesn't know whether it's pointing to a single `int` or to an array of `int`. It's up to the coder to know what their pointers are pointing to.
Alright.. in the below Video,
from 8:00 to 9:00 , it is shown if,
int* p = new int(10);
then pointer p has the address of a memory location in a heap where an int with initial value of 10 is created.
So the explanations here on Learncpp and the video seem opposite to each other. Or am I missing something?
Please help.
Noted! Thanks!
new int[10]; // Square brackets (Array)
All 10 ints will have garbage values here right?
If the array type is a fundamental type (int, float, etc.), yes.
If not, the elements will be default-initialized (You'll learn about classes later).
If the array is made of `std::string` for example, all elements are empty strings.
If you want to initialize the array, you can use empty curly braces as always.
Hi, Alex and Nascardriver!
Is it true that stack memory allocation has automatic duration and scope-dependent and heap memory allocation has dynamic duration and scope-independent?
C++ doesn't put any requirements on the underlying memory model.
But yes, dynamic allocations usually use the heap and automatic allocations use the stack.
Hi, Alex (and Nascardriver too)
I don't undestand your mailboxes analogy. I can't imagine and just don't get it. Your plate analogy is perfect by the way. I'm sorry Alex, in my place there is no any mailboxes.
Hello Alex (or Nascardriver),
I am also unable to understand the mailboxes analogy, but I want to understand it. Can you please elaborate that in detail so that I can understand.
Thanks in advance.
And a huge thank you for this amazing tutorial also.
Try drawing a picture of a bunch of mailboxes stacked on top of each other, and an arrow pointed at the lowest mailbox. For each item pushed on the stack, put the item in the mailbox being pointed to and move the arrow up. For each item being popped, move the arrow down and remove the item from the mailbox that is now being pointed to. That's all it is.
The plate analogy is great for understanding how to use a stack, but it does not quite accurately represent what the stack looks like. I suppose this isn't important for learning c++ (apart from stack overflows) but it's important for learning how a computer works.
(note: I am a student, so I may not be 100% accurate in my explanation)
For computers, you only get a certain amount of memory. As Alex says, "On Windows, the default stack size is 1MB. On some unix machines, it can be as large as 8MB."
You cannot create more memory (like the old joke "download more RAM"), nor can you destroy memory (or else you wouldn't be able to get it back).
In the plate analogy, you can keep adding/pushing more plates without ever hitting a cap. Also, when you take/pop plates, the volume of the plate stack takes up decreases. Since memory in the stack has to stay constant (not right either because it shares space with the heap, but close enough), the plate analogy fails to take this into consideration.
The mailbox analogy is like the plate analogy except that the memory is already laid out. Each mailbox represents space to put a chunk of data into and the entire stack of mailbox represents all of the memory given to you. You cannot remove mailboxes nor can you add more because you would anger the authorities. Instead of putting a plate on top of the plate stack, you place it inside the bottom-most available mailbox.
But where is the bottom-most available mailbox? You can't look at the top of the stack (while the plate analogy incorrectly allows you to) because all of the unused mailboxes are blocking your view. So you have to keep track of where the bottom-most available mailbox is by yourself (the push() and pop() functions keep track for you), i.e. through a marker or post-it. Whenever you push or pop something, you know which mailbox is the bottom-most available mailbox because of the post-it. When you push, remember to move the post-it up 1 mailbox for the new bottom-most available mailbox. Same for popping except you move 1 down.
The reason I say "available" instead of "empty" is because when you pop some data, you can't empty it out because it needs to have some 0s and 1s in it. You could set the memory to all 0s, but I think it would take too long and maybe a few other restraints. So instead, we just move the post-it down 1 mailbox and just overwrite the data in the bottom-most available mailbox if we have to re-use it.
"Because the stack is not large enough to handle this array, the array allocation overflows into portions of memory the program is not allowed to use. Consequently, the program crashes."
So, does the memory overflowed by the array allocation is restored to the previous contents or the overflow does'nt really changes the content of the memory.
The code in question doesn't change the memory, it merely reserves it, no data has been overridden.
Data is never automatically restored, and there's no reason to do so, your program crashed anyway.
Hey Alex, in the foo example option c seems to be lacking code tags around main().
Great lesson, really interesting reading about the low level code and functionality!
Thanks! The formatting has been fixed.
Hi Alex,
In advantages and disadvantages of stacks, following line is mentioned:
"All memory allocated on the stack is known at compile time."
As per my understanding, the stack frames are created/destroyed when a function is called/returned.
So how memory allocated on the stack is known at compile time.??
The compiler is responsible for laying out all of the stack frames at compile time (with local variables translated into an memory address offset from the start of the stack frame).
It doesn't necessarily know (or care) which stack frame is actively loaded into memory at a given point.
hi Alex,
i am missing something under the bss segment(unitialized data segment)?
my point here is that "aren't zero initialid g_variables really initialized?"
i think they should be stored in the data segment where initialize g_variables are stored.
If I'm wrong, then Wikipedia is wrong too.
The first line of this page is written "The memory a program", should be "The memory *of* a program".
outPut:
Size of Class: 8
1. here it shows size of the class as 8, it means memory allocated for class variable (variable1 & variable2),but what about the class member function? where it gets allocate memory for member function?
2. But if I made member function virtual
then it shows
size of Class: 16
Why is that so?
Hi Akshay!
> what about the class member function?
Functions aren't members of class instances, because the function is the same across all instances. It is created once, somewhere in memory, and every occurrence of the function is replaces with the address the function is located at.
> But if I made member function virtual
Classes with functions store a pointer to their virtual function table. You're compiling in 64 bit mode, so that pointer is 8 bytes in size, 8 (variables) + 8 (vtable pointer) = 16. You can add more virtual functions without increasing the class size, because each instance only stores a pointer to the vtable, not the vtable itself. More about this in chapter 12.
Thanks @nascardriver, my confusion got cleared now.
Hi Alex,
here it is mentioned that
"The call stack is a fixed-size chunk of memory addresses."
but in
it mentioned that 'The stack grows towards the lower part of memory that is towards heap. In other hand heap grows towards higher part of memory that is towards stack.'
that means call stack size is not fixed.
The maximum stack size is fixed, but the stack itself can grow and shrink.
Readers might assume that a stack grows from lower memory addresses to higher if they follow the plate analogy. On the standard PC x86 architecture the stack grows toward address zero. That could lead to confusion if looking at actual addresses. It might be something worth mentioning.
Added a technical note about this to the article. Thanks for mentioning!
Alex,
1. Since dynamically created arrays (or any variable type) stay on the heap and not on the stack, does it mean that, their scopes are global? Say, we have a main() function that calls another function createArray() that creates an array dynamically. When the program returns to the main, will it be able to refer to the array created without createArray() having to return the dynamically created array?
2. In case returning the array to the main() from createArray() is mandatory, is the return type (for the createArray() function), when returning lets say a 2 dimensional array with integer elements, int** ?
Hi Mahmud!
1:
The array still exists after @createArray has finished, however there is reliable way for accessing it anymore, since you don't know where it is.
2:
Yes.
Bonus (1):
@nascardriver
Thanks a lot for the clarification!
Hi nascardriver!
I found this code (Bonus (1)) interesting so I started playing with it.
In the '// Use @arr here' section I added the following:
This outputs '16d' (hex) rather than '365'.
However, what we seem to want from this code is an accessible/usable two dimensional 'normal' int array (unless I'm mistaken).
I'm guessing the issue is being caused by the iostream and use of std::hex somehow?
How could it be resolved?
Once you set a flag in a stream, it will stay set.
You can switch back to decimal using @std::dec
Everything after this line will be decimal
1) Nascardriver's answer is good. But it sounds like you're confusing the concepts of scope and duration. Duration has to do with how long something lives for. Scope has to do with where you can access something. So for a dynamically allocated array, it has dynamic duration (meaning it lives until you explicitly destroy it). Dynamically allocated memory doesn't have "scope", since you never access it directly. Instead, you access it through pointers. Those pointers have scope.
//is it stack - friendly code?
#include <vector>
using std::vector;
void foo( void );
int main( void )
{
foo();
return 0;
};
void foo( void )
{
const uint32_t N = 1E9;
vector < float > vec;
vec.resize( N );
};
Yes, because vectors allocate their elements on the heap. However, if this were a std::array, that would not be the case.
An easy way to tell is this: if you take the sizeof() your variable, that's how many bytes on the stack that variable is taking. Even though your vector has a lot of elements, the sizeof(vec) should be relatively small.
Hi Alex!
I have a question. Does making a new block creates a new stack aswell? Or variables only delete because they're out of scope?
No, blocks inside a function don't create an additional stack frame. Compilers have some leeway in terms of how they handle the variables inside functions, but most allocate room for variables defined in the function (regardless of nested blocks) on the stack at the start of the function. The compiler can enforce the proper scoping rules.
Two questions I've had for a while now, but haven't been able to find answers to.
1. Are the stack and the heap (and other types of memory) physically different? Ie. Could you open up you machine and point to the stack as being distinct from the heap? Or are they just different software allocations within a homogeneous RAM chip?
2. Why is LIFO an efficient way to order memory? From a naive perspective, it seems like it would be most inefficient, ie. if I want to access something on the bottom, I have to remove every item from the top.
1) It depends on what you mean by distinct. But generally, yes, they're distinct portions of memory. The stack is fixed in size, and the heap has a variable size. A given memory address could belong to one or the other, but not both simultaneously. How these are mapped to actual RAM is determined by the OS (virtual memory makes this complicated).
2) LIFO is useful because it's very simple and fast to add or remove the top element, and it ensures all items in the stack are contiguous. If you need to access something on the bottom, you probably shouldn't use a stack in the first place.
Could you describe me what a "register" is, or give me an example?
A register is a small piece of memory that is part of the CPU. Many CPUs load data from main memory into registers, operate on the registers, and then output the result of the register back to main memory.
1) Saved copies of any registers modified by the function that need to be restored when the function returns.
2) Return values can be handled in a number of different ways, depending on the computer’s architecture. Some architectures include the return value as part of the stack frame. Others use CPU registers.
I didn't understand these 2 points.
Can you explain me these? Reply Please.
This would require explaining a lot about computer architectures, which isn't really the point of these tutorials. Just note this as an interesting aside and move on.
Should i read on net about registers?
If you're super interested in computer architecture and the intimate details of how all this stuff works, sure. But if your primary interest is learning C++ then I'd bookmark it to return to later.
Hello Alex,
When is the best time to use vector instead of array and vice versa?
Use std::array when you know the array's length at compile-time and it won't change. Use std::vector if you don't know the array's length until runtime, or the length needs to change.
Hello alex ,
First i want to thank you for this tutorial, but i deployed a multithreaded program which when allocate all memory on the stack use arrays instead of vectors for example
func x(some parameters){
vector <double >a;
a.resize(4)
vector <double> e;
a.resize(4)
vector <double> c;
c.resize(4)
vector <bool> d;
d.resize(4)
vector <bool> x;
x.resize(4)
for( int i =0 ; i <100000 ; i++ )
for (int j =0 ; j <100000; j++ )
{
double inphase[4];
double quadrature[4];
bool status [4];
bool hidden [4];
here some reading and writng in those arrays and vectors
}
}
program executes well but one transforming the vecyors to arrays program becomes 10x slower how is that and when switching array with vectors and vectors wth arrays program executes fast again , i am using open mp and visual studio 2010 by the way the parallization is for this function each thread call this function seperately.
I'm not sure. In general, std::array should be faster than std::vector (at the cost of being less flexible). You must be accruing some kind of inefficiency with std::array or stack allocation, perhaps because multiple threads don't share a stack.
and that exactly what i need to that threads don't share the stack what is the drawback of threads not sharing same stack ?
I'm not the right person to ask. I haven't done multithreaded programming in a long time, and am not all that familiar with the pros and cons of it.
Thanks lot for your concern alex
That was an interesting read. I had a random idea popped in my head that might sound dumb. I understand the difference between the stack and the heap (I think) but I wonder is it possible to use a pointer to point to another stack instead of grab memory from the heap? Instead of calling a function and pushing onto the main stack, I was thinking maybe the pointer can allocate the function call to another stack and then you could push additional functions onto that.
I guess the only thing that might be wrong there is that the OS will terminate the program for trying to access memory it wasn't given. Is that right?
I'm not sure if I explained that clearly but in my mind I imagine something like:
A program is only set up with one stack, and the compiler is set up to automatically use it. So when you call a function, it goes on the main stack. I'm not aware of any way to reroute to a second stack.
Dear Alex,
I have a module and I want to save it in a persistent memory in order not to be lost or crashed when I call it again starting from the point where I left it, and I found the only way to do that is to transfer it to a disk after saving it in a file using mmap command, but the problem is I found this command can not be executed under Windows OS because it's included in <sys/mman.h> which is one of the Linux header files. So is there an alternative command can be executed under Windows platform using vc++.
Regards.
I can't really speak to this, as it's OS specific and not in my core area of knowledge. Google search or Stack Overflow is probably your best bet here.
Dear Alex,
I couldn't because each project has its header files and main function! so how can I merge these two projects together in one project. Actually, I'm looking for steps to do that.
The header files you should just be able to copy into your interface project. The main function I'm not sure about.
Hi Alex,
Thanks for the very detailed guide. Very easy to understand for someone who's not a CS major.
Name (required)
Website
Save my name, email, and website in this browser for the next time I comment. | https://www.learncpp.com/cpp-tutorial/79-the-stack-and-the-heap/ | CC-MAIN-2020-29 | refinedweb | 3,988 | 72.26 |
Details
- Type:
Improvement
- Status:
Closed
- Priority:
Minor
- Resolution: Fixed
- Affects Version/s: 1.0-JSR-3
- Fix Version/s: 1.1-beta-2
- Component/s: None
- Labels:None
- Number of attachments :
Description
allow
data = new groovy.sql.DataSet(tablename)
List rows = data.rows()
to fetch all rows at once analogous to Sql.rows(...)
Activity
Came to the realization that I could just re-use the rows() method on the SQL class and just pass in a SQL
string. D**mba@@ to go and cut-and paste code. I uploaded the revised dataset class. The changes still work and the unit tests pass.
implemented
I think John Carnell's query is an important one here:
"... However, what happens if the user is trying to filter based on a where. Shouldn't the call really return all of the rows for the DataSet."
I think the answer to this is yes.
Here is some code:
def athletes = [ [firstname: 'Paul', lastname:'Tergat', dateOfBirth:'1969-06-17'], [firstname: 'Khalid', lastname:'Khannouchi', dateOfBirth:'1971-12-22'], [firstname: 'Ronaldo', lastname:'da Costa', dateOfBirth:'1970-06-07'], [firstname: 'Paula', lastname:'Radcliffe', dateOfBirth:'1973-12-17'] ] athleteSet = db.dataSet('Athlete') athletes.each { a -> athleteSet.add(a) } youngsters = athleteSet.findAll{ it.dateOfBirth > '1970-1-1'} def rows = [] youngsters.each { rows += it.lastname } println rows.join(', ') // => Khannouchi, da Costa, Radcliffe println youngsters.rows().size() // => 4
I would expect rows() for youngsters to return the three youngster rows not the entire original table (4 rows here).
do I see it right, that we just need to replace
public List rows() throws SQLException { String sql = "SELECT * FROM " + table; return rows(sql); }
with
public List rows() throws SQLException { String sql = getSql(); return rows(sql); }
in DataSet?
Thought you would have already clocked off, so went ahead and fixed it.
Yes, your suggestion was almost what I did - just:
rows(getSql(), getParameters())
because sql is stored proc with params.
Added two patches for the proposed enhancement to the DataSet class. I have added the rows() method
and the firstRow() method. I have also updated the SqlCompleteTest class to test these two methods.
However, I do have a question on this open improvement. The way I read the "improvement" is that the rows() method
will return all of the rows on the underlying tableset. To implement this I simply did a SELECT * FROM the table to
build the data. However, what happens if the user is trying to filter based on a where. Shouldn't the call really return all of the rows for the DataSet.
Please let me know how exactly this code should work. I can change the code to behave whatever way needed. | http://jira.codehaus.org/browse/GROOVY-1109?attachmentSortBy=dateTime | CC-MAIN-2013-20 | refinedweb | 441 | 68.36 |
#include <stdio.h>
int scanf( char* string, ... );
This function reads formatted input from stdio. The data is converted
and stored at addresses given by the arguments that follow string.
These arguments should also contain format specifications for the data read. If
the end of the file is reached before any conversion takes place, scanf() will
return EOF. If successful, it will return the number of items
read and stored.
See also: fscanf().
Back to Essential C Functions.
See also: fscanf().
Back to Essential C Functions.
Log in or registerto write something here or to contact authors.
Need help? accounthelp@everything2.com | http://everything2.com/title/scanf | CC-MAIN-2017-13 | refinedweb | 101 | 70.9 |
Use Python for Scientific Computing
which would give us a 500x500 element matrix initialized with zeros. We access the real and imaginary parts of each element using:
a.real[0,0]=1.0 a.imag[0,0]=2.0
which would set the value 1+2j into the [0,0] element.
There also are functions to give us more complicated results. These include dot products, inner products, outer products, inverses, transposes, traces and so forth. Needless to say, we have a great deal of tools at our disposal to do a fair amount of science already. But is that enough? Of course not.
Now that we can do some math, how do we get some “real” science done? This is where we start using the features of our second package of interest, scipy. With this package, we have quite a few more functions available to do some fairly sophisticated computational science. Let's look at an example of simple data analysis to show what kind of work is possible.
Let's assume you've collected some data and want to see what form this data has, whether there is any periodicity. The following code lets us do that:
import scipy inFile = file('input.txt', r) inArray = scipy.io.read_array(inFile) outArray = fft(inArray) outFile = file('output.txt', w) scipy.io.write_array(outFile, outArray)
As you can see, reading in the data is a one-liner. In this example, we use the FFT functions to convert the signal to the frequency domain. This lets us see the spread of frequencies in the data. The equivalent C or FORTRAN code is simply too large to include here.
But, what if we want to look at this data to see whether there is anything interesting? Luckily, there is another package, called matplotlib, which can be used to generate graphics for this very purpose. If we generate a sine wave and pass it through an FFT, we can see what form this data has by graphing it (Figures 1 and 2).
We see that the sine wave looks regular, and the FFT confirms this by having a single peak at the frequency of the sine wave. We just did some basic data analysis.
This shows us how easy it is to do fairly sophisticated scientific programming. And, if we use an interactive Python environment, we can do this kind of scientific analysis in an exploratory way, allowing us to experiment on our data in near real time.
Luckily for us, the people at the SciPy Project have thought of this and have given us the program ipython. This also is available at the main SciPy site. ipython has been written to work with scipy, numpy and matplotlib in a very seamless way. To execute it with matplotlib support, type:
ipython -pylab
The interface is a simple ASCII one, as shown in Figure 3.
If we use it to plot the sine wave from above, it simply pops up a display window to draw in the plot (Figure 4).
The plot window allows you to save your brilliant graphs and plots, so you can show the entire world your scientific breakthrough. All of the plots for this article actually were generated this way.
So, we've started to do some real computational science and some basic data analysis. What do we do next? Why, we go bigger, of course.
So far, we have looked at relatively small data sets and relatively straightforward computations. But, what if we have really large amounts of data, or we have a much more complex analysis we would like to run? We can take advantage of parallelism and run our code on a high-performance computing cluster.
The good people at the SciPy site have written another module called mpi4py. This module provides a Python implementation of the MPI standard. With it, we can write message-passing programs. It does require some work to install, however.
The first step is to install an MPI implementation for your machine (such as MPICH, OpenMPI or LAM). Most distributions have packages for MPI, so that's the easiest way to install it. Then, you can build and install mpi4py the usual way with the following:
python setup.py build python setup.py install
To test it, execute:
mpirun -np 5 python tests/helloworld.py. | http://www.linuxjournal.com/magazine/use-python-scientific-computing?page=0,1&quicktabs_1=0 | CC-MAIN-2015-27 | refinedweb | 721 | 66.03 |
Content-type: text/html
symlink - Makes a symbolic link to a file
#include <unistd.h>
int symlink (
const char *path1,
const char *path2 );
Interfaces documented on this reference page conform to industry standards as follows:
symlink(): XPG4-UNIX
Refer to the standards(5) reference page for more information about industry standards and associated tags.
Specifies the contents of the symbolic link to create. Names the symbolic link to be created.
The symlink() function creates a symbolic link with the name specified by the path2 parameter which refers to the file named by the path1 parameter.
Like a hard link (described in the link() function),. Unlike hard links, a symbolic link can cross file system boundaries.
When a component of a pathname refers to a symbolic link rather than a directory, the pathname contained in the symbolic link is resolved. If the pathname in the symbolic link starts with a / (slash), the symbolic link pathname is resolved relative to the process root directory. If the pathname in the symbolic link does not start with a / (slash), the symbolic link pathname is resolved relative to the directory that contains the symbolic link.
If the symbolic link is the last component of the original pathname, remaining components of the original pathname are appended to the contents of the link and pathname resolution continues.
The symbolic link pathname may or may not be traversed, depending on which function is being performed. Most functions traverse the link.
The functions which refer only to the symbolic link itself, rather than to the object to which the link refers, are: An error is returned if a symbolic link is named by the path2 parameter. If the file specified is a symbolic link, the status of the link itself is returned. An error is returned if a symbolic link is named as the path parameter. This call applies only to symbolic links. A symbolic link can be removed by invoking the remove() function. If the file to be renamed is a symbolic link, the symbolic link is renamed. If the new name refers to an existing symbolic link, the symbolic link is destroyed. An error is returned if a symbolic link is named as the path parameter. An error is returned if the symbolic link named by the path2 parameter already exists. A symbolic link can be created that refers to another symbolic link; that is, the path1 parameter can refer to a symbolic link. A symbolic link can be removed by invoking unlink().
Search access to the symbolic link is required to traverse the pathname contained therein. Normal permission checks are made on each component of the symbolic link pathname during its resolution.
Upon successful completion, the symlink() function returns a value of 0 (zero). If the symlink() function fails, a value of -1 is returned and errno is set to indicate the error.
If the symlink() function fails, errno may be set to one of the following values: The requested operation requires writing in a directory with a mode that denies write permission, or search permission is denied on a component of path2. The directory in which the entry for the symbolic link is being placed cannot be extended because the user's quota of disk blocks on the file system containing the directory has been exhausted. The path specified by the path2 parameter already exists. Too many symbolic links are found in translating path2. The length of the path1 parameter or path2 parameter exceeds PATH_MAX, or a pathname component of path2 is longer than NAME_MAX while {_POSIX_NO_TRUNC} is in effect. The path2 parameter points to a null pathname, or a component of path2 does not exist. The directory in which the entry for the symbolic link is being placed cannot be extended because there is no space left on the file system containing the directory.
Functions: link(2), readlink(2), unlink(2)
Commands: ln(1)
Standards: standards(5) delim off | http://backdrift.org/man/tru64/man2/symlink.2.html | CC-MAIN-2017-22 | refinedweb | 655 | 53.31 |
Hi
Go through this,_<program name><3 digit suffix>"
-
Also Check this out--
1. Storage Location Determination at Sales Order:
You can do it through user exit... this is the exit...
USEREXIT_SOURCE_DETERMINATION (MV45AFZB)
2. Pricing Authorisation:
Refer OSS Note 105621
3. Making some of the Fields as Mandatory in Customer master based on the Business Rule.
Enhancement is SAPMF02D. Activate it through CMOD. The Following are the example to stop the Customer master creation with Tax jurisdiction Code.
&----
*& Include ZXF04U01 *
&----
*This user exit has been written for not to save the customer master *
*to whom the Tax Juridiction code is not updated. The Export Customer *
*is exceptional. Exit written by M. Murali on 18.05.2004. *
IF i_knvv-vkorg ne '2000' .
if i_kna1-TXJCD is initial .
Message ID 'ZK' type 'E' number '005' .
endif .
endif .
4. Different Billing Document Number range based on Company code / Plant Code for the same billing type.
The Following are the example for different invoice number based on Plant.
Use the Exit RV60AFZZ.
FORM USEREXIT_NUMBER_RANGE USING US_RANGE_INTERN.
if (
xvbrk-fkart = 'ZF3' or xvbrk-fkart = 'ZF2' )
and xvbrk-vkorg = '1000'.
case xvbrp-werks.
when 'AGR'.
us_range_intern = '22'.
when 'AHD'.
us_range_intern = '28'.
Endif.
ENDFORM.
5. Changing the billing Date to Current date or Changing some of the Field values of the Billing data.
Create a Data transfer routine for the billing and do the changes what ever you want. The following are the example for changing the billing date to current date.
&----
*& Form DATEN_KOPIEREN_611
&----
text
----
FORM DATEN_KOPIEREN_611.
IF VBRK-FKDAT NE SY-DATUM .
SY-SUBRC = 0 .
MESSAGE ID 'ZSD1' TYPE 'I' NUMBER '014' .
VBRK-FKDAT = SY-DATUM .
ENDIF .
ENDFORM. "DATEN_KOPIEREN_611
-
Userexits are places in the SAP standard code that are designed to insert code by the customer. It happens often that some values are set by default by SAP but are not appropriate for the business
For more information on User Exits, please check this links.>>>>>>
-
Some more information on userexits and enhancements.-----
Enhancement/Modifications
1) Execute tcode SMOD to find available enhancement/modifications.
2) Create a project for the enhancement in tcode CMOD.
3) You must activate your project first in order to hit a break-point or get into debug mode for your existing enhancements/modifications, if you do not, the best you will be able to do is step through the main program until you hit the call for that particular customer enhancement.
4) To get into debug, you can enter a hard break-point in the enhancement itself, set a soft break-point with the stop sign, or the long way, before you execute your transaction or while you are in your transaction, you can place a /h in the ok code area (this is the area of your gui where you can type in a tcode). Once you have the /h, hit enter and that will take you into debug, from there, you can do many different things to find exactly what you are looking for.
User Exits
1) Identify the main program you want to locate a user exit/debug.
2) For example, go to SE80 and do a search by program or dev class (SAPMV45A sales order or Dev Class VMOD, most SD user exits are in this dev class). In SE80 if you go by program, most user exit programs end in a 'Z' on a rare occasion 'X' or 'Y'.
3) If you are looking at including MV45AFZZ, you can see where there are different forms. These forms will get called at times within the program. If you are looking to fill the storage location on the sales order, you will probably want to take a look at the perform that fills in a field in vbap.
4) If this is what you are trying to accomplish, you will need to do the select against the config Table TVKOL based on the shipping point/plant and possibly storage condition based on your picking strategies.
5) For the debug part, you can do the same as in the enhancements/modifications but you will not need to activate any projects
-
PDF which is available with me Userexits in FICO. We can do the same in SD also.------>
-
Go to transaction SE81.
Click on SD .
Click "edit" on the menu bar .
Choose select subtree.
Click on "information system" .
Open Environment node, customer exits, and enhancements.
Press F8 to get all the user exits for that module.
In brief: SE81->SD->Select subtree->Information System->Envir->Exit Techniques->Customers exits->enhancements->Execute(F8)
- 'NNN'.
Mohan
Award points if it helps
Add comment | https://answers.sap.com/questions/2056850/index.html | CC-MAIN-2019-18 | refinedweb | 761 | 66.03 |
Agenda
See also: IRC log
<JeniT>
RESOLUTION: approve previous minutes
<danbri> I'll note that I requested mon/tuesday TPAC per my action.
<JeniT>
jenit: this section is a sketch of different
methods of finding a metadata document that provides metadata about a
CSV, or finding it within the CSV itself.
... The metadata document can tell an application how to deal with that file, in particular, how to transform into different formats.
... In that document, there are five different methods listed with issues.
... 3.5, use a standard path
<JeniT>
yakovsh: is 3.5 specifically when used with HTTP?
<danbri> http not https, ftp, gopher, … ?
yakovsh: If so, why is the standard name considered? If using HTTP, then the Link header can describe it.
<AndyS> and "file:"
jenit: yes, and 3.4 talks about using the Link header. When we discussed on the list, people felt that having a standard location relative to the CSV would be easier than controlling the Link header.
<danbri> ."
yakovsh: Can 3.5 also be used when files are on disk? Why HTTP only?
jenit: No particular reason, and that's a good point.
<danbri> nearby:
andys: I think we also need to be adhere
when we're deadlining with packages of CSV files, in which case a
package description file will be needed. Something to address that will
be needed.
... When I mentioned being able to work out a file given a CSV file, I was thinking of one per CSV, such as given foo.csv, it might be foo.csvm.
jenit: something about being in a similar
directory
... Where I've seen metadata files used with CSV, such as simple data format, or googles, the metadata file has always been describing several related CSV files.
<danbri> (this? -> DSPL )
jenit: I took it as a strength that a metadata file would describe several CSV files, as that matched current usage.
<JeniT> danbri: yes
<yakovsh> for favicon, here is the link to the w3c doc:
andys: that's good when there's one publisher, but CSV files may come from a number of different publishers, and the publisher is just mechanically moving them into place.
<AndyS> "the final publisher" putting up the files on the directory.
<danbri> erg
andys: In a lot of environments, it's either impossible to control, or very difficult to control in terms of technology
jenit: what about if you use a suffix on a file name; if you want to use it on all files in a directory, use a suffix on the directory name.
andys: perhaps we document both, and in an issue say that the WG is likely to pick one, so people have a warning. It should depend on actual user experience.
jenit: let's change the document to cover both cases. I think it's reasonable for both to be possible: somewhere you look for an individual CSV file, and another default location.
<ivan> +1 to Jeni, we should have several documents in a priority order
<danbri> makes sense avoiding .xyz
<ivan> +1 to danbri + jeni, too
jenit: I'm inclined to use a suffix that doesn't look like ".foo", as those are associated with different formats.
<yakovsh> +q
jenit: do we anticipate a single mime type for CSV metadata, or not? We'll take this to the list.
<Zakim> danbri, you wanted to ask about "/.well-known/" (". Why aren't per-directory well-known locations defined?")
<yakovsh> its an rfc:
<yakovsh>
<yakovsh> the actual registry
danbri: There's an IETF draft from mnot and
friends. As I understand it, it's really one place per site. I wonder if
we could consider extending it to be per-directory.
... Personally, I'm not excited about well known paths, but we should look at site-map files.
jenit: yes, .well-known is one-per-site.
... Given we're trying to do something really easy, I think it's unlikely they could access either .well-known or site-map.
yakovsh: Are we sure that every OS uses file extensions? I think MacOS uses something in the file itself.
<ivan> OS X uses extensions I believe
<danbri> osx is hybrid now
jenit: I think Mac uses a combination of both. When we're talking about a default method, I don't think that's relevant.
yakovish: regarding .well-known URIs, it's tied into AWWW. It might be prudent to reach out to mnot. It's not clear how widely it's used, such as robots.txt
jenit: some of these (e.g. robots.txt) came before .well-known. I'm not sure it's a relevant notion.
<danbri> even if something's not in the .well-known/ registry, it can still provide a safe sub-namespace to put such names where they'll only clash with other would-be-well-known names, and not with publisher names
<JeniT>
jenit: we'll consider a standard path and a backup, possibly using a file extension.
<JeniT>
jenit: Moving on to 3.4, I think this is fairly straight forward.
jenit: Just to be sure rel=describedby is the right header
andys: we have the two cases again, a description per CSV, or one for a group.
<JeniT>
jenit: It doesn't make a case, as the Link header describes the resource (which could be multi-part?)
<danbri> describedby is registered in
jenit: Perhaps we can assume that we always have a package description.
andys: if multiple people are dropping files into a directory, this might not be a good assumption.
<danbri> +1 for one type of metadata file
<danbri> we can have conventions evolve over time
jenit: Andy seemed to be saying there would be two different types of files (packages, and individual). I'm suggesting there should be just one, but sometimes the package might just have one file in it.
<danbri> there might be a few of these that get composed
<danbri> (i.e. merged)
andys: there might be one directory with mixed information. Perhaps it should be either one or the other, a package or an individual file. Going down every path might be exhausing
ivan: from a syntactic point, does describedby allow me to use a list of URIs or just one?
jenit: I think you can have multiple Link headers, with different types and locations.
ivan: that's also related to Andy's
question: the various access methods. We have to allow for different
routes to get metadata with a prioritization.
... In this sense, if it's one link header with a list of references, they are in priority order, and if some are metadata for the package, and some individual, falls back to priority.
... I can imagine a system setting up a standard describedby for all CSV files, and the user adds more metadata with a well-known URI.
<JeniT>
jenit: I tried to put in something about
cascade in section 3. That might not satisfy your requirements.
... for the Link header, we should say you can have multiple link headers and that they are merged with the one at the top being the highest priority.
ivan: the problem is, what does priority really mean? Suppose it's all in RDF. The "RDF way" would say that all statements are accumulated and do not hide each other. Other systems would do occlusion.
<danbri> oops!
<danbri> 3 pixel difference between selecting a browser tab vs closing it
jenit: it says if the same property is specified in two different locations, information closer to the document should override that which is further away.
yakovsh: In RFC4180 I started defining metadata as part of the mime type. If the mime type is a good place to stash metadata?
jenit: Probably not, as it gets lost when it moves around.
<danbri> (isdescribedby seems ok to me.)
<JeniT>
danbri: it seems everyone wants to talk
about RDF mappings, but we've been putting that off.
... Also, XML, JSON, ...
jenit: the best way to structure discussion
is to have a spec to discuss and "kick".
... I'd like to have people step forward to edit a document and have others contribute.
... On CSV to RDF
<danbri> I'd like to help, and relay in some ideas from
<AndyS> We have two already -- and
<danbri> also we have a backwards sparql proof of concept,
<Zakim> danbri, you wanted to suggest we pick some concrete CSV files (from the UC work) to focus the mapping design
jenit: I find it hard to be able to say that one direction is definitely the way to go. I think the next step is for someone the characterize the difference between the different approaches so we can have an educated discussion in order to make a discussion.
danbri: I'm feeling a bit overwhelmed by the different threads using a set of CSV files. Then we can compare different designs.
andys: There already are examples in the different examples.
danbri: I think we should have some core examples.
<danbri> what can we take from WD-csvw-ucr-20140327 ?
jenit: I think the first step is to focus on
the direct mapping, i.e. with zero metadata. If we can get that down,
we're in a good position.
... Who'd like to take forward direct mapping for CSV to RDF, the possibilities and advantages/disadvantages with proper examples.
... can Andys and gkellogg get together on this?
andys: I don't think this quite touches on
the fundamental differences between the two approaches.
... Gregg's very much based on JSON-LD, and I'm interested in a mapping to RDF triples.
<danbri> ACTION: danbri try expressing a direct mapping expressed using [recorded in]
<trackbot> Created ACTION-10 - Try expressing a direct mapping expressed using [on Dan Brickley - due 2014-04-02].
<AndyS>
andys: I found three classes of JSON-style output. I have no idea which are commonly used. I understand the first (one row to object), I understand column arrays, I don't know what the background is about turning everything into arrays without objects inside.
jenit: given that the main difference
between the two approaches is about the syntax of the metadata document,
I'd like to get something down as a starting point, being just a direct
mapping. This would be really helpful.
... Andy, if you could do this?
gkellogg: why don't we work together.
andys: I'd like feedback on what I've written.
jenit: something in ReSpec on GitHub; copy/paste is fine.
andys: I'm looking at a mapping to RDF, gregg's looks to both RDF and JSON through JSON-LD. When you compare and contrast, it might not be as useful.
ivan: In a way, the JSON vs RDF model is
just one dimension of the differences. There was another discussion is
what level of complexity do we want to allow and define within that?
... I'm a little concerned that we're having the same discussion as we had in RDB2RDF; I'm a bit worried we're just repeating the same arguments.
... Before going beyond that, I'd like to have an understanding on how the RDF conversions are done in the use cases. There are only 2-3 that really rely on an RDF mapping. R2RML can be quite complex, with a full SQL language inside. It's a level of complexity I'm quite afraid of.
... It's a kind of difference between the proposals I'd like to examine.
... Mappings of URIs and properties and much complexity.
<JeniT> +1 on defining the RDF output without dictating a particular serialisation of that RDF
<JeniT> I think there are layers: direct mapping using no metadata, mapping to RDF graph using metadata, mapping to RDF syntax using metadata
jenit: a clear document that says we need a decision would help focus discussion.
danbri: we should just say we're choosing one, say left to write, top to bottom.
jenit: also, commas are used as syntactic marker, and not as text.
ivan: I had a conversation with Richard, our
i18n guy; the best way to do that would be to contact the i18n mailing
list and ask them to look at use cases to see if there's something too
latin-biased.
... apart from that, we should try to collect use cases outside of the US-Europe world.
ivan: I can try to reach out to Chinese colleagues, or google has some aribic people.
yakovsh: I'm a hebrew speaker, but I've never seen a hebrew CSV, but I'll poke around.
<danbri> can we add "We particularly seek feedback and suggestions on the Internationalization aspects of this work" to the Status section?
ivan: next time, I don't want to touch it now, it's in the webmaster's control. | http://www.w3.org/2014/03/26-csvw-minutes.html | CC-MAIN-2016-30 | refinedweb | 2,127 | 64.61 |
12.10. Transposed Convolution¶
The layers we introduced so far for convolutional neural networks, including convolutional layers (Section 6.2) and pooling layers (Section 6.5), often reducethe input width and height, or keep them unchanged. Applications such as semantic segmentation (Section 12.9) and generative adversarial networks (Section 14.2), however, require to predict values for each pixel and therefore needs to increase input width and height. Transposed convolution, also named fractionally-strided convolution Dumoulin.Visin.2016 or deconvolution Long.Shelhamer.Darrell.2015, serves this purpose.
from mxnet import nd, init from mxnet.gluon import nn import d2l
12.10.1. Basic 2D Transposed Convolution¶
Let’s consider a basic case that both input and output channels are 1, with 0 padding and 1 stride. Fig. 12.10.1 illustrates how transposed convolution with a \(2\times 2\) kernel is computed on the \(2\times 2\) input matrix.
We can implement this operation by giving matrix kernel \(K\) and matrix input \(X\).
def trans_conv(X, K): h, w = K.shape Y = nd.zeros((X.shape[0] + h - 1, X.shape[1] + w - 1)) for i in range(X.shape[0]): for j in range(X.shape[1]): Y[i: i + h, j: j + w] += X[i, j] * K return Y
Remember the convolution computes results by
Y[i, j] = (X[i: i + h, j: j + w] * K).sum() (refer to
corr2d in
Section 6.2), which summarizes input values through
the kernel. While the transposed convolution broadcasts input values
through the kernel, which results in a larger output shape.
Verify the results in Fig. 12.10.1.
X = nd.array([[0,1], [2,3]]) K = nd.array([[0,1], [2,3]]) trans_conv(X, K)
[[ 0. 0. 1.] [ 0. 4. 6.] [ 4. 12. 9.]] <NDArray 3x3 @cpu(0)>
Or we can use
nn.Conv2DTranspose to obtain the same results. As
nn.Conv2D, both input and kernel should be 4-D tensors.
X, K = X.reshape((1, 1, 2, 2)), K.reshape((1, 1, 2, 2)) tconv = nn.Conv2DTranspose(1, kernel_size=2) tconv.initialize(init.Constant(K)) tconv(X)
[[[[ 0. 0. 1.] [ 0. 4. 6.] [ 4. 12. 9.]]]] <NDArray 1x1x3x3 @cpu(0)>
12.10.2. Padding, Strides, and Channels¶
We apply padding elements to the input in convolution, while they are applied to the output in transposed convolution. A \(1\times 1\) padding means we first compute the output as normal, then remove the first/last rows and columns.
tconv = nn.Conv2DTranspose(1, kernel_size=2, padding=1) tconv.initialize(init.Constant(K)) tconv(X)
[[[[4.]]]] <NDArray 1x1x1x1 @cpu(0)>
Similarly, strides are applied to outputs as well.
tconv = nn.Conv2DTranspose(1, kernel_size=2, strides=2) tconv.initialize(init.Constant(K)) tconv(X)
[[[[0. 0. 0. 1.] [0. 0. 2. 3.] [0. 2. 0. 3.] [4. 6. 6. 9.]]]] <NDArray 1x1x4x4 @cpu(0)>
The multi-channel extension of the transposed convolution is the same as the convolution. When the input has multiple channels, denoted by \(c_i\), the transposed convolution assigns a \(k_h\times k_w\) kernel matrix to each input channel. If the output has a channel size \(c_o\), then we have a \(c_i\times k_h\times k_w\) kernel for each output channel.
As a result, if we feed \(X\) into a convolutional layer \(f\) to compute \(Y=f(X)\) and create a transposed convolution layer \(g\) with the same hyper-parameters as \(f\) except for the output channel set to be the channel size of \(X\), then \(g(Y)\) should has the same shape as \(X\). Let’s verify this statement.
X = nd.random.uniform(shape=(1, 10, 16, 16)) conv = nn.Conv2D(20, kernel_size=5, padding=2, strides=3) tconv = nn.Conv2DTranspose(10, kernel_size=5, padding=2, strides=3) conv.initialize() tconv.initialize() tconv(conv(X)).shape == X.shape
True
12.10.3. Analogy to Matrix Transposition¶
The transposed convolution takes its name from the matrix transposition.
In fact, convolution operations can also be achieved by matrix
multiplication. In the example below, we define a \(3\times\) input
\(X\) with a \(2\times 2\) kernel \(K\), and then use
corr2d to compute the convolution output.
X = nd.arange(9).reshape((3,3)) K = nd.array([[0,1], [2,3]]) Y = d2l.corr2d(X, K) Y
[[19. 25.] [37. 43.]] <NDArray 2x2 @cpu(0)>
Next, we rewrite convolution kernel \(K\) as a matrix \(W\). Its shape will be \((4,9)\), where the \(i\)-th row present applying the kernel to the input to generate the \(i\)-th output element.
def kernel2matrix(K): k, W = nd.zeros(5), nd.zeros((4, 9)) k[:2], k[3:5] = K[0,:], K[1,:] W[0, :5], W[1, 1:6], W[2, 3:8], W[3, 4:] = k, k, k, k return W W = kernel2matrix(K) W
[[0. 1. 0. 2. 3. 0. 0. 0. 0.] [0. 0. 1. 0. 2. 3. 0. 0. 0.] [0. 0. 0. 0. 1. 0. 2. 3. 0.] [0. 0. 0. 0. 0. 1. 0. 2. 3.]] <NDArray 4x9 @cpu(0)>
Then the convolution operator can be implemented by matrix multiplication with proper reshaping.
Y == nd.dot(W, X.reshape((-1))).reshape((2,2))
[[1. 1.] [1. 1.]] <NDArray 2x2 @cpu(0)>
We can implement transposed convolution as a matrix multiplication as
well by reusing
kernel2matrix. To reuse the generated \(W\), we
construct a \(2\times 2\) input, so the corresponding weight matrix
will have a shape \((9,4)\), which is \(W^T\). Let’s verify the
results.
X = nd.array([[0,1], [2,3]]) Y = trans_conv(X, K) Y == nd.dot(W.T, X.reshape((-1))).reshape((3,3))
[[1. 1. 1.] [1. 1. 1.] [1. 1. 1.]] <NDArray 3x3 @cpu(0)>
12.10.4. Summary¶
Compared to convolutions that reduce inputs through kernels, transposed convolutions broadcast inputs.
If a convolution layer reduces the input width and height by \(n_w\) and \(h_h\) time, respectively. Then a transposed convolution layer with the same kernel sizes, padding and strides will increase the input width and height by \(n_w\) and \(n_h\), respectively.
We can implement convolution operations by the matrix multiplication, the corresponding transposed convolutions can be done by transposed matrix multiplication. | http://classic.d2l.ai/chapter_computer-vision/tranposed-conv.html | CC-MAIN-2020-16 | refinedweb | 1,021 | 59.5 |
This is your resource to discuss support topics with your peers, and learn from each other.
03-16-2009 06:52 AM
I have just started playing with 8800 simulator (on OS 4.5) and have found this strange behavior.
When user presses loudspeaker key ($ or € - in euro version) I get a popup
"Use this key to insert a dollar, pound..."
I don't want this to happen.
Why is it appearing, and how can I disable it ?
BTW, my screen extends MainScreen and implements KeyListener for handling key events.
Solved! Go to Solution.
03-16-2009 07:17 AM
Further research:
It never showed up in the Dialer !? How can that be ? How could the dialer disable that popup ? I want to do that also
.
But after I selected and option ($, €...) it never showed it self again.
How to disable/skip it in the first place ?
03-16-2009 07:36 AM
03-24-2009 10:14 AM
Thanks for the effort Simon but this is still not enough.
It's rather silly.
I implement key listener and the user presses the currency key (ludspeaker/$) and the popup goes on top.
The strange thing is that the listener is still active and since I consume most of the keys the popup remains on top until i close my screen (the listener unregisters)
So I am stuck ;(
I also tried to use the Keypad.hasCurrencyKey() to inform the user (or to prepare my app) of the problem, but that didn't work either. It always returned true.
Has anyone found a workaround?
(I hesitate to go to keyInjection techniques as I don't want to set the key to some value the user didn't choose. )
So I am kinda stuck to writing to the manual: "The currency key must be set before starting the application"
03-30-2009 08:32 AM
myraddin wrote:
The strange thing is that the listener is still active and since I consume most of the keys the popup remains on top until i close my screen (the listener unregisters)
I haven't been able to reproduce this part. What is the exact 4 digit 4.5.0.x version you are testing with? Can you post sonme sample code?
03-31-2009 03:27 AM
It's rather easy to reproduce.
I am testing in simulator v 2.9.0.52. OS 4.5.0.7 with a 8800 device.
But I have already tested on the real device on both 8800 (4.5.0) and Bold (4.6.0)
There are two ways to get this behavior:
1) Application derived app. pushing global screen which implements key listener.
2) Application derived app. implementing phone listener pushing global screen implementing key listener
The behaviour is somewhat different:
Here's the code:
(main app)
public class MainApp extends Application implements PhoneListener { private static MainApp app;
public static int SCENARIO_VERSION = 1;
public static int SCENARIO_NO_PHONE_LISTENER = 1;
public static int SCENARIO_PHONE_LISTENER = 2;
public static void main(String[] args) { app = new MainApp(); Phone.addPhoneListener(app);
if (MainApp.SCENARIO_VERSION == MainApp.SCENARIO_NO_PHONE_LISTENER)
{
Application.getApplication().invokeLater(new Runnable() { public void run() { // Scenario 1: Ui.getUiEngine().pushGlobalScreen(new SomeScreen(), 4, UiEngine.GLOBAL_SHOW_LOWER); } });
}
app.enterEventDispatcher(); } public void callInitiated(int arg0) { if (MainApp.SCENARIO_VERSION == MainApp.SCENARIO_PHONE_LISTENER)
{
// Scenario 2:
Ui.getUiEngine().pushGlobalScreen(new SomeScreen(), 4, UiEngine.GLOBAL_SHOW_LOWER);
}
} . . . }
(the popup to be shown)
final public class SomeScreen extends MainScreen implements KeyListener { public void sublayout(int width, int height) { super.sublayout(width, height); int overlayScreenSize = 200; setExtent(200, overlayScreenSize - 50); setPosition(50, Display.getHeight() - overlayScreenSize); } protected void paintBackground(Graphics g) { g.setBackgroundColor(0xFAFAD2); g.clear(); } protected void applyTheme() { // This popup doesn't want any theme influence } public boolean keyChar(char aKeyCaracter, int aKeyStatus, int aTime) { return false; } public boolean keyDown(int aKeyCode, int aTime) { System.out.println("KeyListener " + Integer.toHexString(aKeyCode)); if (aKeyCode == 0x1b0000) { // ESC key pressed if (MainApp.SCENARIO_VERSION == MainApp.SCENARIO_NO_PHONE_LISTENER) { // Scenario 1: System.exit(0); } else { // Scenario 2: Ui.getUiEngine().popScreen(this); } } return true; } public boolean keyUp(int arg0, int arg1) { return false; } public boolean keyRepeat(int arg0, int arg1) { return false; } public boolean keyStatus(int arg0, int arg1) { System.out.println("keyStatus " + Integer.toHexString(arg0)); return false; } }
Here are the results of scenario 1:
And this is scenario 2: (Make a call then press currency key)
The strangest thing is that in scenario 1 the popup isn't shown ! But if you press the currency key 5-10 times in a row, you will get a 'too many threads exception'.
The second anomaly is that my app never receives any key action event when the user presses the currency key in both scenarios while something is in fact sent. Why is the app's key listener being eluded ?
I don't want to go into RIM's details of implementation, the only thing that interests me is how to get the event that the currency key was pressed on a globally pushed screen
if it was not set previously. Or how to tell the system not to bother me with it while my screen is on top.
Thanks for your interest
P.S. I have played arround the screen's priorities and flags but could never achieve my final goal
04-10-2009 11:51 AM
You shouldn't implement KeyListener in a class that extends MainScreen. MainScreen already has its own KeyListener. This means that you now have 2 KeyListeners in the class, but are only overriding one set of KeyListener methods.
You can remove the second KeyListener implementation or implement the KeyListener in a different class, which you can then add to MainScreen.
04-30-2009 08:12 AM
Sorry for the long delay. I just noticed you post (how rude of me - I've been avoiding the currency key for a long time - now I must deal with it)
Once again, very informative help from you, thanks.
But unfortunately, the problem is not resolved.
I removed the ... implements KeyListener... part but the same problem is still present.
The 'Select currency key' dialog still pops up infront but doesn't catch the key inputs (everything is sent to my app).
How can I give control to the popup and later return to my screen ? (or how to remove the currency key popup in general ?)
Thanks again,
Myraddin
05-12-2009 04:36 PM
I have sent this issue to our development team. The popupscreen should appear on top of your screen automatically. | http://supportforums.blackberry.com/t5/Java-Development/Use-this-key-to-insert-a-dollar-pound-message/m-p/223494 | CC-MAIN-2015-48 | refinedweb | 1,070 | 57.77 |
I'm trying to execute a Python script from my Controller. In my controllers directory I created a folder called pythonScripts in which I have placed my script. From my controller function I am getting data from the database, in json format, and passing it to python to create a network and then do some analysis on the network using networkx. The problem is that i'm unable to execute the script. The following is my code for runing the script.
$output = array();
exec("/pythonScripts/createNetwork.py '{$connections}' ", $output);
return json_encode($output);
$connections
import json
import networkx as nx
import sys
from networkx.readwrite import json_graph
def main():
userJSONData = sys.argv[1] #the JSON format of user data will be passed into this script.
print userJSONData
if __name__ == '__main__':
main()
You must get the return value from your script execution to check if it was successful (== 0) and if your script is not executable, you must also call python:
exec("/usr/bin/python /pythonScripts/createNetwork.py '{$connections}'", $output, $return); if ($return) { throw new \Exception("Error executing command - error code: $return"); } var_dump($output); | https://codedump.io/share/QImrL4xffPlX/1/executing-python-script-from-laravel-4-controller | CC-MAIN-2017-26 | refinedweb | 182 | 56.15 |
go to bug id or search bugs for
New Comment:
Description:
------------
if you round with money_format, some small numbers with 0.005 are rounded down instead of up.
Numbers bigger than 63.5 are all rounded falsly down like with floor()
Test script:
---------------
for($i=0;$i<70;$i++) {
echo($i.'.005 =>'.money_format('%.2i',$i+0.005).'<br>');
if($i<10) echo($i.'.00500000000000001=>'.money_format('%.2i',$i+0.00500000000000001).'<br>');
}
Expected result:
----------------
0.005 =>0.01
0.00500000000000001=>0.01
1.005 =>1.01
1.00500000000000001=>1.01
2.005 =>2.01
2.00500000000000001=>2.01
3.005 =>3.01
3.00500000000000001=>3.01
4.005 =>4.01
4.00500000000000001=>4.01
5.005 =>5.01
5.00500000000000001=>5.01
6.005 =>6.01
6.00500000000000001=>6.01
7.005 =>7.01
7.00500000000000001=>7.01.01
...
Actual result:
--------------
0.00500000000000001=>0.01
1.005 =>1.00
1.00500000000000001=>1.01
2.005 =>2.00
2.00500000000000001=>2.00
3.005 =>3.00
3.00500000000000001=>3.00
4.005 =>4.00
4.00500000000000001=>4.00
5.005 =>5.00
5.00500000000000001=>5.00
6.005 =>6.00
6.00500000000000001=>6.00
7.005 =>7.00
7.00500000000000001=>7.00.00
17.005 =>17.00
18.005 =>18.00
19.005 =>19.00
20.005 =>20.00
21.005 =>21.00
22.005 =>22.00
23.005 =>23.00
24.005 =>24.00
25.005 =>25.00
26.005 =>26.00
27.005 =>27.00
28.005 =>28.00
29.005 =>29.00
30.005 =>30.00
31.005 =>31.00
32.005 =>32.01
33.005 =>33.01
34.005 =>34.01
35.005 =>35.01
36.005 =>36.01
37.005 =>37.01
38.005 =>38.01
39.005 =>39.01
40.005 =>40.01
41.005 =>41.01
42.005 =>42.01
43.005 =>43.01
44.005 =>44.01
45.005 =>45.01
46.005 =>46.01
47.005 =>47.01
48.005 =>48.01
49.005 =>49.01
50.005 =>50.01
51.005 =>51.01
52.005 =>52.01
53.005 =>53.01
54.005 =>54.01
55.005 =>55.01
56.005 =>56.01
57.005 =>57.01
58.005 =>58.01
59.005 =>59.01
60.005 =>60.01
61.005 =>61.01
62.005 =>62.01
63.005 =>63.01
64.005 =>64.00
65.005 =>65.00
66.005 =>66.00
67.005 =>67.00
68.005 =>68.00
69.005 =>69.00
from 64 on all numbers are rounded down like floor().
.
so why does php round 7.005 down to 7.00 while it rounds 8.005 correctly up to 8.01 ?
and why from 16 on again wrong and from 32 on again correct, from 64 on again wrong but now
not a change from 128 on?
Did you visit the site we provided that explains this? What you think is 7.005 is
actually represented as 7.004999999999999999 whereas 8.005 is
8.005000000000000000001
That's what limited precision means. Go read
for a complete explanation.
ah ok,
that explains the behaviour.
wouldnt it be senseful, if php would always floor the float variable then?
or at least throw a notice error, if you parse a float through "%.2i"?
i understand this is not a senseful programm-style. cause you get an unpredictable behaviour.
if money_format() would round down always, then it would be predictable
are there more functions that would show the same behaviour like in this example? what about printf()
How does always rounding down fix anything? What if you are trying to do 3.0 but
you actually get 2.999999999999999, if you always round down you would get 2
instead of 3.
now that i uderstand how floating numbers work in computers, this is sure a real bug then:
they say:
"a tiny error in the 17th decimal place doesn’t matter at all "
here the bug is at the 3rd place!
and:
" If you just don’t want to see all those extra decimal places: simply format your result rounded to a fixed number of decimal places when displaying it."
this is what i did here and that caused the bug
think about it: NO ONE wants this behaviour if he or she uses '%.2i' in a function
This function should be fixed, so it rounds mathematically correct like the round() function.
If not, then it should throw an error like
"cannot shorten a float variable with this syntax" or so.
Rounding errors compound when you loop like that. There is still no bug here.
money_format('%.2i',1.005) results in 1.00
and
money_format('%.2i',7.005) results in 7.01
this is definitely a bug. you only showed me, the reason, how the bug was created. but it should still be solved!
Or, if you say, the syntax is not correct, cause you provoke an unpredictable behaviour with '%.2i' then shouldnt that throw an error notice at least?
why cannot the php-supporters use the same algorythm, like in function round(), where there IS rounded correctly
(i meant: money_format('%.2i',8.005) results in 8.01)
This is getting a bit tiresome. There is no bug here.
Maybe you will understand it this way. Try running this:
<?php
ini_set('precision',32);
$ns = array(1.005, 7.005, 8.005);
foreach($ns as $n) {
echo "$n ".money_format('%.2i', $n)."\n";
}
The output on my machine is:
1.004999999999999893418589635985 1.00
7.004999999999999893418589635985 7.00
8.0050000000000007815970093361102 8.01
That is, 1.005 can't actually be represented accurately and it ends up being
slightly below 1.005 which means when you round it you get 1.00. And 8.005 ends
up being represented as slightly larger than 8.005 so when you round it you get
8.01. It makes perfect sense. Simply add a fuzz factor to your floating point
values to the appropriate precision you care about. Or, as most people know,
when dealing with money, don't use floating point at all. Work in integers.
i am sorry, i didnt want to do any harm.
i perfectly understood it before already.
But what do you think about my suggestions?
1. a notice error
or
2. round it like the round() function
There is no efficient way to "notice the error"
If you want slow and accurate floating point manipulation it is available via the
arbitrary precision extensions like bcmath and gmp.
maybe an efficient way would be to see if the desired amount of digits is less than the amount of digits in the float.
in my example where i tried to show 1.005 with only 2 fractional digits, the notice could be like:
"Error: input float exceeds maximum number of fractional digits"
this would save thousands of hours of debugging worldwide.
i am so keen on this issue, cause it cost me already about 100 hours of work to find out what was the cause, why in my project my bills every now and then where not correctly rounded.
it is a stupid bug really hard to track and i am sure am not the only one that had this problem.
if php would throw a notice, that would save lots of hazzle for not-so-experiensed programmers.
how can you know, that you shouldn't use floats in money_format. this is a kind of secret knowledge right now especially cause the manual says: "string money_format ( string $format , float $number )"
No, that won't work. Often the value comes from expressions that always generate
lots of digits. eg. number_format('%.2i', 1/3)
What would you expect that to do? You wold get a notice every single time even
though there may not be any relevant loss of precision.
There are just certain things you need to eventually learn when you start
programming.
how does the function round(7.005,2) do the trick to get 7.01? is it so much slower than how money_format does it?
could you please set the Status back to "Bug"
all people i talked to, said, that this should be solved. cause it causes follow-up bugs in your applications that are very hard to find, cause they happen so seldom
The same bug is in
sprintf("%01.2f",7.005);
which results in
7.00
while
sprintf("%01.2f",8.005);
results in
8.01
Please stop. This is not a PHP-specific issue. This is just how floating point
works in computers. You will need to learn how to deal with it. For example,
compile this C program:
#include <stdio.h>
int main(char *argv[], int argc) {
printf("%01.2f\n",7.005);
printf("%01.2f\n",8.005);
}
cc f.c -o f
then run it:
./f
you get:
7.00
8.01
Which is exactly the same as the PHP results. PHP is a thin layer over C. | https://bugs.php.net/bug.php?id=61787&edit=2 | CC-MAIN-2018-30 | refinedweb | 1,483 | 84.68 |
...one of the most highly
regarded and expertly designed C++ library projects in the
world. — Herb Sutter and Andrei
Alexandrescu, C++
Coding Standards
There is, at the moment, one back-end. This back-end contains the library engine and defines the performance and functionality trade-offs. The currently available back-end implements most of the functionality defined by the UML 2.0 standard at very high runtime speed, in exchange for longer compile-time. The runtime speed is due to a constant-time double-dispatch and self-adapting capabilities allowing the framework to adapt itself to the features used by a given concrete state machine. All unneeded features either disable themselves or can be manually disabled. See section 5.1 for a complete description of the run-to-completion algorithm.
MSM being divided between front and back-end, one needs to first define a front-end. Then, to create a real state machine, the back-end must be declared:
typedef msm::back::state_machine<my_front_end> my_fsm;
We now have a fully functional state machine type. The next sections will describe what can be done with it.
The
start() method starts the state machine, meaning it will
activate the initial state, which means in turn that the initial state's
entry behavior will be called. We need the start method because you do not
always want the entry behavior of the initial state to be called immediately
but only when your state machine is ready to process events. A good example
of this is when you use a state machine to write an algorithm and each loop
back to the initial state is an algorithm call. Each call to start will make
the algorithm run once. The iPodSearch example uses this possibility.
The
stop() method works the same way. It will cause the exit
actions of the currently active states(s) to be called.
Both methods are actually not an absolute need. Not calling them will simply cause your first entry or your last exit action not to be called.
The main reason to exist for a state machine is to dispatch events. For MSM, events are objects of a given event type. The object itself can contain data, but the event type is what decides of the transition to be taken. For MSM, if some_event is a given type (a simple struct for example) and e1 and e2 concrete instances of some_event, e1 and e2 are equivalent, from a transition perspective. Of course, e1 and e2 can have different values and you can use them inside actions. Events are dispatched as const reference, so actions cannot modify events for obvious side-effect reasons. To dispatch an event of type some_event, you can simply create one on the fly or instantiate if before processing:
my_fsm fsm; fsm.process_event(some_event()); some_event e1; fsm.process_event(e1)
Creating an event on the fly will be optimized by the compiler so the performance will not degrade.
The backend also offers a way to know which state is active, though you will normally only need this for debugging purposes. If what you need simply is doing something with the active state, internal transitions or visitors are a better alternative. If you need to know what state is active, const int* current_state() will return an array of state ids. Please refer to the internals section to know how state ids are generated..
Sometimes, one needs to customize states to avoid repetition and provide a common functionality, for example in the form of a virtual method. You might also want to make your states polymorphic so that you can call typeid on them for logging or debugging. It is also useful if you need a visitor, like the next section will show. You will notice that all front-ends offer the possibility of adding a base type. Note that all states and state machines must have the same base state, so this could reduce reuse. For example, using the basic front end, you need to:
Add the non-default base state in your msm::front::state<> definition, as first template argument (except for interrupt_states for which it is the second argument, the first one being the event ending the interrupt), for example, my_base_state being your new base state for all states in a given state machine:
struct Empty : public msm::front::state<my_base_state>
Now, my_base_state is your new base state. If it has a virtual
function, your states become polymorphic. MSM also provides a
default polymorphic base type,
msm::front::polymorphic_state
Add the user-defined base state in the state machine frontend definition, as a second template argument, for example:
struct player_ : public msm::front::state_machine<player_,my_base_state>
You can also ask for a state with a given id (which you might have gotten
from current_state()) using
const base_state* get_state_by_id(int id)
const where base_state is the one you just defined. You can now
do something polymorphically.
In some cases, having a pointer-to-base of the currently active states is not enough. You might want to call non-virtually a method of the currently active states. It will not be said that MSM forces the virtual keyword down your throat!
To achieve this goal, MSM provides its own variation of a visitor pattern
using the previously described user-defined state technique. If you add to
your user-defined base state an
accept_sig typedef giving the
return value (unused for the moment) and parameters and provide an accept
method with this signature, calling visit_current_states will cause accept
to be called on the currently active states. Typically, you will also want
to provide an empty default accept in your base state in order in order not
to force all your states to implement accept. For example your base state
could be:
struct my_visitable_state { // signature of the accept function typedef args<void> accept_sig; // we also want polymorphic states virtual ~my_visitable_state() {} // default implementation for states who do not need to be visited void accept() const {} };
This makes your states polymorphic and visitable. In this case, accept is made const and takes no argument. It could also be:
struct SomeVisitor {…}; struct my_visitable_state { // signature of the accept function typedef args<void,SomeVisitor&> accept_sig; // we also want polymorphic states virtual ~my_visitable_state() {} // default implementation for states who do not need to be visited void accept(SomeVisitor&) const {} };
And now,
accept will take one argument (it could also be
non-const). By default,
accept takes up to 2 arguments. To get
more, set #define BOOST_MSM_VISITOR_ARG_SIZE to another value before
including state_machine.hpp. For example:
#define BOOST_MSM_VISITOR_ARG_SIZE 3 #include <boost/msm/back/state_machine.hpp>
Note that accept will be called on ALL active states and also automatically on sub-states of a submachine.
Important warning: The method visit_current_states takes its parameter by value, so if the signature of the accept function is to contain a parameter passed by reference, pass this parameter with a boost:ref/cref to avoid undesired copies or slicing. So, for example, in the above case, call:
SomeVisitor vis; sm.visit_current_states(boost::ref(vis));
This example uses a visiting function with 2 arguments.
Flags is a MSM-only concept, supported by all front-ends, which base themselves on the functions:
template <class Flag> bool is_flag_active() template <class Flag,class BinaryOp> bool is_flag_active()
These functions return true if the currently active state(s) support the Flag property. The first variant ORs the result if there are several orthogonal regions, the second one expects OR or AND, for example:
my_fsm.is_flag_active<MyFlag>() my_fsm.is_flag_active<MyFlag,my_fsm_type::Flag_OR>()
Please refer to the front-ends sections for usage examples.
It is sometimes necessary to have the client code get access to the states' data. After all, the states are created once for good and hang around as long as the state machine does so why not use it? You simply just need sometimes to get information about any state, even inactive ones. An example is if you want to write a coverage tool and know how many times a state was visited. To get a state, use the get_state method giving the state name, for example:
player::Stopped* tempstate = p.get_state<player::Stopped*>();
or
player::Stopped& tempstate2 = p.get_state<player::Stopped&>();
depending on your personal taste.
You might want to define a state machine with a non-default constructor. For example, you might want to write:
struct player_ : public msm::front::state_machine_def<player_> { player_(int some_value){…} };
This is possible, using the back-end as forwarding object:
typedef msm::back::state_machine<player_ > player; player p(3);
The back-end will call the corresponding front-end constructor upon creation.
You can pass arguments up to the value of the BOOST_MSM_CONSTRUCTOR_ARG_SIZE macro (currently 5) arguments. Change this value before including any header if you need to overwrite the default.
You can also pass arguments by reference (or const-reference) using boost::ref (or boost::cref):
struct player_ : public msm::front::state_machine_def<player_> { player_(SomeType& t, int some_value){…} }; typedef msm::back::state_machine<player_ > player; SomeType data; player p(boost::ref(data),3);.
MSM is optimized for run-time speed at the cost of longer compile-time. This can become a problem with older compilers and big state machines, especially if you don't really care about run-time speed that much and would be satisfied by a performance roughly the same as most state machine libraries. MSM offers a back-end policy to help there. But before you try it, if you are using a VC compiler, deactivate the /Gm compiler option (default for debug builds). This option can cause builds to be 3 times longer... If the compile-time still is a problem, read further. MSM offers a policy which will speed up compiling in two main cases:
many transition conflicts
submachines
The back-end
msm::back::state_machine has a policy argument
(first is the front-end, then the history policy) defaulting to
favor_runtime_speed. To switch to
favor_compile_time, which is declared in
<msm/back/favor_compile_time.hpp>, you need to:
switch the policy to
favor_compile_time for the
main state machine (and possibly submachines)
move the submachine declarations into their own header which
includes
<msm/back/favor_compile_time.hpp>
add for each submachine a cpp file including your header and calling a macro, which generates helper code, for example:
#include "mysubmachine.hpp" BOOST_MSM_BACK_GENERATE_PROCESS_EVENT(mysubmachine)
configure your compiler for multi-core compilation
You will now compile your state machine on as many cores as you have submachines, which will greatly speed up the compilation if you factor your state machine into smaller submachines.
Independently, transition conflicts resolution will also be much faster.
This policy uses boost.any behind the hood, which means that we will lose a feature which MSM offers with the default policy, event hierarchy. The following example takes our iPod example and speeds up compile-time by using this technique. We have:
our main state machine and main function
PlayingMode moved to a separate header
MenuMode moved to a separate header
events move to a separate header as all machines use. | http://www.boost.org/doc/libs/1_55_0/libs/msm/doc/HTML/ch03s05.html | CC-MAIN-2013-48 | refinedweb | 1,829 | 52.29 |
Shapely Window in PyQTFri 27 May 2011
Recently I needed to make a PyQt app where the
window is the shape of an image and doesn’t have a border. I say needed, that’s not strictly true, more like wanted to because it’d be more interesting.
It wasn’t as straight forward as I had hoped but I managed to get it working in the end and here I am about to share it with the world. Here’s what we will need:
- python, Qt, and PyQt installed
- The image you want to use as the background with transparency. PNG will do nicely
- A texteditor
- A cup of coffee
Set the coffee aside so you don’t knock it over, but keep it within reach. First we need to open up Designer.
Once open create a new Dialog or MainWindow. Open the resource browser and add your background image to the list of resources. Now right click on the
window and select
Change styleSheet (I am using Designer4 in case it looks different for you). Enter the following in the popup, adjusting for names and path as appropriate.
#Dialog{ background-color: rgb(0, 0, 0); background-image: url(:/img/windowshape.png); }
Adjusting the
`#Dialog` to the
window name and the url path to your image.
You should now see the image in the
window with a black surround. At this point it’s also good to set the
window size to something suitable for your image.
As per usual you will need to compile your .ui and resource files with pyuic and pyrcc, but details on that are outside the scope of this post. So in order to remove the border we call this in our
__init__ function
self.win = QtGui.QMainWindow() self.win.setWindowFlags(self.win.windowFlags() | QtCore.Qt.FramelessWindowHint | QtCore.Qt.WindowSystemMenuHint)
It’s important to derive your class from
QtGui.QMainWindow otherwise this won’t work.
Don’t be surprised if under certain
window managers you still see a border, not all `window` managers will render windows without borders from
Qt. The settings are just a request rather than a demand. In Fluxbox for instance the border still shows. Gnome and KDE work fine.
Right, time for a sip of coffee. You earned it. But wait, what about the black colour? We want the
window in the shape of the image. Ah yes, we do that by adding this function
def resizeEvent(self, event): pixmap = QtGui.QPixmap(":/img/windowshape.png") region = QtGui.QRegion(pixmap.mask()) self.setMask(pixmap.mask());
This will tell
Qt to use the image as a mask for the `window` region, effectively hiding any parts where the image is transparent. More coffee, we're almost there. Right now, as you're swallowing that last sip, you're wondering "How do I move or close the
window without a border?". Fear not fellow coder for there is a solution for each of these:
To close the
window we just need to add a context menu. Simply done by adding one more function to your class:
def contextMenuEvent(self, event): menu = QtGui.QMenu(self) quitAction = menu.addAction("Quit") action = menu.exec_(self.mapToGlobal(event.pos())) if action == quitAction: self.close()
You can of course replace the
self.close() with a call to some confirmation dialog if you want, but that will now enable a right-click menu on your GUI with a quit option. Also you can add keyboard shortcuts to the application as well if you are so inclined.
Now for moving the thing. Here we need two extra functions, one for the mouse move and one for the mouse press:
def mouseMoveEvent(self, event): if (event.buttons() == QtCore.Qt.LeftButton): self.move(event.globalPos().x() - self.drag_position.x(), event.globalPos().y() - self.drag_position.y()); event.accept(); def mousePressEvent(self, event): if (event.button() == QtCore.Qt.LeftButton): self.drag_position* = event.globalPos() - self.pos(); event.accept();
We just get the click position when the mouse button is pressed, work out the offset from the top left corner of the
window (this is what the move function uses) and then when the mouse moves, we move the
window with it.
That's it really. Pretty simple once you know how. Now go and enjoy the rest of that coffee. | https://www.unlogic.co.uk/2011/05/27/shapely-window-in-pyqt/ | CC-MAIN-2018-22 | refinedweb | 714 | 66.23 |
Updating...
CodePlex is shutting down.
Read about the shutdown plan, including archive and migration information, on
Brian Harry's blog.
Code
Plex
Project Hosting for Open Source Software
This project has moved.
For the latest updates, please
go here
.
source code
downloads
documentation
discussions
issues
people
license
All Project Updates
Discussions
Issue Tracker
Downloads
Reviews
Source Code
Wiki & Documentation
Subscribe
0.1.0.0 Beta
Rating:
No reviews yet
Downloads:
2929
Released:
Dec 31, 2006
Updated:
Oct 2, 2007
by
crashlander
Dev status:
-not yet defined by owner-
Recommended Download
FarseerPhysicsSDK_0.1.0.0
application, 2383K, uploaded
Dec 31, 2006
- 2929 downloads
Release Notes
New version coming Oct 3rd, 2007. Don't bother with this version....
This release contains Demos 4,5,6,7,and 8. Only 8 is new, the others are just updated to all work with the 1.0 release of XNA and are included because they show some of the more basic operation of the engine.
In the root of the release folder you will find a change log describing all that has changed for this version. I'm including it below as well.
Have fun!
Farseer Physics Engine Change Log
Changes For Version 0.1.0.0
FarseerXNAGame
• Refactored RectangleEntity, PolygonEntity, and PointEntity to inherit from a common base class: PolygonEntityBase.
• Added CircleEntity to go along with the other entity types.
• Made all private class level variables use the naming convention: _name
• Added a “RigidBodyDiagnosticEntityView” object to the Entities namespace. This object creates a diagnostic view of a RigidBody. Currently, it shows the vertices of the RigidBody and the collision points. (See FarseerDemo8)
FarseerXNAPhysics
• Added CollisionEnabled boolean property to RigidBody.
• Added CollisionResponseEnabled boolean to RigidBody. This will allow collision info to be generated without the bodies interacting physically.
• Modified PolygonRigidBody, RectangleRigidBody, and CircleRigidBody. Hopefully, the process of creating a rigid body and setting the way its collision grid is constructed is a little cleaner. I added a ‘CollisionPrecisionType’. If this value is set to “Relative”, the collision precision value will be multiplied by the shortest side of the rigid body’s geometry and this value will be used for the grid cell size. If the collision precision type is set to “Absolute”, the collision precision parameter will be used directly for the grid cell size. Currently, the Rigid Body Entities in the FarseerXNAGame project all use the “Relative” setting, but this change will allow for more flexibility if desired.
• Removed the PhysicsSimulatorBase class. It really wasn’t being used.
• Made all PRIVATE class level variables use the naming convention: _name
• Added RemoveJoint and RemoveSpring methods to the PhysicsSimulator class.
• Added Reset method to the PhysicsSimulator class. This method simply clears out all the list objects: RigidBody, Joint, Spring.
• Added CollisionEvent to RigidBody. The CollisionEventArgs object that is returned when the event fires will return a list of contact points.
• Added AngularJoint. This acts to keep 2 bodies at a specified relative angle. It is usually used in conjunction with a RevoluteJoint. Very similar to AngularSpring, but stiffer.
• Added Collide(Vector2 point) method to RigidBody. This adds the ability to check collision of a point with a rigid body. The method returns a bool.
• Added a “CollisonPoint” object. This object contains a boolean and a pointer to a RigidBody. It is used for doing collision tests between a point and the physics world. (see next bullet item.)
• Added CollidePoint(Vector2 point) to the PhysicsSimulator class. This method will return a CollisionPoint object. If the CollisionPoint’s “IsCollision” property is true, then you can use the CollisionPoint.RigidBody property to determine what body was collided with.
• The above two bullet points give the ability to do mouse dragging of rigid bodies. I’ve implemented a version of this in FarseerDemo8.
Reviews for this release
No reviews yet for this release.
Opera does not support ClickOnce
X
To install this application, save it and then open it. Opening it directly from Opera will not work correctly.
Other Downloads
Released
|
Planned
Farseer Physics Engine 3.5
Aug 26, 2013
, Stable
Farseer Physics Engine 3.3.1
Apr 9, 2011
, Stable
Farseer Physics Engine 3.3
Mar 19, 2011
, Stable
Farseer Physics Engine 3.2
Jan 1, 2011
, Stable
Farseer Physics Engine 3.1
Nov 12, 2010
, Stable
Farseer Physics Engine 3.0
Aug 22, 2010
, Stable
Farseer Physics Engine 2.1.3
Nov 5, 2009
, Stable
Farseer Physics Engine 2.1.2
Oct 25, 2009
, Stable.
Release notifications
to display notification settings.
X
(change e-mail address)
Unsubscribe
Also stop notifications for
individual
issue(s) I subscribed to.
© 2006-2017 Microsoft
Get Help
Privacy Statement
Code of Conduct
Advertise With Us
Version 3.28.2017.21050 | http://farseerphysics.codeplex.com/releases/view/1420 | CC-MAIN-2017-22 | refinedweb | 777 | 60.11 |
Writing Server-rendered React Apps with Next.js
Free JavaScript Book!
Write powerful, clean and maintainable JavaScript.
RRP $11.95
The dust has settled a bit as far as the JavaScript front-end ecosystem is concerned. React has arguably the biggest mindshare at this point, but has a lot of bells and whistles you need to get comfortable with. Vue offers a considerably simpler alternative. And then there are Angular and Ember — which, while still popular, are not the first recommendations for starting a new project.
So, while React is the most popular option, it still requires lot of tooling to write production-ready apps. Create React App solves many of the pain points of starting, but the jury is still out on how long you can go without ejecting. And when you start looking into the current best practices around front-end, single-page applications (SPAs) — like server-side rendering, code splitting, and CSS-in-JS — it’s a lot to find your way through.
That’s where Next comes in.
Why Next?
Next provides a simple yet customizable solution to building production-ready SPAs. Remember how web apps were built in PHP (before “web apps” was even a term)? You create some files in a directory, write your script and you’re good to deploy. That’s the kind of simplicity Next aims at, for the JavaScript ecosystem.
Next is not a brand new framework per se. It fully embraces React, but provides a framework on top of that for building SPAs while following best practices. You still write React components. Anything you can do with Next, you can do with a combination of React, Webpack, Babel, one of 17 CSS-in-JS alternatives, lazy imports and what not. But while building with Next, you aren’t thinking about which CSS-in-JS alternative to use, or how to set up Hot Module Replacement (HMR), or which of many routing options to choose. You’re just using Next — and it just works.
I’d like to think I know a thing or two about JavaScript, but Next.JS saves me an ENORMOUS amount of time. — Eric Elliott
Getting Started
Next requires minimal setup. This gets you all the dependencies you need for starting:
$ npm install next react react-dom --save
Create a directory for your app, and inside that create a directory called
pages. The file system is the API. Every
.js file becomes a route that gets automatically processed and rendered.
Create a file
./pages/index.js inside your project with these contents:
export default () => ( <div>Hello, Next!</div> )
Populate
package.json inside your project with this:
{ "scripts": { "dev": "next", "build": "next build", "start": "next start" } }
Then just run
npm run dev in the root directory of your project. Go to and you should be able to see your app, running in all its glory!
Just with this much you get:
- automatic transpilation and bundling (with Webpack and Babel)
- Hot Module Replacement
- server-side rendering of
./pages
- static file serving:
./static/is mapped to
/static/.
Good luck doing that with Vanilla React with this much setup!
Features
Let’s dig into some of the features of modern SPA apps, why they matter, and how they just work with Next.
Automatic code splitting
Why it Matters?
Code splitting is important for ensuring fast time to first meaningful paint. It’s not uncommon to have JavaScript bundle sizes reaching up to several megabytes these days. Sending all that JavaScript over the wire for every single page is a huge waste of bandwidth.
How to get it with Next
With Next, only the declared imports are served with each page. So, let’s say you have 10 dependencies in your
package.json, but
./pages/index.js only uses one of them.
pages/login.js
import times from 'lodash.times' export default () => ( return <div>times(5, <h2> Hello, there! </h2>)</div>; )
Now, when the user opens the login page, it’s not going to load all the JavaScript, but only the modules required for this page.
So a certain page may have fat imports, like this:
import React from 'react' import d3 from 'd3' import jQuery from 'jquery'
But this won’t affect the performance of the rest of the pages. Faster load times FTW.
Scoped CSS
Why it Matters?
CSS rules, by default, are global. Say you have a CSS rule like this:
.title { font-size: 40px; }
Now, you might have two components,
Profile, both of which may have a div with class
title. The CSS rule you defined is going to apply to both of them. So, you define two rules now, one for selector
.post .title, the other for
.profile .title. It’s manageable for small apps, but you can only think of so many class names.
Scoped CSS lets you define CSS with components, and those rules apply to only those components, making sure that you’re not afraid of unintended effects every time you touch your CSS.
With Next
Next comes with styled-jsx, which provides support for isolated scoped CSS. So, you just have a
<style> component inside your React Component render function:
export default () => ( <div> Hello world <p>These colors are scoped!</p> <style jsx>{\ p { color: blue; } div { background: red; } `}</style> </div> )
You also get the colocation benefits on having the styling (CSS), behavior (JS), and the template (JSX) all in one place. No more searching for the relevant class name to see what styles are being applied to it.
Dynamic Imports
Why it matters?
Dynamic imports let you dynamically load parts of a JavaScript application at runtime. There are several motivations for this, as listed in the proposal:
This could be because of factors only known at runtime (such as the user’s language), for performance reasons (not loading code until it is likely to be used), or for robustness reasons (surviving failure to load a non-critical module).
With Next
Next supports the dynamic import proposal and lets you split code into manageable chunks. So, you can write code like this that dynamically loads a React component after initial load:
import dynamic from 'next/dynamic' const DynamicComponentWithCustomLoading = dynamic( import('../components/hello2'), { loading: () => <p>The component is loading...</p> } ) export default () => <div> <Header /> <DynamicComponentWithCustomLoading /> <p>Main content.</p> </div>
Routing
Why it matters?
A problem with changing pages via JavaScript is that the routes don’t change with that. During their initial days, SPAs were criticized for breaking the web. These days, most frameworks have some robust routing mechanism. React has the widely used
react-router package. With Next, however, you don’t need to install a separate package.
With Next
Client-side routing can be enabled via a
next/link component. Consider these two pages:
// pages/index.js import Link from 'next/link' export default () => <div> Click{' '} <Link href="/contact"> <a>here</a> </Link>{' '} to find contact information. </div>
// pages/contact.js export default () => <p>The Contact Page.</p>
Not only that, you can add
prefetch prop to
Link component, to prefetch pages even before the links are clicked. This enables super-fast transition between routes.
Server rendering
Most of the JavaScript-based SPAs just don’t work without JavaScript disabled. However, it doesn’t have to be that way. Next renders pages on the server, and they can be loaded just like good old rendered web pages when JavaScript is disabled. Every component inside the
pages directory gets server-rendered automatically and their scripts inlined. This has the added performance advantage of very fast first loads, since you can just send a rendered page without making additional HTTP requests for the JavaScript files.
Next Steps
That should be enough to get you interested in Next, and if you’re working on a web app, or even an Electron-based application, Next provides some valuable abstractions and defaults to build upon.
To learn more about Next, Learning Next.js is an excellent place to start, and may be all you’ll need.
Get practical advice to start your career in programming!
Master complex transitions, transformations and animations in CSS! | https://www.sitepoint.com/writing-server-rendered-react-apps-next-js/ | CC-MAIN-2021-04 | refinedweb | 1,348 | 65.83 |
Adds GraphQL support to your Flask application
Project description
Adds GraphQL support to your Flask application.
Usage
Just use the GraphQLView view from flask_graphql
from flask_graphql import GraphQLView app.add_url_rule('/graphql', view_func=GraphQLView.as_view('graphql', schema=schema, graphiql=True)) # Optional, for adding batch query support (used in Apollo-Client) app.add_url_rule('/graphql/batch', view_func=GraphQLView.as_view('graphql', schema=schema, batch=True))
This will add /graphql and /graphiql endpoints to your app.
Supported options
- schema: The GraphQLSchema object that you want the view to execute when it gets a valid request.
- context: A value to pass as the context to the graphql() function.
- root_value: The root_value you want to provide to executor.execute.
- pretty: Whether or not you want the response to be pretty printed JSON.
- executor: The Executor that you want to use to execute queries.
- graphiql: If True, may present GraphiQL when loaded directly from a browser (a useful tool for debugging and exploration).
- graphiql_template: Inject a Jinja template string to customize GraphiQL.
- batch: Set the GraphQL view as batch (for using in Apollo-Client or ReactRelayNetworkLayer)
You can also subclass GraphQLView and overwrite get_root_value(self, request) to have a dynamic root value per request.
class UserRootValue(GraphQLView): def get_root_value(self, request): return request.user
Project details
Release history Release notifications
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages. | https://pypi.org/project/Flask-GraphQL/ | CC-MAIN-2019-30 | refinedweb | 233 | 57.27 |
Results 1 to 5 of 5
Thread: C++, hello world?
C++, hello world?PHP Code:
#include <iostream>
int main()
{
std::cout << "Hello world!\n";
return 0;
}
I try to compile it and I get the following error
c:/cpp/hello.cpp:7:2: Warning: no newline at end of file
- Join Date
- Jun 2002
- 66
- Thanks
- 0
- Thanked 0 Times in 0 Posts
For some reason you need a newline at the end of file to compile. Not an error, just a warning. Compilers are picky. Just press enter after the closing brace.
That's the most stupid thing I've ever known lol Spent ages trying to figure it out too
ta for that
Bah, ok it's compiled into an *.exe now, but if I try to run it from dos, it's too big to fit in memory?
so when I try to run it by clicking on the *.exe, it shows a popup saying that it's not a valid win32 file
Is this just my compiler trying to wind me up and playing silly buggers with me, or is c++ not meant to work on windows?
nm, fixed it. I needed the windows.h file thingy | http://www.codingforums.com/computer-programming/17297-c-hello-world.html | CC-MAIN-2017-26 | refinedweb | 201 | 84.17 |
SchedGet(), SchedGet_r()
Get the scheduling policy for a thread
Synopsis:
#include <sys/neutrino.h> int SchedGet( pid_t pid, int tid, struct sched_param *param ); int SchedGet_r( pid_t pid, int tid, struct sched_param *param );
Arguments:
- pid
- 0 or a process ID; see below.
- tid
- 0 or a thread ID; see below.
- param
- A pointer to a sched_param structure where the function can store the scheduling parameters.
Library:
libc
Use the -l c option to qcc to link against this library. This library is usually included automatically.
Description:
The SchedGet() and SchedGet_r() kernel calls return the current scheduling policy and.
- Instead of using these kernel calls directly, consider calling pthread_getschedparam().
- In order to get the scheduling policy for a process whose real or saved user ID is different from the calling process's real or effective user ID, your process must have the PROCMGR_AID_SCHEDULE ability enabled. For more information, see procmgr_ability().
The scheduling policy is returned on success and is one of SCHED_FIFO, SCHED_RR, SCHED_SPORADIC, or SCHED_OTHER.
Blocking states
These calls don't block.
Returns:
The only difference between these functions is the way they indicate errors:
Errors:
- EFAULT
- A fault occurred when the kernel tried to access the buffers provided.
- EPERM
- The calling process doesn't have the required permission; see procmgr_ability().
- ESRCH
- The process indicated by pid or thread indicated by tid doesn't exist.
Classification:
Last modified: 2013-09-30 | http://developer.blackberry.com/native/reference/core/com.qnx.doc.neutrino.lib_ref/topic/s/schedget.html | CC-MAIN-2013-48 | refinedweb | 230 | 57.37 |
Building A Web App With React, Redux And Sanity.io
The fast evolution of digital platforms have placed serious limitations on traditional CMS like Wordpress. These platforms are coupled, inflexible and are focused on the project, rather than the product. Thankfully, several headless CMS have been developed to tackle these challenges and many more.
Unlike traditional CMS, headless CMS, which can be described as Software as a Service (SaaS), can be used to develop websites, mobile apps, digital displays, and many more. They can be used on limitless platforms. If you are looking for a CMS that is platform independent, developer-first, and offers cross platform support, you need not look farther from headless CMS.
A headless CMS is simply a CMS without a head. The
head here refers to the frontend or the presentation layer while the
body refers to the backend or the content repository. This offers a lot of interesting benefits. For instance, it allows the developer to choose any frontend of his choice and you can also design the presentation layer as you want.
There are lots of headless CMS out there, some of the most popular ones include Strapi, Contentful, Contentstack, Sanity, Butter CMS, Prismic, Storyblok, Directus, etc. These headless CMS are API-based and have their individual strong points. For instance, CMS like Sanity, Strapi, Contentful, and Storyblok are free for small projects.
These headless CMS are based on different tech stacks as well. While Sanity.io is based on React.js, Storyblok is based on Vue.js. As a React developer, this is the major reason why I quickly picked interest in Sanity. However, being a headless CMS, each of these platforms can be plugged on any frontend, whether Angular, Vue or React.
Each of these headless CMS has both free and paid plans which represent significant price jump. Although these paid plans offer more features, you wouldn’t want to pay all that much for a small to mid-sized project. Sanity tries to solve this problem by introducing pay-as-you-go options. With these options, you will be able to pay for what you use and avoid the price jump.
Another reason why I choose Sanity.io is their GROQ language. For me, Sanity stands out from the crowd by offering this tool. Graphical-Relational Object Queries (GROQ) reduces development time, helps you get the content you need in the form you need it, and also helps the developer to create a document with a new content model without code changes.
Moreover, developers are not constrained to the GROQ language. You can also use GraphQL or even the traditional
axios and
fetch in your React app to query the backend. Like most other headless CMS, Sanity has comprehensive documentation that contains helpful tips to build on the platform.
Note: This article requires a basic understanding of React, Redux and CSS.
Getting Started With Sanity.io
To use Sanity in your machine, you’ll need to install the Sanity CLI tool. While this can be installed locally on your project, it is preferable to install it globally to make it accessible to any future applications.
To do this, enter the following commands in your terminal.
npm install -g @sanity/cli
The
-g flag in the above command enables global installation.
Next, we need to initialize Sanity in our application. Although this can be installed as a separate project, it is usually preferable to install it within your frontend app (in this case React).
In her blog, Kapehe explained in detail how to integrate Sanity with React. It will be helpful to go through the article before continuing with this tutorial.
Enter the following commands to initialize Sanity in your React app.
sanity init
The
sanity command becomes available to us when we installed the Sanity CLI tool. You can view a list of the available Sanity commands by typing
sanity or
sanity help in your terminal.
When setting up or initializing your project, you’ll need to follow the prompts to customize it. You’ll also be required to create a dataset and you can even choose their custom dataset populated with data. For this listing app, we will be using Sanity’s custom sci-fi movies dataset. This will save us from entering the data ourselves.
To view and edit your dataset,
cd to the Sanity subdirectory in your terminal and enter
sanity start. This usually runs on. You may be required to login to access the interface (make sure you login with the same account you used when initializing the project). A screenshot of the environment is shown below.
Sanity-React Two-way Communication
Sanity and React need to communicate with each other for a fully functional application.
CORS Origins Setting In Sanity Manager
We’ll first connect our React app to Sanity. To do this, login to and locate
CORS origins under
API Settings in the
Settings tab. Here, you’ll need to hook your frontend origin to the Sanity backend. Our React app runs on by default, so we need to add that to the CORS.
This is shown in the figure below.
Connecting Sanity To React
Sanity associates a
project ID to every project you create. This ID is needed when connecting it to your frontend application. You can find the project ID in your Sanity Manager.
The backend communicates with React using a library known as
sanity client. You need to install this library in your Sanity project by entering the following commands.
npm install @sanity/client
Create a file
sanitySetup.js (the filename does not matter), in your project
src folder and enter the following React codes to set up a connection between Sanity and React.
import sanityClient from "@sanity/client" export default sanityClient({ projectId: PROJECT_ID, dataset: DATASET_NAME, useCdn: true });
We passed our
projectId,
dataset name and a boolean
useCdn to the instance of the sanity client imported from
@sanity/client. This works the magic and connects our app to the backend.
Now that we’ve completed the two-way connection, let’s jump right in to build our project.
Setting Up And Connecting Redux To Our App
We’ll need a few dependencies to work with Redux in our React app. Open up your terminal in your React environment and enter the following bash commands.
npm install redux react-redux redux-thunk
Redux is a global state management library that can be used with most frontend frameworks and libraries such as React. However, we need an intermediary tool
react-redux to enable communication between our Redux store and our React application. Redux thunk will help us to return a function instead of an action object from Redux.
While we could write the entire Redux workflow in one file, it is often neater and better to separate our concerns. For this, we will divide our workflow into three files namely,
actions,
reducers, and then the
store. However, we also need a separate file to store the
action types, also known as
constants.
Setting Up The Store
The store is the most important file in Redux. It organizes and packages the states and ships them to our React application.
Here is the initial setup of our Redux store needed to connect our Redux workflow.
import { createStore, applyMiddleware } from "redux"; import thunk from "redux-thunk"; import reducers from "./reducers/"; export default createStore( reducers, applyMiddleware(thunk) );
The
createStore function in this file takes three parameters: the
reducer (required), the initial state and the enhancer (usually a middleware, in this case,
thunk supplied through
applyMiddleware). Our reducers will be stored in a
reducers folder and we’ll combine and export them in an
index.js file in the
reducers folder. This is the file we imported in the code above. We’ll revisit this file later.
Introduction To Sanity’s GROQ Language
Sanity takes querying on JSON data a step further by introducing GROQ. GROQ stands for Graph-Relational Object Queries. According to Sanity.io, GROQ is a declarative query language designed to query collections of largely schema-less JSON documents.
Sanity even provides the GROQ Playground to help developers become familiar with the language. However, to access the playground, you need to install sanity vision.
Run
sanity install @sanity/vision on your terminal to install it.
GROQ has a similar syntax to GraphQL but it is more condensed and easier to read. Furthermore, unlike GraphQL, GROQ can be used to query JSON data.
For instance, to retrieve every item in our movie document, we’ll use the following GROQ syntax.
*[_type == "movie"]
However, if we wish to retrieve only the
_ids and
crewMembers in our movie document. We need to specify those fields as follows.
`*[_type == 'movie']{ _id, crewMembers }
Here, we used
* to tell GROQ that we want every document of
_type movie.
_type is an attribute under the movie collection. We can also return the type like we did the
_id and
crewMembers as follows:
*[_type == 'movie']{ _id, _type, crewMembers }
We’ll work more on GROQ by implementing it in our Redux actions but you can check Sanity.io’s documentation for GROQ to learn more about it. The GROQ query cheat sheet provides a lot of examples to help you master the query language.
Setting Up Constants
We need constants to track the action types at every stage of the Redux workflow. Constants help to determine the type of action dispatched at each point in time. For instance, we can track when the API is loading, fully loaded and when an error occurs.
We don’t necessarily need to define constants in a separate file but for simplicity and clarity, this is usually the best practice in Redux.
By convention, constants in Javascript are defined with uppercase. We’ll follow the best practices here to define our constants. Here is an example of a constant for denoting requests for moving movie fetching.
export const MOVIE_FETCH_REQUEST = "MOVIE_FETCH_REQUEST";
Here, we created a constant
MOVIE_FETCH_REQUEST that denotes an action type of
MOVIE_FETCH_REQUEST. This helps us to easily call this action type without using
strings and avoid bugs. We also exported the constant to be available anywhere in our project.
Similarly, we can create other constants for fetching action types denoting when the request succeeds or fails. A complete code for the
movieConstants.js is given in the code below.
export const MOVIE_FETCH_REQUEST = "MOVIE_FETCH_REQUEST"; export const MOVIE_FETCH_SUCCESS = "MOVIE_FETCH_SUCCESS"; export const MOVIE_FETCH_FAIL = "MOVIE_FETCH_FAIL"; export const MOVIES_FETCH_REQUEST = "MOVIES_FETCH_REQUEST"; export const MOVIES_FETCH_SUCCESS = "MOVIES_FETCH_SUCCESS"; export const MOVIES_FETCH_FAIL = "MOVIES_FETCH_FAIL"; export const MOVIES_FETCH_RESET = "MOVIES_FETCH_RESET"; export const MOVIES_REF_FETCH_REQUEST = "MOVIES_REF_FETCH_REQUEST"; export const MOVIES_REF_FETCH_SUCCESS = "MOVIES_REF_FETCH_SUCCESS"; export const MOVIES_REF_FETCH_FAIL = "MOVIES_REF_FETCH_FAIL"; export const MOVIES_SORT_REQUEST = "MOVIES_SORT_REQUEST"; export const MOVIES_SORT_SUCCESS = "MOVIES_SORT_SUCCESS"; export const MOVIES_SORT_FAIL = "MOVIES_SORT_FAIL"; export const MOVIES_MOST_POPULAR_REQUEST = "MOVIES_MOST_POPULAR_REQUEST"; export const MOVIES_MOST_POPULAR_SUCCESS = "MOVIES_MOST_POPULAR_SUCCESS"; export const MOVIES_MOST_POPULAR_FAIL = "MOVIES_MOST_POPULAR_FAIL";
Here we have defined several constants for fetching a movie or list of movies, sorting and fetching the most popular movies. Notice that we set constants to determine when the request is
successful and
failed.
Similarly, our
personConstants.js file is given below:
export const PERSONS_FETCH_REQUEST = "PERSONS_FETCH_REQUEST"; export const PERSONS_FETCH_SUCCESS = "PERSONS_FETCH_SUCCESS"; export const PERSONS_FETCH_FAIL = "PERSONS_FETCH_FAIL"; export const PERSON_FETCH_REQUEST = "PERSON_FETCH_REQUEST"; export const PERSON_FETCH_SUCCESS = "PERSON_FETCH_SUCCESS"; export const PERSON_FETCH_FAIL = "PERSON_FETCH_FAIL"; export const PERSONS_COUNT = "PERSONS_COUNT";
Like the
movieConstants.js, we set a list of constants for fetching a person or persons. We also set a constant for counting persons. The constants follow the convention described for
movieConstants.js and we also exported them to be accessible to other parts of our application.
Finally, we’ll implement light and dark mode in the app and so we have another constants file
globalConstants.js. Let’s take a look at it.
export const SET_LIGHT_THEME = "SET_LIGHT_THEME"; export const SET_DARK_THEME = "SET_DARK_THEME";
Here we set constants to determine when light or dark mode is dispatched.
SET_LIGHT_THEME determines when the user switches to the light theme and
SET_DARK_THEME determines when the dark theme is selected. We also exported our constants as shown.
Setting Up The Actions
By convention, our actions are stored in a separate folder. Actions are grouped according to their types. For instance, our movie actions are stored in
movieActions.js while our person actions are stored in
personActions.js file.
We also have
globalActions.js to take care of toggling the theme from light to dark mode.
Let’s fetch all movies in
moviesActions.js.
import sanityAPI from "../../sanitySetup"; import { MOVIES_FETCH_FAIL, MOVIES_FETCH_REQUEST, MOVIES_FETCH_SUCCESS } from "../constants/movieConstants"; const fetchAllMovies = () => async (dispatch) => { try { dispatch({ type: MOVIES_FETCH_REQUEST }); const data = await sanityAPI.fetch( `*[_type == 'movie']{ _id, "poster": poster.asset->url, } ` ); dispatch({ type: MOVIES_FETCH_SUCCESS, payload: data }); } catch (error) { dispatch({ type: MOVIES_FETCH_FAIL, payload: error.message }); } };
Remember when we created the
sanitySetup.js file to connect React to our Sanity backend? Here, we imported the setup to enable us to query our sanity backend using GROQ. We also imported a few constants exported from the
movieConstants.js file in the
constants folder.
Next, we created the
fetchAllMovies action function for fetching every movie in our collection. Most traditional React applications use
axios or
fetch to fetch data from the backend. But while we could use any of these here, we’re using Sanity’s
GROQ. To enter the
GROQ mode, we need to call
sanityAPI.fetch() function as shown in the code above. Here,
sanityAPI is the React-Sanity connection we set up earlier. This returns a
Promise and so it has to be called asynchronously. We’ve used the
async-await syntax here, but we can also use the
.then syntax.
Since we are using
thunk in our application, we can return a function instead of an action object. However, we chose to pass the return statement in one line.
const fetchAllMovies = () => async (dispatch) => { ... }
Note that we can also write the function this way:
const fetchAllMovies = () => { return async (dispatch)=>{ ... } }
In general, to fetch all movies, we first dispatched an action type that tracks when the request is still loading. We then used Sanity’s GROQ syntax to asynchronously query the movie document. We retrieved the
_id and the poster url of the movie data. We then returned a payload containing the data gotten from the API.
Similarly, we can retrieve movies by their
_id, sort movies, and get the most popular movies.
We can also fetch movies that match a particular person’s reference. We did this in the
fetchMoviesByRef function. }); } };
This function takes an argument and checks if
person._ref in either the
castMembers or
crewMembers matches the passed argument. We return the movie
_id,
poster url, and
title alongside. We also dispatch an action of type
MOVIES_REF_FETCH_SUCCESS, attaching a payload of the returned data, and if an error occurs, we dispatch an action of type
MOVIE_REF_FETCH_FAIL, attaching a payload of the error message, thanks to the
try-catch wrapper.
In the
fetchMovieById function, we used
GROQ to retrieve a movie that matches a particular
id passed to the function.
The
GROQ syntax for the function is shown below.]` );
Like the
fetchAllMovies action, we started by selecting all documents of type
movie but we went further to select only those with an id supplied to the function. Since we intend to display a lot of details for the movie, we specified a bunch of attributes to retrieve.
We retrieved the movie
id and also a few attributes in the
castMembers array namely
ref,
characterName, the person’s name, and the person’s image. We also changed the alias from
castMembers to
cast.
Like the
castMembers, we selected a few attributes from the
crewMembers array, namely
ref,
department,
job, the person’s name and the person’s image. we also changed the alias from
crewMembers to
crew.
In the same way, we selected the overview text, popularity, movie’s poster url, movie’s release date and title.
Sanity’s GROQ language also allows us to sort a document. To sort an item, we pass order next to a pipe operator.
For instance, if we wish to sort movies by their
releaseDate in ascending order, we could do the following.
const data = await sanityAPI.fetch( `*[_type == 'movie']{ ... } | order(releaseDate, asc)` );
We used this notion in the
sortMoviesBy function to sort either by ascending or descending order.
Let’s take a look at this function below.
const sortMoviesBy = (item, type) => async (dispatch) => { try { dispatch({ type: MOVIES_SORT_REQUEST }); const data = await sanityAPI.fetch( `*[_type == 'movie']{ _id, "poster" : poster.asset->url, title } | order( ${item} ${type})` ); dispatch({ type: MOVIES_SORT_SUCCESS, payload: data }); } catch (error) { dispatch({ type: MOVIES_SORT_FAIL, payload: error.message }); } };
We began by dispatching an action of type
MOVIES_SORT_REQUEST to determine when the request is loading. We then used the
GROQ syntax to sort and fetch data from the
movie collection. The item to sort by is supplied in the variable
item and the mode of sorting (ascending or descending) is supplied in the variable
type. Consequently, we returned the
id, poster url, and title. Once the data is returned, we dispatched an action of type
MOVIES_SORT_SUCCESS and if it fails, we dispatch an action of type
MOVIES_SORT_FAIL.
A similar
GROQ concept applies to the
getMostPopular function. The
GROQ syntax is shown below.
const data = await sanityAPI.fetch( ` *[_type == 'movie']{ _id, "overview": { "text": overview[0].children[0].text }, "poster" : poster.asset->url, title }| order(popularity desc) [0..2]` );
The only difference here is that we sorted the movies by popularity in descending order and then selected only the first three. The items are returned in a zero-based index and so the first three items are items 0, 1 and 2. If we wish to retrieve the first ten items, we could pass
[0..9] to the function.
Here’s the complete code for the movie actions in the
movieActions.js file.
import sanityAPI from "../../sanitySetup"; import { MOVIE_FETCH_FAIL, MOVIE_FETCH_REQUEST, MOVIE_FETCH_SUCCESS, MOVIES_FETCH_FAIL, MOVIES_FETCH_REQUEST, MOVIES_FETCH_SUCCESS, MOVIES_SORT_REQUEST, MOVIES_SORT_SUCCESS, MOVIES_SORT_FAIL, MOVIES_MOST_POPULAR_REQUEST, MOVIES_MOST_POPULAR_SUCCESS, MOVIES_MOST_POPULAR_FAIL, MOVIES_REF_FETCH_SUCCESS, MOVIES_REF_FETCH_FAIL, MOVIES_REF_FETCH_REQUEST } from "../constants/movieConstants"; const fetchAllMovies = () => async (dispatch) => { try { dispatch({ type: MOVIES_FETCH_REQUEST }); const data = await sanityAPI.fetch( `*[_type == 'movie']{ _id, "poster" : poster.asset->url, } ` ); dispatch({ type: MOVIES_FETCH_SUCCESS, payload: data }); } catch (error) { dispatch({ type: MOVIES_FETCH_FAIL, payload: error.message }); } }; }); } }; const fetchMovieById = (id) => async (dispatch) => { try { dispatch({ type: MOVIE_FETCH_REQUEST });]` ); dispatch({ type: MOVIE_FETCH_SUCCESS, payload: data }); } catch (error) { dispatch({ type: MOVIE_FETCH_FAIL, payload: error.message }); } }; const sortMoviesBy = (item, type) => async (dispatch) => { try { dispatch({ type: MOVIES_MOST_POPULAR_REQUEST }); const data = await sanityAPI.fetch( `*[_type == 'movie']{ _id, "poster" : poster.asset->url, title } | order( ${item} ${type})` ); dispatch({ type: MOVIES_SORT_SUCCESS, payload: data }); } catch (error) { dispatch({ type: MOVIES_SORT_FAIL, payload: error.message }); } }; const getMostPopular = () => async (dispatch) => { try { dispatch({ type: MOVIES_SORT_REQUEST }); const data = await sanityAPI.fetch( ` *[_type == 'movie']{ _id, "overview": { "text": overview[0].children[0].text }, "poster" : poster.asset->url, title }| order(popularity desc) [0..2]` ); dispatch({ type: MOVIES_MOST_POPULAR_SUCCESS, payload: data }); } catch (error) { dispatch({ type: MOVIES_MOST_POPULAR_FAIL, payload: error.message }); } }; export { fetchAllMovies, fetchMovieById, sortMoviesBy, getMostPopular, fetchMoviesByRef };
Setting Up The Reducers
Reducers are one of the most important concepts in Redux. They take the previous state and determine the state changes.
Typically, we’ll be using the switch statement to execute a condition for each action type. For instance, we can return
loading when the action type denotes loading, and then the payload when it denotes success or error. It is expected to take in the
initial state and the
action as arguments.
Our
movieReducers.js file contains various reducers to match the actions defined in the
movieActions.js file. However, each of the reducers has a similar syntax and structure. The only differences are the
constants they call and the values they return.
Let’s start by taking a look at the
fetchAllMoviesReducer in the
movieReducers.js file.
import { MOVIES_FETCH_FAIL, MOVIES_FETCH_REQUEST, MOVIES_FETCH; } };
Like all reducers, the
fetchAllMoviesReducer takes the initial state object (
state) and the
action object as arguments. We used the switch statement to check the action types at each point in time. If it corresponds to
MOVIES_FETCH_REQUEST, we return loading as true to enable us to show a loading indicator to the user.
If it corresponds to
MOVIES_FETCH_SUCCESS, we turn off the loading indicator and then return the action payload in a variable
movies. But if it is
MOVIES_FETCH_FAIL, we also turn off the loading and then return the error. We also want the option to reset our movies. This will enable us to clear the states when we need to do so.
We have the same structure for other reducers. The complete
movieReducers.js is shown below.
import { MOVIE_FETCH_FAIL, MOVIE_FETCH_REQUEST, MOVIE_FETCH_SUCCESS, MOVIES_FETCH_FAIL, MOVIES_FETCH_REQUEST, MOVIES_FETCH_SUCCESS, MOVIES_SORT_REQUEST, MOVIES_SORT_SUCCESS, MOVIES_SORT_FAIL, MOVIES_MOST_POPULAR_REQUEST, MOVIES_MOST_POPULAR_SUCCESS, MOVIES_MOST_POPULAR_FAIL, MOVIES_FETCH_RESET, MOVIES_REF_FETCH_REQUEST, MOVIES_REF_FETCH_SUCCESS, MOVIES_REF_FETCH; } }; const fetchMoviesByRefReducer = (state = {}, action) => { switch (action.type) { case MOVIES_REF_FETCH_REQUEST: return { loading: true }; case MOVIES_REF_FETCH_SUCCESS: return { loading: false, movies: action.payload }; case MOVIES_REF_FETCH_FAIL: return { loading: false, error: action.payload }; default: return state; } }; const fetchMovieByIdReducer = (state = {}, action) => { switch (action.type) { case MOVIE_FETCH_REQUEST: return { loading: true }; case MOVIE_FETCH_SUCCESS: return { loading: false, movie: action.payload }; case MOVIE_FETCH_FAIL: return { loading: false, error: action.payload }; default: return state; } }; const sortMoviesByReducer = (state = {}, action) => { switch (action.type) { case MOVIES_SORT_REQUEST: return { loading: true }; case MOVIES_SORT_SUCCESS: return { loading: false, movies: action.payload }; case MOVIES_SORT_FAIL: return { loading: false, error: action.payload }; default: return state; } }; const getMostPopularReducer = (state = {}, action) => { switch (action.type) { case MOVIES_MOST_POPULAR_REQUEST: return { loading: true }; case MOVIES_MOST_POPULAR_SUCCESS: return { loading: false, movies: action.payload }; case MOVIES_MOST_POPULAR_FAIL: return { loading: false, error: action.payload }; default: return state; } }; export { fetchAllMoviesReducer, fetchMovieByIdReducer, sortMoviesByReducer, getMostPopularReducer, fetchMoviesByRefReducer };
We also followed the exact same structure for
personReducers.js. For instance, the
fetchAllPersonsReducer function defines the states for fetching all persons in the database.
This is given in the code below.
import { PERSONS_FETCH_FAIL, PERSONS_FETCH_REQUEST, PERSONS_FETCH_SUCCESS, } from "../constants/personConstants"; const fetchAllPersonsReducer = (state = {}, action) => { switch (action.type) { case PERSONS_FETCH_REQUEST: return { loading: true }; case PERSONS_FETCH_SUCCESS: return { loading: false, persons: action.payload }; case PERSONS_FETCH_FAIL: return { loading: false, error: action.payload }; default: return state; } };
Just like the
fetchAllMoviesReducer, we defined
fetchAllPersonsReducer with
state and
action as arguments. These are standard setup for Redux reducers. We then used the switch statement to check the action types and if it’s of type
PERSONS_FETCH_REQUEST, we return loading as true. If it’s
PERSONS_FETCH_SUCCESS, we switch off loading and return the payload, and if it’s
PERSONS_FETCH_FAIL, we return the error.
Combining Reducers
Redux’s
combineReducers function allows us to combine more than one reducer and pass it to the store. We’ll combine our movies and persons reducers in an
index.js file within the
reducers folder.
Let’s take a look at it.
import { combineReducers } from "redux"; import { fetchAllMoviesReducer, fetchMovieByIdReducer, sortMoviesByReducer, getMostPopularReducer, fetchMoviesByRefReducer } from "./movieReducers"; import { fetchAllPersonsReducer, fetchPersonByIdReducer, countPersonsReducer } from "./personReducers"; import { toggleTheme } from "./globalReducers"; export default combineReducers({ fetchAllMoviesReducer, fetchMovieByIdReducer, fetchAllPersonsReducer, fetchPersonByIdReducer, sortMoviesByReducer, getMostPopularReducer, countPersonsReducer, fetchMoviesByRefReducer, toggleTheme });
Here we imported all the reducers from the movies, persons, and global reducers file and passed them to
combineReducers function. The
combineReducers function takes an object which allows us to pass all our reducers. We can even add an alias to the arguments in the process.
We’ll work on the
globalReducers later.
We can now pass the reducers in the Redux
store.js file. This is shown below.
import { createStore, applyMiddleware } from "redux"; import thunk from "redux-thunk"; import reducers from "./reducers/index"; export default createStore(reducers, initialState, applyMiddleware(thunk));
Having set up our Redux workflow, let’s set up our React application.
Setting Up Our React Application
Our react application will list movies and their corresponding cast and crewmembers. We will be using
react-router-dom for routing and
styled-components for styling the app. We’ll also use Material UI for icons and some UI components.
Enter the following
bash command to install the dependencies.
npm install react-router-dom @material-ui/core @material-ui/icons query-string
Here’s what we’ll be building:
Connecting Redux To Our React App
React-redux ships with a Provider function that allows us to connect our application to the Redux store. To do this, we have to pass an instance of the store to the Provider. We can do this either in our
index.js or
App.js file.
Here’s our index.js file.
import React from "react"; import ReactDOM from "react-dom"; import "./index.css"; import App from "./App"; import { Provider } from "react-redux"; import store from "./redux/store"; ReactDOM.render( <Provider store={store}> <App /> </Provider>, document.getElementById("root") );
Here, we imported
Provider from
react-redux and
store from our Redux store. Then we wrapped our entire components tree with the Provider, passing the store to it.
Next, we need
react-router-dom for routing in our React application.
react-router-dom comes with
BrowserRouter,
Switch and
Route that can be used to define our path and routes.
We do this in our
App.js file. This is shown below.
import React from "react"; import Header from "./components/Header"; import Footer from "./components/Footer"; import { BrowserRouter as Router, Switch, Route } from "react-router-dom"; import MoviesList from "./pages/MoviesListPage"; import PersonsList from "./pages/PersonsListPage"; function App() { return ( <Router> <main className="contentwrap"> <Header /> <Switch> <Route path="/persons/"> <PersonsList /> </Route> <Route path="/" exact> <MoviesList /> </Route> </Switch> </main> <Footer /> </Router> ); } export default App;
This is a standard setup for routing with react-router-dom. You can check it out in their documentation. We imported our components
Header,
Footer,
PersonsList and
MovieList. We then set up the
react-router-dom by wrapping everything in
Router and
Switch.
Since we want our pages to share the same header and footer, we had to pass the
<Header /> and
<Footer /> component before wrapping the structure with
Switch. We also did a similar thing with the
main element since we want it to wrap the entire application.
We passed each component to the route using
Route from
react-router-dom.
Defining Our Pages And Components
Our application is organized in a structured way. Reusable components are stored in the
components folder while Pages are stored in the
pages folder.
Our
pages comprise
movieListPage.js,
moviePage.js,
PersonListPage.js and
PersonPage.js. The
MovieListPage.js lists all the movies in our Sanity.io backend as well as the most popular movies.
To list all the movies, we simply
dispatch the
fetchAllMovies action defined in our
movieAction.js file. Since we need to fetch the list as soon as the page loads, we have to define it in the
useEffect. This is shown below.
import React, { useEffect } from "react"; import { fetchAllMovies } from "../redux/actions/movieActions"; import { useDispatch, useSelector } from "react-redux"; const MoviesListPage = () => { const dispatch = useDispatch(); useEffect(() => { dispatch(fetchAllMovies()); }, [dispatch]); const { loading, error, movies } = useSelector( (state) => state.fetchAllMoviesReducer ); return ( ... ) }; export default MoviesListPage;
Thanks to the
useDispatch and
useSelector Hooks, we can dispatch Redux actions and select the appropriate states from the Redux store. Notice that the states
error and
movies were defined in our Reducer functions and here selected them using the
useSelector Hook from React Redux. These states namely
error and
movies become available immediately we dispatched the
fetchAllMovies() actions.
Once we get the list of movies, we can display it in our application using the
map function or however we wish.
Here is the complete code for the
moviesListPage.js file.
import React, {useState, useEffect} from 'react' import {fetchAllMovies, getMostPopular, sortMoviesBy} from "../redux/actions/movieActions" import {useDispatch, useSelector} from "react-redux" import Loader from "../components/BackdropLoader" import {MovieListContainer} from "../styles/MovieStyles.js" import SortIcon from '@material-ui/icons/Sort'; import SortModal from "../components/Modal" import {useLocation, Link} from "react-router-dom" import queryString from "query-string" import {MOVIES_FETCH_RESET} from "../redux/constants/movieConstants" const MoviesListPage = () => { const location = useLocation() const dispatch = useDispatch() const [openSort, setOpenSort] = useState(false) useEffect(()=>{ dispatch(getMostPopular()) const {order, type} = queryString.parse(location.search) if(order && type){ dispatch({ type: MOVIES_FETCH_RESET }) dispatch(sortMoviesBy(order, type)) }else{ dispatch(fetchAllMovies()) } }, [dispatch, location.search]) const {loading: popularLoading, error: popularError, movies: popularMovies } = useSelector(state => state.getMostPopularReducer) const { loading: moviesLoading, error: moviesError, movies } = useSelector(state => state.fetchAllMoviesReducer) const { loading: sortLoading, error: sortError, movies: sortMovies } = useSelector(state => state.sortMoviesByReducer) return ( <MovieListContainer> <div className="mostpopular"> { popularLoading ? <Loader /> : popularError ? popularError : popularMovies && popularMovies.map(movie => ( <Link to={`/movie?id=${movie._id}`} <h2>{movie.title}</h2> <p>{movie.overview.text.substring(0, 50)}…</p> </div> </Link> )) } </div> <div className="moviespanel"> <div className="top"> <h2>All Movies</h2> <SortIcon onClick={()=> setOpenSort(true)} /> </div> <div className="movieslist"> { moviesLoading ? <Loader /> : moviesError ? moviesError : movies && movies.map(movie =>( <Link to={`/movie?id=${movie._id}`} key={movie._id}> <img className="movie" src={movie.poster} alt={movie.title} /> </Link> )) } { ( sortLoading ? !movies && <Loader /> : sortError ? sortError : sortMovies && sortMovies.map(movie =>( <Link to={`/movie?id=${movie._id}`} key={movie._id}> <img className="movie" src={movie.poster} alt={movie.title} /> </Link> )) ) } </div> </div> <SortModal open={openSort} setOpen={setOpenSort} /> </MovieListContainer> ) } export default MoviesListPage
We started by dispatching the
getMostPopular movies action (this action selects the movies with the highest popularity) in the
useEffect Hook. This allows us to retrieve the most popular movies as soon as the page loads. Additionally, we allowed users to sort movies by their
releaseDate and
popularity. This is handled by the
sortMoviesBy action dispatched in the code above. Furthermore, we dispatched the
fetchAllMovies depending on the query parameters.
Also, we used the
useSelector Hook to select the corresponding reducers for each of these actions. We selected the states for
error and
movies for each of the reducers.
After getting the
movies from the reducers, we can now display them to the user. Here, we have used the ES6
map function to do this. We first displayed a loader whenever each of the movie states is loading and if there’s an error, we display the error message. Finally, if we get a movie, we display the movie image to the user using the
map function. We wrapped the entire component in a
MovieListContainer component.
The
<MovieListContainer> … </MovieListContainer> tag is a
div defined using styled components. We’ll take a brief look at that soon.
Styling Our App With Styled Components
Styled components allow us to style our pages and components on an individual basis. It also offers some interesting features such as
inheritance,
Theming,
passing of props, etc.
Although we always want to style our pages on an individual basis, sometimes global styling may be desirable. Interestingly, styled-components provide a way to do that, thanks to the
createGlobalStyle function.
To use styled-components in our application, we need to install it. Open your terminal in your react project and enter the following
bash command.
npm install styled-components
Having installed styled-components, Let’s get started with our global styles.
Let’s create a separate folder in our
src directory named
styles. This will store all our styles. Let’s also create a
globalStyles.js file within the styles folder. To create global style in styled-components, we need to import
createGlobalStyle.
import { createGlobalStyle } from "styled-components";
We can then define our styles as follows:
export const GlobalStyle = createGlobalStyle` ... `
Styled components make use of the template literal to define props. Within this literal, we can write our traditional
CSS codes.
We also imported
deviceWidth defined in a file named
definition.js. The
deviceWidth holds the definition of breakpoints for setting our media queries.
import { deviceWidth } from "./definition";
We set overflow to hidden to control the flow of our application.
html, body{ overflow-x: hidden; }
We also defined the header style using the
.header style selector.
.header{ z-index: 5; background-color: ${(props)=>props.theme.midDarkBlue}; display:flex; align-items:center; padding: 0 20px; height:50px; justify-content:space-between; position:fixed; top:0; width:100%; @media ${deviceWidth.laptop_lg} { width:97%; } ... }
Here, various styles such as the background color, z-index, padding, and lots of other traditional CSS properties are defined.
We’ve used the styled-components
props to set the background color. This allows us to set dynamic variables that can be passed from our component. Moreover, we also passed the theme’s variable to enable us to make the most of our theme toggling.
Theming is possible here because we have wrapped our entire application with the
ThemeProvider from styled-components. We’ll talk about this in a moment. Furthermore, we used the
CSS flexbox to properly style our header and set the position to
fixed to make sure it remains fixed with respect to the browser. We also defined the breakpoints to make the headers mobile friendly.
Here is the complete code for our
globalStyles.js file.
import { createGlobalStyle } from "styled-components"; import { deviceWidth } from "./definition"; export const GlobalStyle = createGlobalStyle` html{ overflow-x: hidden; } body{ background-color: ${(props) => props.theme.lighter}; overflow-x: hidden; min-height: 100vh; display: grid; grid-template-rows: auto 1fr auto; } #root{ display: grid; flex-direction: column; } h1,h2,h3, label{ font-family: 'Aclonica', sans-serif; } h1, h2, h3, p, span:not(.MuiIconButton-label), div:not(.PrivateRadioButtonIcon-root-8), div:not(.tryingthis){ color: ${(props) => props.theme.bodyText} } p, span, div, input{ font-family: 'Jost', sans-serif; } .paginate button{ color: ${(props) => props.theme.bodyText} } .header{ z-index: 5; background-color: ${(props) => props.theme.midDarkBlue}; display: flex; align-items: center; padding: 0 20px; height: 50px; justify-content: space-between; position: fixed; top: 0; width: 100%; @media ${deviceWidth.laptop_lg}{ width: 97%; } @media ${deviceWidth.tablet}{ width: 100%; justify-content: space-around; } a{ text-decoration: none; } label{ cursor: pointer; color: ${(props) => props.theme.goldish}; font-size: 1.5rem; } .hamburger{ cursor: pointer; color: ${(props) => props.theme.white}; @media ${deviceWidth.desktop}{ display: none; } @media ${deviceWidth.tablet}{ display: block; } } } .mobileHeader{ z-index: 5; background-color: ${(props) => props.theme.darkBlue}; color: ${(props) => props.theme.white}; display: grid; place-items: center; width: 100%; @media ${deviceWidth.tablet}{ width: 100%; } height: calc(100% - 50px); transition: all 0.5s ease-in-out; position: fixed; right: 0; top: 50px; .menuitems{ display: flex; box-shadow: 0 0 5px ${(props) => props.theme.lightshadowtheme}; flex-direction: column; align-items: center; justify-content: space-around; height: 60%; width: 40%; a{ display: flex; flex-direction: column; align-items:center; cursor: pointer; color: ${(props) => props.theme.white}; text-decoration: none; &:hover{ border-bottom: 2px solid ${(props) => props.theme.goldish}; .MuiSvgIcon-root{ color: ${(props) => props.theme.lightred} } } } } } footer{ min-height: 30px; margin-top: auto; display: flex; flex-direction: column; align-items: center; justify-content: center; font-size: 0.875rem; background-color: ${(props) => props.theme.midDarkBlue}; color: ${(props) => props.theme.white}; } `;
Notice that we wrote pure CSS code within the literal but there are a few exceptions. Styled-components allows us to pass props. You can learn more about this in the documentation.
Apart from defining global styles, we can define styles for individual pages.
For instance, here is the style for the
PersonListPage.js defined in
PersonStyle.js in the
styles folder.
import styled from "styled-components"; import { deviceWidth, colors } from "./definition"; export const PersonsListContainer = styled.div` margin: 50px 80px; @media ${deviceWidth.tablet} { margin: 50px 10px; } a { text-decoration: none; } .top { display: flex; justify-content: flex-end; padding: 5px; .MuiSvgIcon-root { cursor: pointer; &:hover { color: ${colors.darkred}; } } } .personslist { margin-top: 20px; display: grid; place-items: center; grid-template-columns: repeat(5, 1fr); @media ${deviceWidth.laptop} { grid-template-columns: repeat(4, 1fr); } @media ${deviceWidth.tablet} { grid-template-columns: repeat(3, 1fr); } @media ${deviceWidth.tablet_md} { grid-template-columns: repeat(2, 1fr); } @media ${deviceWidth.mobile_lg} { grid-template-columns: repeat(1, 1fr); } grid-gap: 30px; .person { width: 200px; position: relative; img { width: 100%; } .content { position: absolute; bottom: 0; left: 8px; border-right: 2px solid ${colors.goldish}; border-left: 2px solid ${colors.goldish}; border-radius: 10px; width: 80%; margin: 20px auto; padding: 8px 10px; background-color: ${colors.transparentWhite}; color: ${colors.darkBlue}; h2 { font-size: 1.2rem; } } } } `;
We first imported
styled from
styled-components and
deviceWidth from the
definition file. We then defined
PersonsListContainer as a
div to hold our styles. Using media queries and the established breakpoints, we made the page mobile-friendly by setting various breakpoints.
Here, we have used only the standard browser breakpoints for small, large and very large screens. We also made the most of the CSS flexbox and grid to properly style and display our content on the page.
To use this style in our
PersonListPage.js file, we simply imported it and added it to our page as follows.
import React from "react"; const PersonsListPage = () => { return ( <PersonsListContainer> ... </PersonsListContainer> ); }; export default PersonsListPage;
The wrapper will output a
div because we defined it as a div in our styles.
Adding Themes And Wrapping It Up
It’s always a cool feature to add themes to our application. For this, we need the following:
- Our custom themes defined in a separate file (in our case
definition.jsfile).
- The logic defined in our Redux actions and reducers.
- Calling our theme in our application and passing it through the component tree.
Let’s check this out.
Here is our
theme object in the
definition.js file.
export const theme = { light: { dark: "#0B0C10", darkBlue: "#253858", midDarkBlue: "#42526e", lightBlue: "#0065ff", normal: "#dcdcdd", lighter: "#F4F5F7", white: "#FFFFFF", darkred: "#E85A4F", lightred: "#E98074", goldish: "#FFC400", bodyText: "#0B0C10", lightshadowtheme: "rgba(0, 0, 0, 0.1)" }, dark: { dark: "white", darkBlue: "#06090F", midDarkBlue: "#161B22", normal: "#dcdcdd", lighter: "#06090F", white: "white", darkred: "#E85A4F", lightred: "#E98074", goldish: "#FFC400", bodyText: "white", lightshadowtheme: "rgba(255, 255, 255, 0.9)" } };
We have added various color properties for the light and dark themes. The colors are carefully chosen to enable visibility both in light and dark mode. You can define your themes as you want. This is not a hard and fast rule.
Next, let’s add the functionality to Redux.
We have created
globalActions.js in our Redux actions folder and added the following codes.
import { SET_DARK_THEME, SET_LIGHT_THEME } from "../constants/globalConstants"; import { theme } from "../../styles/definition"; export const switchToLightTheme = () => (dispatch) => { dispatch({ type: SET_LIGHT_THEME, payload: theme.light }); localStorage.setItem("theme", JSON.stringify(theme.light)); localStorage.setItem("light", JSON.stringify(true)); }; export const switchToDarkTheme = () => (dispatch) => { dispatch({ type: SET_DARK_THEME, payload: theme.dark }); localStorage.setItem("theme", JSON.stringify(theme.dark)); localStorage.setItem("light", JSON.stringify(false)); };
Here, we simply imported our defined themes. Dispatched the corresponding actions, passing the payload of the themes we needed. The payload results are stored in the local storage using the same keys for both light and dark themes. This enables us to persist the states in the browser.
We also need to define our reducer for the themes.
import { SET_DARK_THEME, SET_LIGHT_THEME } from "../constants/globalConstants"; export const toggleTheme = (state = {}, action) => { switch (action.type) { case SET_LIGHT_THEME: return { theme: action.payload, light: true }; case SET_DARK_THEME: return { theme: action.payload, light: false }; default: return state; } };
This is very similar to what we’ve been doing. We used the
switch statement to check the type of action and then returned the appropriate
payload. We also returned a state
light that determines whether light or dark theme is selected by the user. We’ll use this in our components.
We also need to add it to our root reducer and store. Here is the complete code for our
store.js.
import { createStore, applyMiddleware } from "redux"; import thunk from "redux-thunk"; import { theme as initialTheme } from "../styles/definition"; import reducers from "./reducers/index"; const theme = localStorage.getItem("theme") ? JSON.parse(localStorage.getItem("theme")) : initialTheme.light; const light = localStorage.getItem("light") ? JSON.parse(localStorage.getItem("light")) : true; const initialState = { toggleTheme: { light, theme } }; export default createStore(reducers, initialState, applyMiddleware(thunk));
Since we needed to persist the theme when the user refreshes, we had to get it from the local storage using
localStorage.getItem() and pass it to our initial state.
Adding The Functionality To Our React Application
Styled components provide us with
ThemeProvider that allows us to pass themes through our application. We can modify our App.js file to add this functionality.
Let’s take a look at it.
import React from "react"; import { BrowserRouter as Router, Switch, Route } from "react-router-dom"; import { useSelector } from "react-redux"; import { ThemeProvider } from "styled-components"; function App() { const { theme } = useSelector((state) => state.toggleTheme); let Theme = theme ? theme : {}; return ( <ThemeProvider theme={Theme}> <Router> ... </Router> </ThemeProvider> ); } export default App;
By passing themes through the
ThemeProvider, we can easily use the theme props in our styles.
For instance, we can set the color to our
bodyText custom color as follows.
color: ${(props) => props.theme.bodyText};
We can use the custom themes anywhere we need color in our application.
For example, to define
border-bottom, we do the following.
border-bottom: 2px solid ${(props) => props.theme.goldish};
Conclusion
We began by delving into Sanity.io, setting it up and connecting it to our React application. Then we set up Redux and used the GROQ language to query our API. We saw how to connect and use Redux to our React app using
react-redux, use styled-components and theming.
However, we only scratched the surface on what is possible with these technologies. I encourage you to go through the code samples in my GitHub repo and try your hands on a completely different project using these technologies to learn and master them.
Resources
- Sanity Documentation
- How to Build a Blog with Sanity.io by Kapehe
- Redux Documentation
- Styled Components Documentation
- GROQ Cheat Sheet
- Material UI Documentation
- Redux Middleware and SideEffects
- Redux Thunk Documentation
| https://www.smashingmagazine.com/2021/02/web-app-react-redux-sanity-io/ | CC-MAIN-2021-39 | refinedweb | 6,747 | 50.63 |
We have upgraded the community system as part of the upgrade a password reset is required for all users before login in.
I2C Resource Busy
- Taylor Howard last edited by
If I use i2cget to read a byte from a connected I2C device I get:
root@Omega-0849:~# i2cget -y 0 0x10 0x00 Error: Could not set address to 0x10: Resource busy
If I use the force option I can successfully read a byte:
root@Omega-0849:~# i2cget -y -f 0 0x10 0x00 0x54
Also when using the python
onionI2Cmodule I'm receiving an error:
root@Omega-0849:~# python test.py Could not find platform dependent libraries <exec_prefix> Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>] Traceback (most recent call last): File "test.py", line 4, in <module> readVal = i2c.readBytes(0x10, 0x00, 1) IOError: I2C transaction failed.
This is my Python code:
from OmegaExpansion import onionI2C i2c = onionI2C.OnionI2C() readVal = i2c.readBytes(0x10, 0x00, 1) print readVal
Any idea why this would be happening? | https://community.onion.io/topic/3439/i2c-resource-busy/ | CC-MAIN-2020-34 | refinedweb | 163 | 61.16 |
django-webtest
django-webtest is an app for instant integration of Ian Bicking's WebTest () with django's testing framework.
Installation
$ pip install webtest $ pip install django-webtest
Usage
from django_webtest import WebTest class MyTestCase(WebTest): # optional: we want some initial data to be able to login fixtures = ['users', 'blog_posts'] # optional: default extra_environ for this TestCase extra_environ = {'HTTP_ACCEPT_LANGUAGE': 'ru'} def testBlog(self): # pretend to be logged in as user `kmike` and go to the index page index = self.app.get('/',Blog</a> link, check that it # works (result page doesn't raise exceptions and returns 200 http # code) and test if result page have 'My Article' text.templates and response.context goodness that is usually only available if you use django's native test client. These attributes contain a list of templates that were used to render the response and the context used to render these templates. All django's native asserts ( assertFormError, assertTemplateUsed, assertTemplateNotUsed, assertContains, assertNotContains, assertRedirects) are also supported for WebTest responses.
The session dictionary is available via self.app.session, and has the content as django's native test client.
Unlike django's native test client CSRF checks are not suppressed by default so missing CSRF tokens will cause test fails (and that's good).
If forms are submitted via WebTest forms API then all form fields (including CSRF token) are submitted automagically:
class AuthTest(WebTest): fixtures = ['users.json'] def test_login(self) form = self.app.get(reverse('auth_login')).form form['username'] = 'foo' form['password'] = 'bar' response = form.submit().follow() self.assertEqual(response.context['user'].username, 'foo')
However if forms are submitted via raw POST requests using app.post then csrf tokens become hard to construct. CSRF checks can be disabled by setting csrf_checks attribute to False in this case:
class MyTestCase(WebTest): csrf_checks = False def test_post(self) self.app.post('/').
Why?
While django.test.client.Client is fine for it's purposes, it is not well-suited for functional or integration testing. From django's test client docstring:
This is not intended as a replacement for Twill/Selenium or the like - it is here to allow testing against the contexts and templates produced by a view, rather than the HTML rendered to the end-user.
WebTest plays on the same field as twill. WebTest. django-webtest also is able to provide access to the names of rendered templates and template context just like native django TestClient. Twill however understands HTML better and is more mature so consider it if WebTest doesn't fit for some reason.
Contributing
Development happens at github and bitbucket:
The issue tracker is at bitbucket.
Feel free to submit ideas, bugs, pull requests (git or hg) or regular patches. | https://bitbucket.org/kmike/django-webtest/src/cf5ffa3f6e21?at=tip | CC-MAIN-2015-40 | refinedweb | 446 | 54.12 |
The captcha control is a self contained Web control. It is easy to use and secure. It prevents automated registration.
I was given a task to add a captcha validation on a registration page. I found an article in CodeProject that is right on the target. The article solved that task, but I wasn't very happy about having to have an ASPX page reference, a need for a session variable, not easily customizable, distorting the whole image instead of letter. I'm in my vacation and spending a night creating a control to share and for future references.
Add an assembly reference to the control on a page:
<%@ Register
Add the CaptchaControl tag to the page:
CaptchaControl
<cc:CaptchaControl
On postback, call captcha.Validate(userinput) -- where userinput is what user typed.
captcha.Validate(userinput)
userinput
The default values for the control are:
public CaptchaControl()
{
IgnoreCase = true;
CharCount = 3;
this.Width = new Unit(120);
this.Height = new Unit(50);
Password = this.GetType().AssemblyQualifiedName;
CharSet = CharSets.NumberAndLetters;
}
When two users or browser windows of the page are open, after a user (user1) input passes the validation and the other user refreshes the page, if user1 hits the submit button again with the input from the new image, the user will get "validation failed".
user1
user1
This is because the code in the viewstate does not match the code in the new image any more. This problem shouldn't be a problem in the real application because mostly we would want a page redirect to a different page on successful validation. This problem can be resolved if we append the image file name with a guid id, but this approach introduces a new problem that is having a lot of images being created.Each character needs about 40px of space, if you increase the CharCount by 1, you should increase Width by 40 to prevent character image cutoffs.
viewstate
CharCount
1
Width
40
It was fun to add noise and distort the character bitmap image. This is one of the major advantage features over the article I read. Any cool way of distorting the image, please post it here, I'll add it to the project.
This control stores the encrypted code in the viewstate, the password can be set in the HTML markup or programmatically.
viewstate
My next step on this subject is trying to create an HTML helper function for ASP.NET MVC.
This article, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)
public class CaptchaImageHandler : IHttpHandler {
public bool IsReusable {
get { return true; }
}
public void ProcessRequest(HttpContext context) {
context.Response.ContentType = "image/jpeg";
context.Response.Cache.SetCacheability(HttpCacheability.Public);
context.Response.BufferOutput = false;
// Get the string you want to generate
Stream stream = ... // generate the image to the stream here
const int buffersize = 1024 * 16;
byte[] buffer = new byte[buffersize];
int count = stream.Read(buffer, 0, buffersize);
while (count > 0) {
HttpContext.Current.Response.OutputStream.Write(buffer, 0, count);
count = stream.Read(buffer, 0, buffersize);
}
}
}
General News Suggestion Question Bug Answer Joke Praise Rant Admin
Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages. | http://www.codeproject.com/Articles/46581/Free-Captcha-Control?fid=1553805&df=90&mpp=10&sort=Position&spc=None&tid=3299342 | CC-MAIN-2016-36 | refinedweb | 532 | 55.03 |
Unit Testing¶
Unit testing means testing components of a program in isolation from each other to make sure every part works on its own before using it with others. Extensive testing helps avoid new updates causing unexpected side effects as well as alleviates general code rot (a more comprehensive wikipedia article on unit testing can be found here).
A typical unit test sets calls some function or method with a given input, looks at the result and makes sure that this result looks as expected. Rather than having lots of stand-alone test programs, Evennia makes use of a central test runner. This is a program that gathers all available tests all over the Evennia source code (called test suites) and runs them all in one go. Errors and tracebacks are reported.
By default Evennia only tests itself. But you can also add your own tests to your game code and have Evennia run those for you.
Running the Evennia test suite¶
To run the full Evennia test suite, go to your game folder and issue the command
evennia test evennia
This will run all the evennia tests using the default settings. You could also run only a subset of all tests by specifying a subpackage of the library:
evennia test evennia.commands.default
A temporary database will be instantiated to manage the tests. If everything works out you will see how many tests were run and how long it took. If something went wrong you will get error messages. If you contribute to Evennia, this is a useful sanity check to see you haven’t introduced an unexpected bug.
Running tests with custom settings file¶
If you have implemented your own tests for your game (see below) you can run them from your game dir with
evennia test .
The period (
.) means to run all tests found in the current directory
and all subdirectories. You could also specify, say,
typeclasses or
world if you wanted to just run tests in those subdirs.
Those tests will all be run using the default settings. To run the tests
with your own settings file you must use the
--settings option:
evennia --settings settings.py test .
The
--settings option of Evennia takes a file name in the
mygame/server/conf folder. It is normally used to swap settings
files for testing and development. In combination with
test, it
forces Evennia to use this settings file over the default one.
Writing new tests¶
Evennia’s test suite makes use of Django unit test system, which in turn relies on Python’s unittest module.
If you want to help out writing unittests for Evennia, take a look at Evennia’s coveralls.io page. There you see which modules have any form of test coverage and which does not.
To make the test runner find the tests, they must be put in a module
named
test*.py (so
test.py,
tests.py etc). Such a test
module will be found wherever it is in the package. It can be a good
idea to look at some of Evennia’s
tests.py modules to see how they
look.
Inside a testing file, a
unittest.TestCase class is used to test a
single aspect or component in various ways. Each test case contains one
ore more test methods - these define the actual tests to run. You can
name the test methods anything you want as long as the name starts with
“
test_”. Your
TestCase class can also have a method
setUp().
This is run before each test, setting up and storing whatever
preparations the test methods need. Conversely, a
tearDown() method
can optionally do cleanup after each test.
To test the results, you use special methods of the
TestCase class.
Many of those start with “
assert”, such as
assertEqual or
assertTrue.
Example of a
TestCase class:
import unittest # the function we want to test from mypath import myfunc class TestObj(unittest.TestCase): "This tests a function myfunc." def test_return_value(self): "test method. Makes sure return value is as expected." expected_return = "This is me being nice." actual_return = myfunc() # test self.assertEqual(expected_return, actual_return) def test_alternative_call(self): "test method. Calls with a keyword argument." expected_return = "This is me being baaaad." actual_return = myfunc(bad=True) # test self.assertEqual(expected_return, actual_return)
You might also want to read the documentation for the unittest module.
Using the EvenniaTest class¶
Evennia offers a custom TestCase, the
evennia.utils.test_resources.EvenniaTest class. This class initiates
a range of useful properties on themselves for testing Evennia systems.
Examples are
.account and
.session representing a mock connected
Account and its Session and
.char1 and
char2 representing
Characters complete with a location in the test database. These are all
useful when testing Evennia system requiring any of the default Evennia
typeclasses as inputs. See the full definition of the
EvenniaTest
class in evennia/utils/test_resources.py.
# in a test module from evennia.utils.test_resources import EvenniaTest class TestObject(EvenniaTest): def test_object_search(self): # char1 and char2 are both created in room1 self.assertEqual(self.char1.search(self.char2.key), self.char2) self.assertEqual(self.char1.search(self.char1.location.key), char1.location) # ...
Testing in-game Commands¶
In-game Commands are a special case. Tests for the default commands are
put in
evennia/commands/default/tests.py. This uses a custom
CommandTest class that inherits from
evennia.utils.test_resources.EvenniaTest described above.
CommandTest supplies extra convenience functions for executing
commands and check that their return values (calls of
msg() returns
expected values. It uses Characters and Sessions generated on the
EvenniaTest class to call each class).
Each command tested should have its own
TestCase class. Inherit this
class from the
CommandTest class in the same module to get access to
the command-specific utilities mentioned.
from evennia.commands.default.tests import CommandTest from evennia.commands.default import general class TestSet(CommandTest): "tests the look command by simple call" def test_mycmd(self): self.call(general.CmdLook(), "here", "Room\nroom_desc")
Unit testing contribs with custom models¶
A special case is if you were to create a contribution to go to the
evennia/contrib folder that uses its own database models. The
problem with this is that Evennia (and Django) will only recognize
models in
settings.INSTALLED_APPS. If a user wants to use your
contrib, they will be required to add your models to their settings
file. But since contribs are optional you cannot add the model to
Evennia’s central
settings_default.py file - this would always
create your optional models regardless of if the user wants them. But at
the same time a contribution is a part of the Evennia distribution and
its unit tests should be run with all other Evennia tests using
evennia test evennia.
The way to do this is to only temporarily add your models to the
INSTALLED_APPS directory when the test runs. here is an example of
how to do it.
Note that this solution, derived from this stackexchange answer is currently untested! Please report your findings.
# a file contrib/mycontrib/tests.py from django.conf import settings import django from evennia.utils.test_resources import EvenniaTest OLD_DEFAULT_SETTINGS = settings.INSTALLED_APPS DEFAULT_SETTINGS = dict( INSTALLED_APPS=( 'contrib.mycontrib.tests', ), DATABASES={ "default": { "ENGINE": "django.db.backends.sqlite3" } }, SILENCED_SYSTEM_CHECKS=["1_7.W001"], ) class TestMyModel(EvenniaTest): def setUp(self): if not settings.configured: settings.configure(**DEFAULT_SETTINGS) django.setup() from django.core.management import call_command from django.db.models import loading loading.cache.loaded = False call_command('syncdb', verbosity=0) def tearDown(self): settings.configure(**OLD_DEFAULT_SETTINGS) django.setup() from django.core.management import call_command from django.db.models import loading loading.cache.loaded = False call_command('syncdb', verbosity=0) # test cases below ... def test_case(self): # test case here
A note on adding new tests¶
Having an extensive tests suite is very important for avoiding code degradation as Evennia is developed. Only a small fraction of the Evennia codebase is covered by test suites at this point. Writing new tests is not hard, it’s more a matter of finding the time to do so. So adding new tests is really an area where everyone can contribute, also with only limited Python skills.
Testing for Game development (mini-tutorial)¶
Unit testing can be of paramount importance to game developers. When starting with a new game, it is recommended to look into unit testing as soon as possible; an already huge game is much harder to write tests for. The benefits of testing a game aren’t different from the ones regarding library testing. For example it is easy to introduce bugs that affect previously working code. Testing is there to ensure your project behaves the way it should and continue to do so.
If you have never used unit testing (with Python or another language), you might want to check the official Python documentation about unit testing, particularly the first section dedicated to a basic example.
Basic testing using Evennia¶
Evennia’s test runner can be used to launch tests in your game directory (let’s call it ‘mygame’). Evennia’s test runner does a few useful things beyond the normal Python unittest module:
- It creates and sets up an empty database, with some useful objects (accounts, characters and rooms, among others).
- It provides simple ways to test commands, which can be somewhat tricky at times, if not tested properly.
Therefore, you should use the command-line to execute the test runner, while specifying your own game directories (not the one containing evennia). Go to your game directory (referred as ‘mygame’ in this section) and execute the test runner:
evennia --settings settings.py test commands
This command will execute Evennia’s test runner using your own settings
file. It will set up a dummy database of your choice and look into the
‘commands’ package defined in your game directory (
mygame/commands
in this example) to find tests. The test module’s name should begin with
‘test’ and contain one or more
TestCase. A full example can be found
below.
A simple example¶
In your game directory, go to
commands and create a new file
tests.py inside (it could be named anything starting with
test).
We will start by making a test that has nothing to do with Commands,
just to show how unit testing works:
# mygame/commands/tests.py import unittest class TestString(unittest.TestCase): """Unittest for strings (just a basic example).""" def test_upper(self): """Test the upper() str method.""" self.assertEqual('foo'.upper(), 'FOO')
This example, inspired from the Python documentation, is used to test the ‘upper()’ method of the ‘str’ class. Not very useful, but it should give you a basic idea of how tests are used.
Let’s execute that test to see if it works.
> evennia --settings settings.py test commands TESTING: Using specified settings file 'server.conf.settings'. (Obs: Evennia's full test suite may not pass if the settings are very different from the default. Use 'test .' as arguments to run only tests on the game dir.) Creating test database for alias 'default'... . ---------------------------------------------------------------------- Ran 1 test in 0.001s OK Destroying test database for alias 'default'...
We specified the
commands package to the evennia test command since
that’s where we put out test file. In this case we could just as well
just said
. to search all of
mygame for testing files. If we
have a lot of tests it may be useful to test only a single set at a time
though. We get an information text telling us we are using our custom
settings file (instead of Evennia’s default file) and then the test
runs. The test passes! Change the “FOO” string to something else in the
test to see how it looks when it fails.
Testing commands¶
This section will test the proper execution of the ‘abilities’ command, as described in the First Steps Coding page. Follow this tutorial to create the ‘abilities’ command, we will need it to test it.
Testing commands in Evennia is a bit more complex than the simple testing example we have seen. Luckily, Evennia supplies a special test class to do just that … we just need to inherit from it and use it properly. This class is called ‘CommandTest’ and is defined in the ‘evennia.commands.default.tests’ package. To create a test for our ‘abilities’ command, we just need to create a class that inherits from ‘CommandTest’ and add methods.
We could create a new test file for this but for now we just append to
the
tests.py file we already have in
commands from before.
# bottom of mygame/commands/tests.py from evennia.commands.default.tests import CommandTest from commands.command import CmdAbilities from typeclasses.characters import Character class TestAbilities(CommandTest): character_typeclass = Character def test_simple(self): self.call(CmdAbilities(), "", "STR: 5, AGI: 4, MAG: 2")
- Line 1-4: we do some importing. ‘CommandTest’ is going to be our base class for our test, so we need it. We also import our command (‘CmdAbilities’ in this case). Finally we import the ‘Character’ typeclass. We need it, since ‘CommandTest’ doesn’t use ‘Character’, but ‘DefaultCharacter’, which means the character calling the command won’t have the abilities we have written in the ‘Character’ typeclass.
- Line 6-8: that’s the body of our test. Here, a single command is tested in an entire class. Default commands are usually grouped by category in a single class. There is no rule, as long as you know where you put your tests. Note that we set the ‘character_typeclass’ class attribute to Character. As explained above, if you didn’t do that, the system would create a ‘DefaultCharacter’ object, not a ‘Character’. You can try to remove line 4 and 8 to see what happens when running the test.
- Line 10-11: our unique testing method. Note its name: it should begin by ‘test_’. Apart from that, the method is quite simple: it’s an instance method (so it takes the ‘self’ argument) but no other arguments is needed. The line 11 uses the ‘call’ method, which is defined in ‘CommandTest’. It’s a useful method that compares a command against an expected result. It would be like comparing two strings with ‘assertEqual’, but the ‘call’ method does more things, including testing the command in a realistic way (calling its hooks in the right order, so you don’t have to worry about that).
Our line 11 could be understood as: test the ‘abilities’ command (first parameter), with no argument (second parameter), and check that the character using it receives his/her abilities (third parameter).
Let’s run our new test:
> evennia test unit [...] Creating test database for alias 'default'... .. ---------------------------------------------------------------------- Ran 2 tests in 0.156s OK Destroying test database for alias 'default'...
Two tests were executed, since we have kept ‘TestString’ from last time. In case of failure, you will get much more information to help you fix the bug. | http://evennia.readthedocs.io/en/latest/Unit-Testing.html | CC-MAIN-2018-13 | refinedweb | 2,465 | 66.54 |
We are given an array arr[] containing integers only. The goal is to find the number of subsequences of arr[] such that they have maximum number distinct elements. If the array is [ 4,1,2,3,4 ] then two subsequences will be [ 4,1,2,3 ] and [ 1,2,3,4 ].
Let us understand with examples
Input − arr[]= { 1,3,5,4,2,3,1 }
Output − Count of subsequences having maximum distinct elements are − 4
Explanation − The maximum distinct elements are 1,2,3,4 and 5. Count is 5. Subsequences will be −
[ 1,3,5,4,2 ], [ 3,5,4,2,1], [ 5,4,2,3,1 ], [ 1,5,4,2,3 ].
Input − arr[]= { 5,4,2,1,3 }
Output − Count of subsequences having maximum distinct elements are − 1
Explanation − All elements are distinct. Number of subsequences will be 1.
In this approach we will find subsequences based on the fact that if all elements are distinct then the number of subsequences is 1 that is the array itself. In case there are repeated elements then each repeated element will be part of a new subsequence. So we will create a unorderdered_map of frequencies of distinct elements. Then for each frequency multiply that frequency to count. At the end count has a total number of frequencies.
Take an integer array arr[] as input.
Function Max_distinct_subseq(int arr[], int size) takes the array and its size and returns the count of subsequences with maximum distinct elements.
Take the initial count as 1 as if all elements are distinct then the array itself is subsequence with maximum distinct elements.
Create unordered_map<int, int> hash; to store frequencies of all distinct elements.
Traverse the array using for loop and update frequencies for each element arr[i] using hash[arr[i]]]++.
Now traverse the hash using a for loop. For each frequency it->second( it=iterator) multiply to previous count. As x elements of the same time will be part of x different subsequences.
At the end count has a total number of frequencies.
Return count as result.
#include <bits/stdc++.h> using namespace std; int Max_distinct_subseq(int arr[], int size){ int count = 1; unordered_map<int, int> hash; for (int i = 0; i < size; i++){ hash[arr[i]]++; } for (auto it = hash.begin(); it != hash.end(); it++){ count = count * (it->second); } return count; } int main(){ int arr[] = { 3, 7, 3, 3, 1, 5, 6, 9 }; int size = sizeof(arr) / sizeof(arr[0]); cout<<"Count of subsequences having maximum distinct elements are: "<<Max_distinct_subseq(arr, size); return 0; }
If we run the above code it will generate the following output −
Count of subsequences having maximum distinct elements are: 3 | https://www.tutorialspoint.com/count-of-subsequences-having-maximum-distinct-elements-in-cplusplus | CC-MAIN-2022-05 | refinedweb | 444 | 55.84 |
Here's my code for a Hangman game I'm developing:
#include <iostream> using namespace std; const char *const Words[]={"America", "Barrack", "country", "doctors","salamat"}; #define NWORDS (sizeof Words/sizeof (char *)) int main() { const char *toguess=Words[0]; char guess[15] = " "; int numguesses = 1; char ch; string str1; string str2; // string mychars(""); unsigned wordno=1; std::cout << "Welcome to Hangman!\nRaymund is to be executed for a crime he committed.\nTo save him, you need to guess any of the five words provided by the judge.\nYou are given only 15 tries to solve any of the 5 words.\nGoodluck!\n\nNow enter a number to select the word to guess[1.." << NWORDS << "]: "; cin >> wordno; toguess = Words[wordno-1]; bool playing = true; while (playing) { cout << "Enter your guess:" ; cin >> ch; if (0>ch>5) { cout << "Choose only a number from 1 to 5!" << endl; } else (1<=ch=<5) { mychars = mychars + ch; cout << "Your wrong guesses so far: " << mychars << endl; cout << endl; const char *tp ; char *gp; for (tp=toguess, gp=guess;*tp; tp++, gp++) { if(ch == *tp) *gp=ch; } str1 = toguess; str2 = guess; for(gp=guess; *gp; gp++) cout << *gp << " "; cout << endl; for(int i=1; i <= str1.length();i++) cout << "- "; cout << endl; if ( str1==str2) { cout << "Conrats! You got it right!" << endl ; playing= false; } else { numguesses = numguesses + 1; if (numguesses>15) { cout<<endl<<endl <<" +----+ "<<endl <<" | | "<<endl <<" | O "<<endl <<" | /|\\ "<<endl <<" | / \\ "<<endl <<" |You ran out of guesses! "<<endl <<" ============"<<endl<<endl; playing=false; } } // else } // else } // loop cout << " Thanks for playing! " << endl; return 0; }
I want to add something in this part
mychars = mychars + ch; cout << "Your wrong guesses so far: " << mychars << endl; cout << endl;
such that I can count the number of tries left by the player. My plan is to count the letters in mychars then store that number in a variable and subtract the variable from 15. I wonder if that's right? Also, is my code generally OK? Thanks for the help! | https://www.daniweb.com/programming/software-development/threads/312181/help-with-hangman-game-c-code | CC-MAIN-2021-04 | refinedweb | 323 | 78.99 |
08 February 2012 17:27 [Source: ICIS news]
LONDON (ICIS)--?xml:namespace>
However, its shale gas activities would soon benefit from a very substantial increase from its own capital on the zlotych (Zl) 500m ($159.2m, €120.2m) committed to exploring for unconventional gas over the next two years, Orlen added.
Orlen said it was too early to enter into shale gas partnerships with foreign entities due to the unclear potential of its shale gas concessions.
Its comments followed a letter sent to Polish state-controlled firms, including Orlen, by recently-appointed new treasury minister Mikolaj Budzanowski in which he advised that shale gas partnerships should currently be limited to domestic companies.
In a note to investors, Prague-based investment bank Wood & Company cautioned that it believed it would be “hard [for the Polish firms] to explore shale gas deposits without foreign partnerships, as Polish companies do not have the special know-how”.
In late January, Wood & Company expressed scepticism about Budzanowski's call for an acceleration of shale gas exploration efforts and drilling activities in
Orlen and Encana's deal would have involved Encana investing $200m (€150m) in Orlen-led shale gas exploration in eastern
($1 = €0.75)
($1 = Zl3.14)
(€1 = Zl | http://www.icis.com/Articles/2012/02/08/9530642/polands-pkn-orlen-rules-out-shale-gas-venture-with-canadas-encana.html | CC-MAIN-2015-06 | refinedweb | 204 | 50.06 |
17 June 2013 10:34 [Source: ICIS news]
(recasts paragraph 12 for clarity)
By Prema Viswanathan
SINGAPORE (ICIS)--Players in Iran’s petrochemical markets are hopeful that international sanctions on the Middle Eastern country will eventually ease following the election of its former top nuclear negotiator – Hassan Rohani – as president, industry sources said on Monday.
China, India, Turkey and Brazil count Iran as a major supplier of petroleum and petrochemical products.
Rohani won the presidential elections held on 14 June, taking the reins from Mahmoud Ahmadinejad - who has a hardline stance on Iran's nuclear issue and has been ruling the country since 2005.
“We are cautiously optimistic that some change will be seen even in the next two months in terms of an easing of trade flows,” a Tehran-based supplier said.
In recent years, international sanctions in Iran have been tightening over the country’s suspected nuclear weapons programme, posing major deterrents in its trade dealings with the rest of the world.
Any material change to the prevailing trading conditions, however, will take time to emerge, market players said.
“The removal of sanctions will be a more long-drawn process which will take a couple of years,” the Iranian supplier said.
Rohani, who was nicknamed “the diplomat sheikh”, served as secretary of the Supreme National Security Council for 16 years, and was the country’s top nuclear negotiator from 6 October 2003 to 15 August 2005.
“We cannot comment on how the political negotiations will go, but we are hoping that given the new president’s past experience as a negotiator, he can open up dialogue on the nuclear issue and help Iran resume normal trade with the rest of the world,” said a market source.
In May, the international sanctions on Iran were reinforced, directly targeted at trade and investments in its petrochemical industry.
In the fiscal year ending 20 March 2013, Iran exported around 15.5m tonnes of petrochemical products valued at $11.6bn (€8.7bn), out of its total domestic capacity of 60m tonnes/year.
Iranian-origin material accounted for about 27% and 16% of China’s low density polyethylene (PE) and high density PE (HDPE imports, respectively, in 2012.
China also relies on Iran for 40% of its methanol import requirements. India, on the other hand, procures more than 70% of its methanol import from the Middle Eastern country.
Before any major policy change can be enforced in Iran, Rohani will need to convince Iranian religious leaders and the United Nations of his broad plans, taking into account recent developments in the Arab world, a Saudi-based market player said.
“We have to keep in mind that his power is limited as the ultimate lead is still with the religious leaders of Iran,” he said
“[The] ongoing unrest in many countries of the Middle East could [prompt] leaders of Iran [to] start a slow reform process to avoid any civil conflicts like we see in Syria and other places,” the Saudi market player said.
“I believe that we have a [clearer] view if we better understand his role towards Syria and how he will negotiate the nuclear program with the UN in the next weeks and how he can convince the religious leaders to support him in his reform ideas. If he will master those two major challenges I see a strong player coming back into the market,” he added.
In the near term, market conditions are likely to stay roughly the same for Iran, market sources said.
Iran’s low-priced polyethylene (PE) cargoes are expected to continue flowing in huge volumes into neighbouring Pakistan via illegal routes because of porous borders, said a Karachi-based trader. Pakistan is located southeast of Iran.
“Things may change if [Rohani] manages to overturn the [US-led] sanctions. But he will take time to change anything – one or two years even, or more. He may face pressure from the Supreme Council,” the Pakistani trader said.
Electing a “moderate leader is good for Iran and the world”, an India-based petrochemical trader said.
“While I don’t expect any immediate change in the situation, I think western nations will wait and watch before taking any further action to curtail trade with Iran.
“In the long run, it will depend on how the equation between the religious and political leadership within Iran emerges, which is anybody’s guess,” he added.
?xml:namespace>
($1 = €0.75)Additional reporting by Muhamad Fadhil, Heng Hui, Chow Bee Lin | http://www.icis.com/Articles/2013/06/17/9678963/iran-chem-players-see-intl-sanctions-easing-after-rohani-win.html | CC-MAIN-2014-42 | refinedweb | 747 | 56.29 |
I recently took the JavaBronze SE 7/8 exam. I will introduce the problems that I personally worried about while studying.
public class Q1 { public static void main(String[] args) { char chr = 65; int num = 'A'; Q1 q1 = new Q1(); q1.func1(chr, num); } void func1(char chr, int num) { System.out.println("chr:" + chr); System.out.println("num:" + num); } }
The primitive types int and char are compatible and can be implicitly cast. When you assign an int type to a char type variable, the character corresponding to the ASCII code table is assigned. Conversely, if you assign a char type to an int type variable, the decimal number corresponding to that character will be assigned.
public class Q2 { public static void main(String[] args) { int num = 1; System.out.print(++num + num++ + --num + num--); } }
A.4
B.6
C.7
D.8
++ num is calculated as 2, num ++ is calculated as 2, --num is calculated as 2, and num-- is calculated as 2, for a total of 8. It should be noted that num ++ is calculated as 2 and then incremented to 3 Shortly thereafter, it is about to be decremented by --num to 2. If the same variable is on the right side of the postfixed increment and decrement operators, the processing is reflected for that variable.
public class Q3 { public static void main(String[] args) { char[] chr = { 'A', 'B', 'C' }; while (true) for (int i = 0; i <= chr.length; i++) System.out.println(chr[i]); } }
An infinite loop occurs in the while statement, and a runtime error (ArrayIndexOutOfBoundsException) occurs in the for statement. However, when a run-time error occurs in the for statement, even if it is in the middle of the processing being executed The process is stopped halfway and is output as a run-time error. If these occur at the same time, the possibility of them occurring in the following order will be determined.
class Super { static void func(String str) { } } class Sub extends Super { String str; void func(String str) { this.str = str; } void print() { System.out.println(str); } } public class Q4 { public static void main(String[] args) { Sub sub = new Sub(); sub.func("Hello World"); sub.print(); } }
The func method of the Super class is overridden by the func method of the Sub class, but since it is statically qualified, a compile error will occur. It's important to note that the problem is that you don't notice the error when you trace the main method. You need to focus on the methods that are related to superclass and subclass overrides, not the processing of the main method.
class Super { void print() { System.out.println("Hello World"); } } class Sub extends Super { } public class Q5 { public static void main(String[] args) { Sub sub; Super spr; sub = new Sub(); spr = sub; sub = spr; sub.print(); } }
No error occurs in the
spr = sub; line that assigns to the subclass type → superclass type.
A compile error occurs on the line
sub = spr; that assigns to the superclass type → subclass type.
For basic data types, an explicit cast is required when assigning a large variable to a small variable.
Example) double double_01 = 100.0; int int_01 = double_01; //Implicit cast(An error occurs) int int_01 = (int)double_01; //Explicit cast(No error occurs)
Then, in the case of a class type, an error does not occur even if it is assigned to a subclass type that is a large type (super class + difference class) → a super class that is a small type (super class only), and vice versa?
The reason is that the compiler checks the compilation for type compatibility to determine if it is an error. In the case of basic data, compatibility is judged by the size of the type, but in the case of class type, it is judged by looking at the contents of the class.
The subclass says which superclass it inherits from, but the superclass doesn't say which subclass it inherits from.
Therefore, if you assign a superclass to a subclass, you will not get an error because you know the compatibility, but if you assign a subclass to a superclass, you will not know the compatibility, so if you do not explicitly cast it, you will get a compile error. It becomes.
What did you think. This is perfect for exam preparation! !! !! Maybe \ _ (: 3 "∠) \ _
Recommended Posts | https://linuxtut.com/java-bronze-5-problems-to-keep-in-mind-d234c/ | CC-MAIN-2020-50 | refinedweb | 729 | 62.27 |
) Would get interesting - unless you are migrating all users at once you are going to have to allow both Exchange systems to accept mail for this domain. Easiest way of doing this would be to allow Exchange 2007 to be the primary target for that domain, and configure it as Non-Authorative, with a send connector for that domain to the old Exchange system. This would then route to the world.
3) seems fine
4) tbh, unless you want to invest heavily in the Quest tools (or similar), yes it is the best method.
Thanks for the tips about Exchange Primary Target I will have a read up on that tomorrow.
I am now going to rebuild this server as a 2003 R2 DC in the existing domain really as a belt and braces solution to make sure the "Mailbox Move" works correctly for Exchange 2007 across Forests.
This part of the MSDN article makes me think I am doing the right thing
"If you have a forest with a previous version of Exchange that contains only Windows 2000 Server domain controllers (not Windows Server 2003 domain controllers), you cannot use the Move-Mailbox cmdlet to move mailboxes to an Exchange 2007 server in another forest. The Move-Mailbox cmdlet can communicate only with domain controllers running Windows Server 2003 with Service Pack 1 or later. To move mailboxes, you must have at least one domain controller in both the source and the destination forests running Windows Server 2003 with Service Pack 1 or later."
So I will do this first and I have also found an excellent article about Exchange 2007 SMTP Namespace Sharing and different Relay Domain Types which should also help with my 2nd concern from the initial question.
Administration of Active Directory does not have to be hard. Too often what should be a simple task is made more difficult than it needs to be.The solution? Hyena from SystemTools Software. With ease-of-use as well as powerful importing and bulk updating capabilities.
Shared namespace caused some issues as I could feed mail through to Exchange but when sending back you would get undeliverable as the exchange 2000 server held the authoritative settings for the shared domain.
However in the Virtual SMTP server settings there is the option to route all unrelsoved names to another host, so I enter the new Exchange 2007 server in as this smart host and mail flows happily between both servers, and they can both send to the internet via our firewall.
Now to migrate 26 sites and 500 users :-)
Move-Mailbox is also quite a minefield, cross-forest moves always bring the mailbox in as linked rather than user, which could have an NDR issue if replying to old mail, but working on that one at the moment.
Thanks goodness we only do this every 8 years .... LOL
1. Migrate User and Groups with ADMT to preserve SID and access to local profile (saves work at the back end) If you want to preseve local profile you need to migrate user before the computer.
2. Migrate the computer with ADMT. Can be a bit fiddly to get working, there is a hotfix for XP machines so they can deal with RODCs in 2008 and even if you don't have RODCs you have to run the hotfix or else the computer migration continually fails (typical MS shenanigans), also run this as an admin in the source domain.
3. Move-Mailbox cmdlet then migrate the mailbox. I run this simply so it moves the mailbox and deletes the mail attributes from the Source Domain. Doing this also removes the GALsync contact already created in the new domain. All you have to do is create a new GALsync contact to the user in the source domain.
4. User logs onto new domain and all settings, printers, documents etc are preserved, the only thing they have to do is connect to the new Exchange server to get to their e-mail. Although in my case I have also renamed the local printers as local fileserver also upgraded to 2008 so they have to re-add the printers, but they would not have had to do this if I had left the names the same.
So this is a Cross Forest migration from a windows 2000 AD multiple forest with an empty root domain to a new windows 2008 forest with a single domain, moving the mailbox from Exchange 2000 to Exchange 2007 !! It was a slog but got there in the end. | https://www.experts-exchange.com/questions/26481152/Exchange-2000-with-only-2000-DCs-to-new-Forest-with-Ex-2007-Server-2008.html | CC-MAIN-2018-26 | refinedweb | 760 | 62.21 |
Plotting datetime charts
PyGMT accepts a variety of datetime objects to plot data and create charts.
Aside from the built-in Python
datetime object, PyGMT supports input using
ISO formatted strings,
pandas,
xarray, as well as
numpy.
These data types can be used to plot specific points as well as get
passed into the
region parameter to create a range of the data on an axis.
The following examples will demonstrate how to create plots using the different datetime objects.
import datetime import numpy as np import pandas as pd import pygmt import xarray as xr
Using Python’s
datetime
In this example, Python’s built-in
datetime module is used
to create data points stored in list
x. Additionally,
dates are passed into the
region parameter in the format
(x_start, x_end, y_start, y_end),
where the date range is plotted on the x-axis.
An additional notable parameter is
style, where it’s specified
that data points are to be plotted in an X shape with a size
of 0.3 centimeters.
x = [ datetime.date(2010, 6, 1), datetime.date(2011, 6, 1), datetime.date(2012, 6, 1), datetime.date(2013, 6, 1), ] y = [1, 2, 3, 5] fig = pygmt.Figure() fig.plot( projection="X10c/5c", region=[datetime.date(2010, 1, 1), datetime.date(2014, 12, 1), 0, 6], frame=["WSen", "afg"], x=x, y=y, style="x0.3c", pen="1p", ) fig.show()
Out:
<IPython.core.display.Image object>
In addition to specifying the date,
datetime supports the exact time at
which the data points were recorded. Using
datetime.datetime the
region parameter as well as data points can be created with both date and
time information.
Some notable differences to the previous example include
Modifying
frameto only include West (left) and South (bottom) borders, and removing grid lines
Using circles to plot data points defined through
cin
styleparameter
x = [ datetime.datetime(2021, 1, 1, 3, 45, 1), datetime.datetime(2021, 1, 1, 6, 15, 1), datetime.datetime(2021, 1, 1, 13, 30, 1), datetime.datetime(2021, 1, 1, 20, 30, 1), ] y = [5, 3, 1, 2] fig = pygmt.Figure() fig.plot( projection="X10c/5c", region=[ datetime.datetime(2021, 1, 1, 0, 0, 0), datetime.datetime(2021, 1, 2, 0, 0, 0), 0, 6, ], frame=["WS", "af"], x=x, y=y, style="c0.4c", pen="1p", color="blue", ) fig.show()
Out:
<IPython.core.display.Image object>
Using ISO Format
In addition to Python’s
datetime library, PyGMT also supports passing
times in ISO format. Basic ISO strings are formatted as
YYYY-MM-DD with
each
- delineated section marking the four digit year value, two digit
month value, and two digit day value respectively.
When including time of day into ISO strings, the
T character is used, as
can be seen in the following example. This character is immediately followed
by a string formatted as
hh:mm:ss where each
: delineated section
marking the two digit hour value, two digit minute value, and two digit
second value respectively. The figure in the following example is plotted
over a horizontal range of one year from 1/1/2016 to 1/1/2017.
Out:
<IPython.core.display.Image object>
Mixing and matching Python
datetime and ISO dates
The following example provides context on how both
datetime and ISO date
data can be plotted using PyGMT. This can be helpful when dates and times are
coming from different sources, meaning conversions do not need to take place
between ISO and datetime in order to create valid plots.
x = ["2020-02-01", "2020-06-04", "2020-10-04", datetime.datetime(2021, 1, 15)] y = [1.3, 2.2, 4.1, 3] fig = pygmt.Figure() fig.plot( projection="X10c/5c", region=[datetime.datetime(2020, 1, 1), datetime.datetime(2021, 3, 1), 0, 6], frame=["WSen", "afg"], x=x, y=y, style="i0.4c", pen="1p", color="yellow", ) fig.show()
Out:
<IPython.core.display.Image object>
Using
pandas.date_range
In the following example,
pandas.date_range produces a list of
pandas.DatetimeIndex objects, which gets is used to pass date
data to the PyGMT figure.
Specifically
x contains 7 different
pandas.DatetimeIndex
objects, with the number being manipulated by the
periods parameter. Each
period begins at the start of a business quarter as denoted by BQS when
passed to the
periods parameter. The initial date is the first argument
that is passed to
pandas.date_range and it marks the first data point
in the list
x that will be plotted.
x = pd.date_range("2018-03-01", periods=7, freq="BQS") y = [4, 5, 6, 8, 6, 3, 5] fig = pygmt.Figure() fig.plot( projection="X10c/10c", region=[datetime.datetime(2017, 12, 31), datetime.datetime(2019, 12, 31), 0, 10], frame=["WSen", "ag"], x=x, y=y, style="i0.4c", pen="1p", color="purple", ) fig.show()
Out:
<IPython.core.display.Image object>
Using
xarray.DataArray
In this example, instead of using a
pandas.date_range,
x is
initialized as a list of
xarray.DataArray objects. This object
provides a wrapper around regular PyData formats. It also allows the data to
have labeled dimensions while supporting operations that use various pieces
of metadata.The following code uses
pandas.date_range object to fill
the DataArray with data, but this is not essential for the creation of a
valid DataArray.
x = xr.DataArray(data=pd.date_range(start="2020-01-01", periods=4, freq="Q")) y = [4, 7, 5, 6] fig = pygmt.Figure() fig.plot( projection="X10c/10c", region=[datetime.datetime(2020, 1, 1), datetime.datetime(2021, 4, 1), 0, 10], frame=["WSen", "ag"], x=x, y=y, style="n0.4c", pen="1p", color="red", ) fig.show()
Out:
<IPython.core.display.Image object>
Using
numpy.datetime64
In this example, instead of using a
pd.date_range,
x is
initialized as an
np.array object. Similar to
xarray.DataArray
this wraps the dataset before passing it as a parameter. However,
np.array objects use less memory and allow developers to specify
datatypes.
x = np.array(["2010-06-01", "2011-06-01T12", "2012-01-01T12:34:56"], dtype="datetime64") y = [2, 7, 5] fig = pygmt.Figure() fig.plot( projection="X10c/10c", region=[datetime.datetime(2010, 1, 1), datetime.datetime(2012, 6, 1), 0, 10], frame=["WS", "ag"], x=x, y=y, style="s0.5c", pen="1p", color="blue", ) fig.show()
Out:
<IPython.core.display.Image object>
Generating an automatic region
Another way of creating charts involving datetime data can be done
by automatically generating the region of the plot. This can be done
by passing the dataframe to
pygmt.info, which will find
maximum and minimum values for each column and create a list
that could be passed as region. Additionally, the
spacing argument
can be passed to increase the range past the maximum and minimum
data points.
data = [ ["20200712", 1000], ["20200714", 1235], ["20200716", 1336], ["20200719", 1176], ["20200721", 1573], ["20200724", 1893], ["20200729", 1634], ] df = pd.DataFrame(data, columns=["Date", "Score"]) df.Date = pd.to_datetime(df["Date"], format="%Y%m%d") fig = pygmt.Figure() region = pygmt.info( data=df[["Date", "Score"]], per_column=True, spacing=(700, 700), coltypes="T" ) fig.plot( region=region, projection="X15c/10c", frame=["WSen", "afg"], x=df.Date, y=df.Score, style="c0.4c", pen="1p", color="green3", ) fig.show()
Out:
<IPython.core.display.Image object>
Setting Primary and Secondary Time Axes
This example focuses on labeling the axes and setting intervals
at which the labels are expected to appear. All of these modifications
are added to the
frame parameter and each item in that list modifies
a specific section of the plot.
Starting off with
WS, adding this string means that only
Western/Left (W) and Southern/Bottom (S) borders of
the plot will be shown. For more information on this, please
refer to frame instructions.
The other important item in the
frame list is
"sxa1Of1D". This string modifies the secondary
labeling (s) of the x-axis (x). Specifically,
it sets the main annotation and major tick spacing interval
to one month (a1O) (capital letter o, not zero). Additionally,
it sets the minor tick spacing interval to 1 day (f1D).
The labeling of this axis can be modified by setting
FORMAT_DATE_MAP to ‘o’ to use the month’s
name instead of its number. More information about configuring
date formats can be found on the
official GMT documentation page.
x = pd.date_range("2013-05-02", periods=10, freq="2D") y = [4, 5, 6, 8, 9, 5, 8, 9, 4, 2] fig = pygmt.Figure() with pygmt.config(FORMAT_DATE_MAP="o"): fig.plot( projection="X15c/10c", region=[datetime.datetime(2013, 5, 1), datetime.datetime(2013, 5, 25), 0, 10], frame=["WS", "sxa1Of1D", "pxa5d", "sy+lLength", "pya1+ucm"], x=x, y=y, style="c0.4c", pen="1p", color="green3", ) fig.show()
Out:
<IPython.core.display.Image object>
The same concept shown above can be applied to smaller as well as larger intervals. In this example, data is plotted for different times throughout two days. Primary x-axis labels are modified to repeat every 6 hours and secondary x-axis label repeats every day and shows the day of the week.
Another notable mention in this example is setting FORMAT_CLOCK_MAP to “-hhAM” which specifies the format used for time. In this case, leading zeros are removed using (-), and only hours are displayed. Additionally, an AM/PM system is being used instead of a 24-hour system. More information about configuring time formats can be found on the official GMT documentation page.
x = pd.date_range("2021-04-15", periods=8, freq="6H") y = [2, 5, 3, 1, 5, 7, 9, 6] fig = pygmt.Figure() with pygmt.config(FORMAT_CLOCK_MAP="-hhAM"): fig.plot( projection="X15c/10c", region=[ datetime.datetime(2021, 4, 14, 23, 0, 0), datetime.datetime(2021, 4, 17), 0, 10, ], frame=["WS", "sxa1K", "pxa6H", "sy+lSpeed", "pya1+ukm/h"], x=x, y=y, style="n0.4c", pen="1p", color="lightseagreen", ) fig.show()
Out:
<IPython.core.display.Image object>
Total running time of the script: ( 0 minutes 10.075 seconds)
Gallery generated by Sphinx-Gallery | https://www.pygmt.org/latest/tutorials/advanced/date_time_charts.html | CC-MAIN-2022-27 | refinedweb | 1,653 | 60.82 |
offset XRef animation
On 09/04/2013 at 08:55, xxxxxxxx wrote:
Hi all,
I am trying to set the 'offset' parameter of an XRef object, but I can't get it to work. Does anybody know what I am doing wrong?
Here's my code:
op[c4d.ID_CA_XREF_OFFSET] = c4d.BaseTime(10,doc.GetFps())
c4d.EventAdd()
Thanks in advance!
On 09/04/2013 at 09:16, xxxxxxxx wrote:
your code is technically correct, however the definition of your basetime is a bit unusual.
the basetime denominator is not meant to be the documents frames per second. basetime
is just a format to store fractional values like 1/3 without rounding errors. the basetime in
your example would be for 30 fps 10.0/30.0 = 0.3_ sec.
On 09/04/2013 at 10:20, xxxxxxxx wrote:
Hi, and thanks for the reply.
So do you know what to change to get it to work? What is the correct use of c4d.basetime?
Grtz!
On 09/04/2013 at 10:36, xxxxxxxx wrote:
it is difficult to say what is right and what is wrong in your context. when you want to set
the xref time offset to 10 seconds it is simply
op[c4d.ID_CA_XREF_OFFSET] = c4d.BaseTime(10)
if your code still does not work as expected you will have to add a more extensive example.
On 09/04/2013 at 11:31, xxxxxxxx wrote:
Hi again,
If I have an XRef object in the scene, select it and run this code:
import c4d
op[c4d.ID_CA_XREF_OFFSET] = c4d.BaseTime(10)
c4d.EventAdd()
Then nothing happens. No changed offset, no errors in the console. Nothing.
On 09/04/2013 at 12:09, xxxxxxxx wrote:
i do not mean to be rude but your code does not make any sense at all. op is a predefined
variable which is pointing to the hosting object for a script. for an actual script this is None,
for a python tag it is the tag and for a python node the GvNode. so op can never be a xref
in a scripting environment. also your snippet is pretty unusual, as you include the import
statements but not any wrapping method for your code.
this script will set the time offset to 10 sec for the currently selected xref. edit: it does not
actually check if obj is a baselist2d (and a xref) so this is something you might want to change
for an actual script ;)
import c4d def main() : obj = doc.GetActiveObject() obj[c4d.ID_CA_XREF_GENERATOR]]= True obj[c4d.ID_CA_XREF_OFFSET] = c4d.BaseTime(10) c4d.EventAdd() if __name__=='__main__': main()
On 09/04/2013 at 12:27, xxxxxxxx wrote:
And Hi again ;)
Hmmm... either I am not getting it or something weird is going on, because if I run your code while an XRef object is selected, the offset does not change. (By the way, there is one bracket too many in your script that caused it to throw an error initially) Have you tried the script yourself? Does it work for you?
I modified your code a bit to test some things. When i put in some lines to change some other parameters, those lines seem to work fine. See the comments in the code. Strange...
Thanks very much for taking the time to help me though! Cheers!
import c4d
def main() :
obj = doc.GetActiveObject()
#this doesn't seem to do anything:
obj[c4d.ID_CA_XREF_GENERATOR] = True
obj[c4d.ID_CA_XREF_OFFSET] = c4d.BaseTime(10)
#this, however, works fine:
obj[c4d.ID_BASEOBJECT_REL_POSITION,c4d.VECTOR_X] = 10
obj[c4d.ID_CA_XREF_SHOWOBJECTS] = False
c4d.EventAdd()
if __name__=='__main__':
main()
On 09/04/2013 at 12:52, xxxxxxxx wrote:
hi,
no i hadn't tested it. there is some strange behaviour for xref objects going on. but if you set
the generator flag manually (not sure why c4d refuses it to be set programmatically) the offset
value will be written correctly. this time i have actually tested it ;)
On 09/04/2013 at 12:55, xxxxxxxx wrote:
Ah! Finally! Thanks a lot man!
On 09/04/2013 at 14:54, xxxxxxxx wrote:
op is referring to the active object in a script, so you can omitt this line: obj = doc.GetActiveObject()
Best,
-Niklas | https://plugincafe.maxon.net/topic/7096/8045_offset-xref-animation | CC-MAIN-2020-40 | refinedweb | 700 | 74.69 |
A ring network¶
In this example, a small network of cells, arranged in a ring, will be created and the simulation distributed over multiple threads or GPUs if available.
Note
Concepts covered in this example:
Building a basic
arbor.cellwith a synapse site and spike generator.
Building a
arbor.recipewith a network of interconnected cells.
Running the simulation and extract the results.
The cell¶
Step (1) shows a function that creates a simple cell with a dendrite. We construct the following morphology and label the soma and dendrite:
A 4-segment cell with a soma (pink) and a branched dendrite (light blue).¶
def make_cable_cell(gid): # (1) Build a segment tree tree = arbor.segment_tree() # Soma (tag=1) with radius 6 μm, modelled as cylinder of length 2*radius s = tree.append( arbor.mnpos, arbor.mpoint(-12, 0, 0, 6), arbor.mpoint(0, 0, 0, 6), tag=1 ) # (b0) Single dendrite (tag=3) of length 50 μm and radius 2 μm attached to soma. b0 = tree.append(s, arbor.mpoint(0, 0, 0, 2), arbor.mpoint(50, 0, 0, 2), tag=3) # Attach two dendrites (tag=3) of length 50 μm to the end of the first dendrite. # (b1) Radius tapers from 2 to 0.5 μm over the length of the dendrite. tree.append( b0, arbor.mpoint(50, 0, 0, 2), arbor.mpoint(50 + 50 / sqrt(2), 50 / sqrt(2), 0, 0.5), tag=3, ) # (b2) Constant radius of 1 μm over the length of the dendrite. tree.append( b0, arbor.mpoint(50, 0, 0, 1), arbor.mpoint(50 + 50 / sqrt(2), -50 / sqrt(2), 0, 1), tag=3, )
In step (2) we create a label for both the root and the site of the synapse. These locations will form the endpoints of the connections between the cells.
We’ll create labels for the root (red) and a synapse_site (black).¶
# Associate labels to tags labels = arbor.label_dict( { "soma": "(tag 1)", "dend": "(tag 3)", # (2) Mark location for synapse at the midpoint of branch 1 (the first dendrite). "synapse_site": "(location 1 0.5)", # Mark the root of the tree. "root": "(root)", } )
After we’ve created a basic
arbor.decor, step (3) places a synapse with an exponential decay (
'expsyn') on the
'synapse_site'.
The synapse is given the label
'syn', which is later used to form
arbor.connection objects terminating at the cell.
Note
Mechanisms can be initialized with their name;
'expsyn' is short for
arbor.mechanism('expsyn').
Mechanisms typically have some parameters, which can be queried (see
arbor.mechanism_info) and set
(see
arbor.mechanism). In particular, the
e parameter of
expsyn defaults to
0, which makes it,
given the typical resting potential of cell membranes of
-70 mV, an excitatory synapse.
Step (4) places a threshold detector at the
'root'. The detector is given the label
'detector', which is later used to form
arbor.connection objects originating from the cell.
Note
The number of synapses placed on the cell in this case is 1, because the
'synapse_sites' locset is an explicit location.
Had the chosen locset contained multiple locations, an equal number of synapses would have been placed, all given the same label
'syn'.
The same explanation applies to the number of detectors on this cell.
# (3) Create a decor and a cable_cell decor = ( arbor.decor() # Put hh dynamics on soma, and passive properties on the dendrites. .paint('"soma"', arbor.density("hh")).paint('"dend"', arbor.density("pas")) # (4) Attach a single synapse. .place('"synapse_site"', arbor.synapse("expsyn"), "syn") # Attach a detector with threshold of -10 mV. .place('"root"', arbor.threshold_detector(-10), "detector") )
The recipe¶
To create a model with multiple connected cells, we need to use a
recipe.
The recipe is where the different cells and the connections between them are defined.
Step (5) shows a class definition for a recipe with multiple cells. Instantiating the class requires the desired
number of cells as input. Compared to the simple cell recipe, the main differences
are connecting the cells (8), returning a configurable number of cells (6) and returning a new cell per
gid (7).
Step (8) creates an
arbor.connection between consecutive cells. If a cell has gid
gid, the
previous cell has a gid
(gid-1)%self.ncells. The connection has a weight of 0.01 (inducing a conductance of 0.01 μS
in the target mechanism
expsyn) and a delay of 5 ms.
The first two arguments to
arbor.connection are the source and target of the connection.
The source is a
arbor.cell_global_label object containing a cell index
gid, the source label
corresponding to a valid detector label on the cell and an optional selection policy (for choosing a single detector
out of potentially many detectors grouped under the same label - remember, in this case the number of detectors labeled
‘detector’ is 1).
The
arbor.cell_global_label can be initialized with a
(gid, label) tuple, in which case the selection
policy is the default
arbor.selection_policy.univalent; or a
(gid, (label, policy)) tuple.
The target is a
arbor.cell_local_label object containing a cell index
gid, the target label
corresponding to a valid synapse label on the cell and an optional selection policy (for choosing a single synapse
out of potentially many synapses grouped under the same label - remember, in this case the number of synapses labeled
‘syn’ is 1).
The
arbor.cell_local_label can be initialized with a
label string, in which case the selection
policy is the default
arbor.selection_policy.univalent; or a
(label, policy) tuple. The
gid
of the target cell doesn’t need to be explicitly added to the connection, it is the argument to the
arbor.recipe.connections_on() method.
Step (9) attaches an
arbor.event_generator on the 0th target (synapse) on the 0th cell; this means it
is connected to the
"synapse_site" on cell 0. This initiates the signal cascade through the network. The
arbor.explicit_schedule in instantiated with a list of times in milliseconds, so here a single event at the 1
ms mark is emitted. Note that this synapse is connected twice, once to the event generator, and once to another cell.
Step (10) places a probe at the
"root" of each cell.
Step (11) instantiates the recipe with 4 cells.
# (5) Create a recipe that generates a network of connected cells. class ring_recipe(arbor.recipe): def __init__(self, ncells): # The base C++ class constructor must be called first, to ensure that # all memory in the C++ class is initialized correctly. arbor.recipe.__init__(self) self.ncells = ncells self.props = arbor.neuron_cable_properties() # (6) The num_cells method that returns the total number of cells in the model # must be implemented. def num_cells(self): return self.ncells # (7) The cell_description method returns a cell def cell_description(self, gid): return make_cable_cell(gid) # The kind method returns the type of cell with gid. # Note: this must agree with the type returned by cell_description. def cell_kind(self, gid): return arbor.cell_kind.cable # (8) Make a ring network. For each gid, provide a list of incoming connections. def connections_on(self, gid): src = (gid - 1) % self.ncells w = 0.01 # 0.01 μS on expsyn d = 5 # ms delay return [arbor.connection((src, "detector"), "syn", w, d)] # (9) Attach a generator to the first cell in the ring. def event_generators(self, gid): if gid == 0: sched = arbor.explicit_schedule([1]) # one event at 1 ms weight = 0.1 # 0.1 μS on expsyn return [arbor.event_generator("syn", weight, sched)] return [] # (10) Place a probe at the root of each cell. def probes(self, gid): return [arbor.cable_probe_membrane_voltage('"root"')] def global_properties(self, kind): return self.props # (11) Instantiate recipe ncells = 4 recipe = ring_recipe(ncells)
The execution¶
To create a simulation, we need at minimum to supply the recipe, and in addition can supply a
arbor.context
and
arbor.domain_decomposition. The first lets Arbor know what hardware it should use, the second how to
destribute the work over that hardware. By default, contexts are configured to use 1 thread and domain decompositons to
divide work equally over all threads.
Step (12) initalizes the
threads parameter of
arbor.context with the
avail_threads flag. By supplying
this flag, a context is constructed that will use all locally available threads. On your local machine this will match the
number of logical cores in your system. Especially with large numbers
of cells you will notice the speed-up. (You could instantiate the recipe with 5000 cells and observe the difference. Don’t
forget to turn of plotting if you do; it will take more time to generate the image then to run the actual simulation!)
If no domain decomposition is specified, a default one distributing work equally is created. This is sufficient for now.
You can print the objects to see what defaults they produce on your system.
Step (13) sets all spike generators to record using the
arbor.spike_recording.all policy.
This means the timestamps of the generated events will be kept in memory. Be default, these are discarded.
In addition to having the timestamps of spikes, we want to extract the voltage as a function of time.
Step (14) sets the probes (step 10) to measure at a certain schedule. This is sometimes described as
attaching a sampler to a probe.
arbor.simulation.sample() expects a probeset id and the
desired schedule (here: a recording frequency of 10 kHz, or a
dt of 0.1 ms). Note that the probeset id is a separate index from those of
connection endpoints; probeset ids correspond to the index of the list produced by
arbor.recipe.probes() on cell
gid.
arbor.simulation.sample() returns a handle to the samples that will be recorded. We store
these handles for later use.
Step (15) executes the simulation for a duration of 100 ms.
# (12) Create an execution context using all locally available threads and simulation ctx = arbor.context("avail_threads") sim = arbor.simulation(recipe, ctx) # (13) Set spike generators to record sim.record(arbor.spike_recording.all) # (14) Attach a sampler to the voltage probe on cell 0. Sample rate of 10 sample every ms. handles = [sim.sample((gid, 0), arbor.regular_schedule(0.1)) for gid in range(ncells)] # (15) Run simulation for 100 ms sim.run(100) print("Simulation finished")
The results¶
Step (16) prints the timestamps of the spikes:
# (16) Print spike times print("spikes:") for sp in sim.spikes(): print(" ", sp)
Step (17) generates a plot of the sampling data.
arbor.simulation, which in this case is 1, so we can take the first element.
(Recall that in step (10) we attached a probe to the
"root", which describes one location.
It could have described a locset.)
# (17) Plot the recorded voltages over time. print("Plotting results ...") df_list = [] for gid in range(ncells): samples, meta = sim.samples(handles[gid])[0] df_list.append( pandas.DataFrame( {"t/ms": samples[:, 0], "U/mV": samples[:, 1], "Cell": f"cell {gid}"} ) ) df = pandas.concat(df_list, ignore_index=True) seaborn.relplot(data=df, kind="line", x="t/ms", y="U/mV", hue="Cell", ci=None).savefig( "network_ring_result.svg" )
Since we have created
ncells cells, we have
ncells traces. We should be seeing phase shifted traces, as the action potential propagated through the network.
We plot the results using pandas and seaborn:
The full code¶
You can find the full code of the example at
python/example/network_ring.py. | https://docs.arbor-sim.org/en/latest/tutorial/network_ring.html | CC-MAIN-2022-40 | refinedweb | 1,870 | 59.4 |
Since this is my first blog entry for Inside Aperture, I thought I’d use a couple of paragraphs to introduce myself to readers of this site, and to also share a little about how I’m using Aperture’s Smart Album feature.
I was elated when Derrick Story invited me to blog for Inside Apertue–after he read and accepted an Aperture product review I wrote. Though I’ve been blogging and writing reviews for MyMac.com for the last couple of years, I especially look forward to writing for Inside Aperture mainly because of what I’m learning as a digital wedding photographer who is developing his workflow using Aperture.
I just started my wedding and event photography business last Summer. Getting into this business at this time seems just about right simply because of the ways digital photography has enhanced and seems to be reshaping the entire direction of the professional photography industry. I’m no where near where I would like to be as a photographer, but I’m so inspired and jazzed about all that I’m learning in this era of digital photography.
Like some of you, I’ve been working with both Adobe’s Lightroom and Apple’s Aperture, but choosing to make Aperture my primary digital photo management and processing work horse was easy. For one, I’ve been a die hard Mac user for over 20 years; and two, I just find Aperture’s features and interface so much better for wedding photography.
Using Smart Albums in Wedding Photography
Being an avid user of Apple’s iPhoto since it was introduced (I still have my dog eared copy of the first edition of Derrick Stories Missing Manual book for iPhoto which introduced me to digital photo management) is where I first started using the Smart Albums feature. I now make the same Smart Album feature the center of my workflow in Aperture. Lightroom lacks this feature, and to me it’s one of several reasons I mostly use Aperture.
I set up Smart Albums primarily based on the keywords I use for my wedding projects. Apple had wedding photographers in mind when it built the program because a full list of keywords for wedding projects comes installed with the software. So it was simple for me to add and delete what I needed to the list and get rolling.
Duplicating Smart Albums
Using keywords to set up Smart Albums can take a little time, but the trick to creating several of them is to control click on the first Smart Album you make and set up for a project and select Duplicate Smart Album. From there, you just change the parameters for each duplicated Smart Album and give each of them a title.
It would be useful to me if the next version of Aperture included a way to configure a set of prefigured Smart Albums for each project so they don’t have to be newly created each time.
But nevertheless, Smart Albums are big time savers. In my workflow, I can change the keywords and ratings for various digital files, and the Smart Albums automatically update based on those changes. So my 5 star Smart Album for each project constantly gets updated without me having to drag new files into it.
With Smart Albums, I can make sure I don’t have a bunch of duplicate files being exported out the program. Additionally, I can use the Smart Albums to study the types of photos I’m taking. If want to see, for example, what kinds of photos I’m getting with my favorite Canon 85mm lens, I just check the Smart Album I created to reflect photos taken with that lens. I can also study how my photos are looking based on say that ISO I was using.
Tucking away a project collection of Smart Albums in a folder keeps my Project Panel less crowded and easier to access.
In future entries I will share more from what I’m learning, but I’m also looking forward to exchanges with other photographers and hearing about your workflow and challenges with Aperture.
Bakari
I worked with Aperture (my business partner is a true Mac fanboy) until LR beta 4, but since then it's LR all the way. The key reason is speed of processing multiple big jobs because of background processing. We do 4/5 800 frame jobs a week in lots of artificial lighting (opera/classical work) plus portraits and weekend weddings. Once a job's on its way to web / export / print, we can't afford the downtime of waiting for it to finish. Even without that lag, LR does these output jobs in 2/3 of the time that it takes Aperture.
Just to explain my comment about XMP, Aperture won't read it in sidecar files or if it's embedded. It only writes it if you export or duplicate the masters. I find it incredible that a program designed after XMP should fail to recognize it, especially when the product manager's background was with Extensis Portfolio which has excellent (though patchy) XMP support. For instance, we use it to add job, client, and billing numbers to finished images - using Adobe File Info panels and our own namespace and flow this into our job accounting. For our arts archive, LR reads and preserves all this information - Aperture hoses it.
I'm very critical of the lack of smart albums in LR. As I said, it makes LR feel really dumbed down (we loved them in Portfolio 8 before we switched over to iView) and frankly not professional. In Aperture, we used them you described but something I don't think you said was using them for versions - my partner's wedding shoots have colour, b&w, and often sickly-sweet sepia versions of every shot. So we used smart albums extensively. These tools are databases - how hard is it to let users write queries?
You know, I'm not sure about wanting Aperture to have more speed. That has been getting better, and machines are getting faster. It's the background processing that's the big delay and which needs fixing. Second, the lift and stamp is light years slower than LR's auto sync mode - in Develop you Cmd-click the Sync button. But IMO what Apple really should be doing is bringing out a PC version - after all, Photoshop sales are 60:40 PC:Mac, which is shifting as new serious/amateur photographers are overwhelmingly PC, Apple need economies of scale (eg to support more camera models), and Aperture would hook into a far bigger developer environment for plug ins, web templates, books etc (the plist architecture is cross platform enough).
Gio
galactusofmyth, yeah, it appears that the specific request you have is not there in Aperture, or at least I can't address it based on my use of the program. You are limited to max. and min. focal lengths.
However, one work around may be, I think, is to view photos in the browser using the list view. Do a Command-J and in the List View option change one of the sets to Photo Info- EXIF. With that view you will get the various focal lengths of each shot. From there, you can view particular focal lengths in groups.
Let me know if that helps.
Gio, I think it's important to remember that programs like Aperture have to be flexiable enough to address the needs of many types of users. It's not that I'm a diehard Apple fan, but I just know what works for me at a particular time. I'm not afraid to explore other options and learn what's out there. But for me right now, Aperture is doing the job I need. I like working in a non-linear fashion with my photos. And I like automation features when the work well and get the job done. That doesn't say I won't change my mind later and try out something else.
But in the early stages of beta Lightroom, I suggested and questioned why there was no Smart Album features. When I posted this question on the Adobe discussion board, one of the moderators didn't seem to even know what I was talking about. I felt and still feel that Smart Folders are essential to the photo management and processing workflow. However, I do know that photographers have different needs and purposes, so a more dedicated folder structures may be necessary, which is an option in Aperture.
What Apple really needs to focus on, IMHO, is getting some better speed in Aperture and not at the expense of having a high priced, high performing computer like the MacPro. Aperture should not be geared just to professionals, but also serious amatures who do and will spend their money on a program like this.
Aperture needs to remain a dedicated image processing and managent program for Apple users and it needs to have continuous development so that works with smoothly with other applications.
As for your comment about Aperture's XMP metadata featrues, I need to look into that more. Again, my needs have been very basic so far, but I realize that there are many other photographers out there who require something more and different.
So let me ask you, which program, Aperture or LR, are you primarily using, and why?
Thanks for the response. I am still not seeing it though. Maybe you could walk us through an example using a 70-200 f2.8 as an example? I can see how to set it up for a prime lens (searching by, say, 85mm focal length), but even that would give back "errors" if I used a zoom at a 85mm focal length. In the Aperture forums, one of the big requests is for something akin to what is in Lightroom, as outlined in the Inside Lightroom blog (look at the diagram):
That gets you EXIF data for a specific lens, such as a 70-200, or a 85mm 1.8. I only see Lens minimum (mm) and Lens maximum (mm) as search queries for Aperture. Am I missing something still?
Linear's great when you're trying to get a job done and not jumping round and tweaking whatever catches your eye. After all, you wouldn't expect the guy at McDonalds to swap role whenever he feels like a change, would you? One guy prepares, another flips the burgers, a third takes the cash. Combine linear with background processing and you've got a much faster workflow and throughput.
As for the "archaic" (see * below) folder/album structure, exposing the folders is Adobe listening to its customers' demands during the beta period. This new paradigm stuff, which Adobe also spouted, fools those who've only come from folder-oriented raw processing programs and haven't had any DAM experience. If you have that discipline, then you always plan for things possibly going belly up, and for the inevitable day when you switch your DAM program - solid folder structure and filenaming conventions are going to get you out of trouble. It is inexcusable that LR has no smart albums (its Collections are flexible enough to function as Aperture's projects, unsmart albums or unsmart web albums), and some weak-minded LR users will go on using folders when they would be better off using metadata properly, but Adobe's exposure of folders lets people work the way they want. I never let Apple believers get away with jam-tomorrow arguments, but Adobe will add smart albums - it's a no brainer. And don't defend Apple too much either. Just like 1.5 abandoned only-managed import, Apple will backflip and expose folders. Import folders as projects is a clue. Just watch.
Gio
* When you consider how Aperture handles XMP metadata, don't be too liberal spraying around the word "archaic"
Hey Gio, what can I say. I'm pretty simple guy when it comes to computers and software. So many of Apertures settings and features of course work like Apple's other programs that I just feel right at home with the program.
I'm not yet printing my own photos from Aperture, and I'm still developing my understanding of the whole printing process, but other than that I just find that the more non-linear approach of Aperture is better than what you find in Lightroom.
I really like the design and preset features of LR, but my workflow slows down in that program because I feel like I have to work in a linear fashion--from Browser to Developer, back to Browswer. And again, the folder/album structure in LR is simply archaic. Photoshop CS3 seems to rock, and I even Adobe Camera Raw is well improved, but for just churning out a thousand+ wedding photos, Aperture gets the job done.
galactusofmyth, pull up the settings for a Smart Album and click on the + button. In the drop down menu, select EXIF. You will get another drop down button that includes settings for lens choices.
Welcome to Inside Aperture. You have some great shots on your web site...
Quick question - how do you set up a smart album for a lens? I thought Aperture didn't collect that EXIF data yet.
"For one, I've been a die hard Mac user for over 20 years;"
Hm, that's one of the weakest reasons for choosing the program. But hey, I won't touch anything written by people from Liverpool....
"and two, I just find Aperture's features and interface so much better for wedding photography."
I think you overstate the case of Aperture being better for wedding work. For instance, the lack of background processing makes it slower to send hundreds of images to print/export and then carry on with the next task, eg preparing b&w versions of the whole shoot.
However, smart albums are a very valuable feature for smart users. Their absence dumbs down Lightroom. | http://www.oreillynet.com/digitalmedia/blog/2007/05/using_smart_albums_for_wedding.html | crawl-002 | refinedweb | 2,356 | 68.1 |
Simon Guest
Microsoft Corporation
January 2004
Applies to:
Microsoft® .NET Framework 1.1
Microsoft Visual Studio® .NET 2003
Intrinsyc Ja.NET 1.5 (build 1.5.1287 or higher)
Summary: Learn and helping them to rapidly deploy .NET applications without having to rip out existing infrastructure or wait until new replacement applications are written.(11 printed pages)
Download the SwingInterop.msi code sample
Introduction
Required Software and Skills to Run the Code Samples
Running the Sample Code
How the Sample Works
Where We Are
Conclusion
When I work with organizations that have existing applications in Java, it is often the case that this investment is not only bound to the server side, but can also include the client. These client-side applications are typically developed in either AWT (Abstract Window Toolkit) or Java SWING.
For desktops that are running Microsoft® Windows®, many of these organizations are looking at how Microsoft .NET technology can be used on the client to either replace or extend the existing Java technology. The client-side equivalent of AWT and SWING in Microsoft .NET is referred to as Windows Forms.
This article will show a sample that uses a messaging layer between an existing Java SWING™ application and Windows Forms. Using a third-party toolkit called Ja.NET from Intrinsyc, you'll discover—helping them to rapidly deploy .NET applications without having to rip out existing infrastructure or wait until new replacement applications are written.
Microsoft Visual Studio® .NET 2003 is recommended but not required for building the solution file included in the sample.
To show how this approach works, you will first install, configure and look at the sample code. This demonstrates a sample Java SWING application invoking and displaying two forms, both based on Windows Forms in .NET.
After you've looked at the sample, we'll discuss how the sample works and can be extended.
The sample code is installed and configured in stages. These are outlined as follows:
First, download and run the SwingInterop.MSI to extract and install the sample code. By default, this will be installed to a C:\SwingInterop directory—but you are free to choose your own location. If you choose to change the location, be sure to replace all occurrences of C:\SwingInterop in this article with your new location.
After downloading and installing Ja.NET 1.5 from Intrinsyc (see pre-requisites), you need to configure the Ja.NET libraries so that they are available to the sample code. To do this, copy the following files from C:\Program Files\Intrinsyc Ja.NET\lib (the default library directory for Ja.NET) and place them in the C:\SwingInterop\Java\lib directory:
Janet.jar
Janetor.jar
Also, place the Ja.NET license file (janet_license.xml) you have received via Email in the C:\SwingInterop\Java\lib directory. This is important to make sure that the application has access to the correct license when running.
With the files copied, it is now time to configure the Ja.NET proxy. Open a command prompt, and navigate to the C:\SWINGInterop\Java\SampleClient directory. From here, run the Janetor.bat batch file.
If this does not run, check to make sure that the Java.exe executable is available on your System Path.
Figure 1. Janetor, the Ja.NET Configuration Tool
Within Janetor (the Ja.NET Configuration Tool), ensure that the default remote object is configured for tcp://localhost:5656 (you can change this once the sample is working). It is also advisable to check that the licensing entry matches your license.
Close Janetor and save the settings when prompted. The Ja.NET setup and configuration is now complete.
To build the Java SWING application, navigate to the C:\SWINGInterop\Java\SampleClient directory and run the build.bat file. This will make the required calls to the J2SE Java compiler to build the application.
Again, if for some reason this fails, check your System Path and CLASSPATH to ensure settings include the required Java executables.
If using Visual Studio .NET 2003, you can open and compile the SwingInterop.sln file, found in C:\SWINGInterop\dotNET. Within the IDE you will notice that the solution contains sub-projects:
SampleForms—a set of two sample Windows Forms
SwingInteropAdmin—the administration utility used to stop the engine and to monitor events
SwingInteropEngineLibrary—contains the components that are exposed to the Java SWING application
SwingInteropInvoker—the project responsible for starting the engine
Part of the compile process will also install the assembly containing the two sample forms into the Global Assembly Cache (GAC).
If you wish to instead compile from the command prompt, run the build.bat file from the C:\SWINGInterop\dotNET directory.
If for some reason either build procedure failures, check your System Path—the C# compiler (CSC.EXE) and GAC Registration Tool (GACUTIL.EXE) must be present in your path for this to work. By default, these can be found in the .NET Framework directory.
The engine (this is one of the projects that was just built) is responsible for listening for incoming messages from the Java SWING application and firing up Windows Forms on request.
To start this engine listening for messages, run StartEngine.bat from the C:\SWINGInterop\dotNET directory.
This executable runs as a background process (such that it could be run on startup in a production environment without displaying anything to the user). You can use the Task Manager in Windows Explorer to look at the processes and ensure that the SwingInteropEngine.exe process is running.
Figure 2. Windows Task Manager
After the SWINGInteropEngine process is running, navigate to the C:\SWINGInterop\Java\SampleClient directory and run StartClient.bat to run the test SWING client.
Upon successfully loading the Java SWING client, you will see the main form:
Figure 3. A test message displayed in the main form
This is a form written using Java SWING. To demonstrate the purpose of this sample— how this form can invoke a Windows Form written in .NET—enter a message into the text field and click on the Open Form1 button.
When you click on the button, the Windows Form will be displayed:
Figure 4. After clicking on the button, the Windows Form is displayed.
You can see that the text message has been passed to the Windows Forms as a parameter from the calling client.
To test the reverse—returning a field from the Windows Form back to the Java SWING client—edit the text in this new Windows Form (maybe append some more text in the text field) and click on the Close button.
After the form has closed, the Java SWING application reports the modified text string:
Figure 5. Reporting the modified text string
What we have seen so far is how the SWING client has the ability to invoke a Windows Form via the running SWINGInteropEngine process, pass parameters to the form and obtain a return result from the form.
You may have noticed however that when the Windows Form was displayed, it was running in a modal (or dialog) style. This means that the underlying Java SWING client application was 'frozen' until the new Windows Form had been closed. This is due to the Java client waiting for the return result from the Windows Form.
For the purposes of this sample, if the Java SWING client is not interested in any return result from the Windows Form, it can open the Windows Forms in a non-modal view. The second button on the in the sample application demonstrates this.
Return back to the Java SWING application, and click on the Open Form2 button. The following Windows Form will open:
Figure 6. Another Sample Form window
This is another example of a Windows Form that can be launched from an existing Java SWING application, however as we mentioned, the difference here is that this form is non-modal. The Java SWING application is not waiting for a response. It is therefore possible to return to the calling application and re-open multiple instances of this form.
The sample code starts a new thread in the SWING Interop Engine process for each new instance of the form that is opened.
To stop the Engine cleanly, navigate to the C:\SWINGInterop\dotNET and run the following command:
admin /kill
This sends a terminate signal to the forms engine that is running in the background.
This administration utility (admin.exe) can also be used to trace output from the engine with the following switch:
admin /trace
This will display each event that is generated by the SWINGInteropEngine process to the console. In addition, all events are logged in the Windows Event Log.
To change the TCP port that is used for communication between the Java SWING application and Windows Forms, ensure that the engine is stopped. Restart the engine with a /p parameter indicating the port to use instead of the default (5656):
StartEngine /p:8888
Navigate to the C:\SWINGInterop\Java\SampleClient directory and re-run Janetor.bat. Adjust the remote object's URL to reflect the new port. For example, tcp://localhost:8888
To run the admin utility to either kill or trace the engine, again use the /p parameter.
admin /p:8888 /kill
To now, we have seen how the sample demonstrates how an existing Java SWING application can invoke and display multiple Windows Forms, pass parameters, expect return types and specify modality. How does this work under the covers?
Figure 7. Overview of how the sample demonstrates how an existing Java SWING application can invoke and display multiple Windows Forms and so on
The Java SWING application uses a proxy class (which contains an API and the necessary Ja.NET proxy files) to construct a message to the .NET SWINGInteropEngine process running on the client.
Once the message is received, the engine performs a lookup in the Global Assembly Cache (GAC) to try and load the appropriate Windows Form that matches the request. Part of this load process also includes setting of parameters to control initial visual elements on the form. If the Windows Form cannot be found in the Global Assembly Cache, the working directory of the SWINGInteropEngine process is used instead.
In our sample, the first form exposes a public property called Message. The Java SWING application can pass a name/value pair containing the name of this property and a value that this property should be set to when the form is loaded. Multiple name/value pairs can be sent from the calling application via a hashtable.
In addition to setting of properties, the Java SWING application can nominate an existing public property on the form as a return value when the form closes. When the form is closed, the engine reflects against the form to find out the values of public properties that were set during the lifetime of the form. In the sample, the Java SWING application can ask for one of these values to be returned as part of the .NET Remoting call. This allows the Windows form to set some values, and for these to be returned to the calling application.
The engine logs calls and new form requests to the event log. The admin utility can be used to monitor these entries in the event log and display them.
The Java SWING application uses a sample API to make the calls to invoke the Windows Forms (it constructs a message, which is sent to the SWINGInteropEngine process), and to abstract the underlying .NET Remoting calls.
This API is currently in a package called com.microsoft.samples.windowsforms.
The main methods exposed by this API are as follows:
WindowsForms.OpenChannel();
This opens a channel to the running SWINGInteropEngine process.
WindowsForms.SetAssemblyDetails(String assemblyName, String namespace);
If all of the forms you are opening are stored in one assembly, this is a convenient way of pre-setting the assembly name and namespace such that simpler DisplayForm methods can be used.
WindowsForm.DisplayForm(String formName, Boolean modal);
This will display the form with the form name of formName. Modality is controlled through the Boolean, modal.
For example, WindowsForm.DisplayForm("Form1",true);
WindowsForm.DisplayForm("Form1",true);
WindowsForm.DisplayForm(String formName, Hashtable props, String
returnProperty, Boolean modal);
This will display the form with the given form name. Any keys in the hashtable that match public properties on the form will be set when the form is loaded. The value of any public property in the form that matches the name of returnProperty will be returned by this method when the form is closed. Modality is controlled through the Boolean, modal.
WindowsForm.DisplayForm("Form1", props, "Message", true);
This will display Form1 as a modal form, passing a hashtable of properties (props) to the form. Any keys in the hashtable that match public properties on the form will be set. The value of the public property (Message) will be returned when the form is closed. The form will be modal (Note: As mentioned, this is required for returned property values from the Windows form).
As with any sample, there are of course limitations and some room for improvement. Through writing this sample, the following could be considered:
Currently, passed parameters and return types from Windows Forms can only be strings. Given mappings of primitive types available in Intrinsyc's Ja.NET, it is possible to extend this to include other primitives such as integers, floats and so forth.
For more complex types, we would need to think about some kind of serialization—formatting the type into a document or style that can be understood by both sides.
When the form loads, the user needs to currently set the form to be "on top" and "in focus" to avoid it appearing underneath the Java SWING form.
As the form is being loaded by a third party (in this case, the SWINGInteropEngine process), on occasions, the new form can appear either underneath other windows, or inactive. Changing the form properties (for example, the TopMost property) can help resolve this.
Just because we are using .NET Remoting, do we need to use a TCP channel for inter-process communication?
When we look at performance of the sample, it is reasonable to argue that a TCP channel between two processes on the same machine could be an overhead. For single instance forms and normal user interaction, this is unlikely to be an issue (as a delay of a few milliseconds to open a new form probably won't annoy that many users!), but the argument is valid if the forms were being opened as part of a more automated approach.
To help address this, Intrinsyc are working on a shared memory channel for the next release of their Ja.NET product. This will prevent the call having to rely on the entirety of the network stack and should make for a more optimized solution.
This sample has demonstrated how .NET Remoting can be used on the same machine to allow Windows Forms to be invoked and displayed from an existing client application that uses Java SWING.
For organizations that are looking to migrate client applications from Java SWING to Windows Forms, this can be a valuable approach to immediately start deploying Windows Forms without the potential expense of waiting until the initial application is completely rewritten.
Note that since publication, the sample code presented in this article has also been made available for use with JNBridge Pro, a .NET to Java bridging solution. You can download this sample code and documentation from. | http://msdn.microsoft.com/en-us/architecture/ms954835.aspx | crawl-002 | refinedweb | 2,583 | 64 |
Cross Platform C#
Did you know you can combine C#, F#, Visual Basic, Razor, HTML and JavaScript within a single app?
It's well known by now that you can use C# to write apps for iOS and Android by leveraging Xamarin tools, especially if you've been following this column. It's also well documented that you can take a lot of that C# over to other platforms such as Windows Phone and Windows Store, but did you know you can use F#, Visual Basic, Razor, HTML, and JavaScript in your apps, as well?
One of the great things about the Microsoft .NET Framework platform is the variety of languages at your disposal, but you don't need to leave that flexibility behind when writing mobile apps. In this article, I'll show how you can combine each of these languages within a single native iOS app, starting with a standard Xamarin.iOS project written in C# as the base, and then bring in each of the other languages to add functionality.
F#
F# has been around for nearly a decade now, but has really been gaining a lot of steam in the last couple years in many areas, especially because it's been extended to a variety of platforms like iOS, Android, Windows Phone, Mac OS X and more. I'll start off by creating a new F# Portable Class Library (PCL) named FSharpLib and add a file named MathHelpers.fs to it:
namespace FSharpLib
type MathHelpers =
static member Factorial (num:int): int64 =
match num with
| 1 -> int64(1)
| num -> int64(num) * MathHelpers.Factorial(num - 1)
One of the nicest features of F# is its powerful type system, which makes it a great fit for subjects that rely heavily on units of measure, such as math and physics. It's often said that once your code satisfies the type system and compiles, it's very likely it will work correctly.
F# code also tends to be quite succinct, allowing developers to express solutions succinctly, resulting in simple, maintainable code. Writing a factorial function like my example wouldn't result in many lines in any language, but the result is readable and easy to understand.
With the class library in place, I can now add a reference to it from the iOS app and start using it from C#. For the UI, I'll create a simple screen with a textbox for taking in a number, a label for displaying the resulting factorial and a button to trigger the calculation, as shown in Figure 1.
With that in place, I can use it from the view controller, written in C# (see Listing 1).
using System;
using MonoTouch.UIKit;
using FSharpLib;
namespace MultiLanguageDemo
{
partial class FSharpViewController : UIViewController
{
public FSharpViewController (IntPtr handle) : base (handle)
{
}
partial void Calculate_TouchUpInside(UIButton sender)
{
if (string.IsNullOrWhiteSpace(Number.Text))
return;
Answer.Text = MathHelpers.Factorial(int.Parse(Number.Text)).ToString();
}
}
}
Because FSharpLib is a portable class library it can be used directly from C# code the same way you would with a C# library. Xamarin also supports F# as a first class language for iOS and Android, so you can actually write your entire app in F# if you want to!
Visual Basic
I actually swore off writing Visual Basic years ago, but for you, dear reader, I'll make an exception. Once again taking advantage of PCLs, I'll create one named VisualBasicLib and add a class named UserRepository, as shown in Listing 2.
Option Strict On
Public Class UserRepository
Private _users As XElement =
<users>
<user>Greg Shackles</user>
<user>Wallace McClure</user>
</users>
Public Function ListUsers() As IEnumerable(Of String)
Return From user In _users...<user>
Select user.Value
End Function
End Class
In this class I take advantage of the powerful XML functionality in Visual Basic, providing a method that queries from XML and returns a collection of users contained in it. For the UI, I'll keep it simple and display the list in a single label, putting a new line in between each item (see Listing 3).
using System;
using MonoTouch.UIKit;
using VisualBasicLib;
namespace MultiLanguageDemo
{
partial class VBViewController : UIViewController
{
private readonly UserRepository _repository = new UserRepository();
public VBViewController (IntPtr handle) : base (handle)
{
}
public override void ViewDidLoad()
{
base.ViewDidLoad();
UserList.Text = string.Join(Environment.NewLine, _repository.ListUsers());
}
}
}
The result is Figure 2.
Razor, HTML and JavaScript
If you've used ASP.NET before you're probably already familiar with the Razor templating engine. Even though it's typically associated with ASP.NET applications, the Razor engine can actually be used independently of it, as well.
One of the main selling points of Xamarin is it provides a fully native experience, which is true, but something that often goes overlooked is that it also provides a great hybrid app experience. Xamarin provides a preprocessed Razor item template type that you can use in your apps, allowing you to weave in Web views wherever they make sense in your apps, giving you the best of both worlds, as shown in Figure 3. Because the template is pre-processed, it generates C# code and, therefore, can be shared across multiple platforms, making it even more powerful.
Just like in ASP.NET apps, Razor templates can accept a model type to keep things strongly typed. For this example, I'll define a simple model that stores data for a person, including their name, job title and a list of interests:
using System.Collections.Generic;
namespace MultiLanguageDemo.Web
{
public class PersonModel
{
public string Name { get; set; }
public string Title { get; set; }
public string Company { get; set; }
public IList<string> Interests { get; set; }
}
}
With the model defined, I can create a simple Razor view that will display the data, as shown in Listing 4.
@model MultiLanguageDemo.Web.PersonModel
<html>
<head>
<style type="text/css">
body {
color: #c92727;
font-family: HelveticaNeue-Bold;
}
</style>
</head>
<body>
<h1>@Model.Name</h1>
<p>
<em>@Model.Title</em><br />
<strong>@Model.Company</strong>
</p>
<h3>Interests:</h3>
<ul>
@foreach (var interest in Model.Interests) {
<li>@interest</li>
}
</ul>
</body>
</html>
Because the HTML will be loaded into a Web view on the target platform, you can take advantage of anything you'd normally use on the client-side of a site like JavaScript and CSS. You can even bring in your favorite libraries for both, such as jQuery or Bootstrap.
The pre-processed template will generate a C# class named Person that exposes a GenerateString method to produce the HTML for the view, which can then be supplied to the Web view, as shown in Listing 5.
using System;
using System.Collections.Generic;
using MonoTouch.Foundation;
using MonoTouch.UIKit;
using MultiLanguageDemo.Web;
namespace MultiLanguageDemo
{
partial class RazorViewController : UIViewController
{
public RazorViewController (IntPtr handle) : base (handle)
{
}
public override void ViewDidLoad()
{
base.ViewDidLoad();
var person = new PersonModel
{
Name = "Greg Shackles",
Title = "Senior Engineer",
Company = "Olo",
Interests = new List<string>
{
"Mobile development",
"Heavy metal",
"Homebrewing"
}
};
var html = (new Person {Model = person}).GenerateString();
WebView.LoadHtmlString(html, NSBundle.MainBundle.BundleUrl);
}
}
}
The resulting view will look like Figure 4.
The fun doesn't stop there, though! It's actually possible to communicate between your apps and the Web view in both directions using JavaScript. To start off, I'll add a button to the top-right corner of the screen that will pass JavaScript code into the view, which will then execute immediately:
NavigationItem.RightBarButtonItem =
new UIBarButtonItem(UIBarButtonSystemItem.Play,
(sender, args) =>
WebView.EvaluateJavascript("alert('Hello from C# land!');"));
When the button is tapped it will raise an alert box through the Web view, which is showin in Figure 5. This is a trivial example, but you can execute any JavaScript you want this way.
In addition to being able to use JavaScript to talk from C# to the Web view, you can also communicate from the Web view back to your application by navigating. To start, I'll update the Web view to customize how navigations are handled:
WebView.ShouldStartLoad += (view, request, type) =>
{
if (!request.Url.ToString().StartsWith("myapp://pidgin"))
return true;
new UIAlertView("Message from JavaScript", request.Url.Query, null, "Ok", null).Show();
return false;
};
When a navigation occurs, this method will be invoked. If the new URL matches the predefined scheme, it will cancel the navigation and execute code based on the data provided in the URL. If the URL doesn't match, the Web view is allowed to process it normally.
Finally, I just need to add some code to the Razor template to trigger this navigation. Just like in normal Web development, this can be done through standard links, but it can also be triggered through JavaScript to enable more complicated scenarios:
<p>
<a href="#" id="TalkToCSharp">Say hello to C#</a>
</p>
<script>
(function() {
document.getElementById("TalkToCSharp").addEventListener("click", function () {
window.location = "myapp://pidgin?hello";
}, false);
})();
</script>
When the link is tapped, an alert box is showed through C# in the view controller, displaying the query string sent with the URL, as shown in Figure 6.
This only scratches the surface of what you can do with Web views in an app, as you can leverage the full power of the Web where it makes sense in your apps. It's also worth noting that you aren't limited to views that take up the entire screen. You can mix Web views into your app any way you like, such as for individual items in a native list.
Wrapping Up
As demonstrated here, there's a wide variety of languages available to you when building mobile apps. You can easily weave them together within a single app to take full advantage of each where they make sense. It also means that you can bring in existing code you have in any of these languages and leverage them in ways and platforms you might not even have thought of when you first wrote it!
Printable Format
> More TechLibrary
I agree to this site's Privacy Policy.
> More Webcasts | https://visualstudiomagazine.com/articles/2014/11/01/polyglot-apps.aspx | CC-MAIN-2018-30 | refinedweb | 1,651 | 52.9 |
mbed library sources
hal/sleep_api.h
- Committer:
- bogdanm
- Date:
- 2013-08-05
- Revision:
- 13:0645d8841f51
- Parent:
- 10:3bc89ef62ce7
File content as of revision 13:0645d8841f51:
/* MBED_SLEEP_API_H #define MBED_SLEEP_API_H #include "device.h" #if DEVICE_SLEEP #ifdef __cplusplus extern "C" { #endif /** Send the microcontroller to sleep * * The processor is setup ready for sleep, and sent to sleep using __WFI(). sleep(void); /** Send the microcontroller to deep sleep * * This processor is setup ready for deep sleep, and sent to sleep using __WFI(). This mode * has the same sleep features as sleep plus it powers down peripherals and clocks. All state * is still maintained. * * The processor can only be woken up by an external interrupt on a pin or a watchdog timer. * * deepsleep(void); #ifdef __cplusplus } #endif #endif #endif | https://os.mbed.com/users/lzbpli/code/mbed-src/file/0645d8841f51/hal/sleep_api.h/ | CC-MAIN-2022-27 | refinedweb | 124 | 65.42 |
If you are new to using NET Reflector, or you are wondering whether it would be useful to you, you'll appreciate Jason Crease's quick run through of the basic functionality. This is also available as a video here
When you first run .NET Reflector, you will be asked pick a version of the NET Framework to populate the assembly list.
When you want to browse an assembly, all you will need to do is to run NET Reflector, drag the file from Windows explorer and drop it onto NET Reflector. You can also open and .exe, .dll or .mcl file from the file menu in .NET Reflector.
Then you can drill down into namespaces, class and functions and down there, you have a bit of information about each of the components – you can see where the assembly is, its name, version, what type of assembly it is…
Here, we are browsing System.Windows.Forms
The main functionality of .NET Reflector is the disassembler. So let’s say I want to understand what happens during a particular event because something is not behaving the right way, or my app is not behaving as expected. You can drill down to the method of your choice, right-click and choose Disassemble. And there you have it:
You may, at this stage, need to choose the version of NET Framework that the assembly was developed with by using the View->options dialog box.
you can see the code for that function and see what it does in the Disassembler pane on the right.
I can choose to view the code in C#, or VB.NET or IL code.
The other cool thing you can do with .NET Reflector is navigating the code view. You can browse the source code by clicking on a function name and that’s really useful because you can explore code quickly, and understand what’s happening behind the scenes.
You can navigate the disassembly view,so if I click here in the onclick method
... it will take me to method OnClick
Right down there, we can see that we have the ability to Expand Methods.
When we click on it, it expands and decompiles all the methods in one go.
The other thing you can do with .NET Reflector, is to Analyze a class or a function, so say we pick the function called OnClick(EventArgs) and we right click and choose the command Analyze, we’ll have the Analyzer come up here on the right and we can see what that function depends on, and see the functions that it uses…so we can see the interaction between the different functions.
It’s the same principle with a Class. If you pick a class, say Button, you can see the same thing plus, we can also see what the type is exposed by and instantiated by.
You can also access documentation with .NET Reflector. So if I pick an assembly for which there’s xml documentation available, we can see the documentation for that class or function appearing down on the right. Here we’ll look at a class from ANTS Profiler.
.NET Reflector also has a search functionality (<View<Search). You can set it so it searches on classes or functions.
You’ll also note that .NET Reflector lets you access and save resource files.
Finally, one last thing I wanted to show you is the Windows Shell integration. We’ll close Reflector, bring up the command line (run – cmd), we’ll drag the Reflector.exe in here and add “ /register”
Now if we find a dll and right-click on it, we get a “Browse with .NET Reflector” in the context menu!
If you press Control Alt C when you have selected an item in an assembly in NET Reflector’s browser, you will have a reference on your clipboard that is just like a URL. If you paste it into any web-page or browser-based page, you can then use the URI as a short-cut to bring that item up in NET Reflector. | https://www.simple-talk.com/dotnet/.net-tools/first-steps-with-.net-reflector/ | CC-MAIN-2014-15 | refinedweb | 682 | 70.94 |
German language chennai työt
Chennai Harvest Ecommerce Website
German voice over...
Develop a 32‐bit single or multi‐cycle CPU capable of performing a search for prime numbers. CPU
I want to create a game for a potential app idea - need help in coding and creating everything. awaits
..
You will be given 8 videos each is about 1 minute, and required to edit them to a 2 minutes only total video duration. You must be good in Arabic language. translation of English article into German [kirjaudu nähdäksesi URL:n] length of article is 2800 words. I Please share your best quote.
...
I need an Android app. I would like it designed and built.
Loooking for copywriters who can write texts about gambling (online casino) in the following languages: German Finnish Norwegian Italian Slovak Portuguese Dutch
want to market our business in online. I want to get somebody from Chennai - Tamil Nadu desire to innovate in serving our clients and.
Website is ready, now need to develop Hybrid app. I am in Chennai, so i prefer only from Chennai and surrounding area Freelancers..
...Technical]
We have an existing system that is developed using C# ASP.NET MVC and uses MS SQL as i..
Hello, I need a virtual assistant for some German Works. German is a necessary point what you need. Work is to upload in our shopsystem new products. Best regards
You have to recreate a map using r geo_sf() function and ggplot2(). Needed data sets are given. Very Simple
.. and budget $60- $80 USD. Let me know if you could do it for me. Thanks!
...week/month? Do you deliver the English texts in a grammatically correct version? Do you have translators or do you know translators who can also translate the texts into German? We also offer Spanish and French. What we don't need are spun texts. Then cooperation would make no sense. We really need professional texts. Thank you in advance for your
Hi I would like to add the Arabic language to the user interface and dashboard for Magento script version 2.3.1 with the directions set from right to left my website link : [kirjaudu nähdäksesi URL:n] Thanks you.
import 2 xml of different language , WPML setup , site is already done
Looking for a full time on premises developer in chennai, India. Salary 25-30K P.M. Not looking for web development firms. Looking to hire an individual.
.. of Tuesday 21st. Please let me know your price, your language and that you agree are looking for native speakers to write and translate* 81. 7. It should be fu... | https://www.fi.freelancer.com/job-search/german-language-chennai/ | CC-MAIN-2019-22 | refinedweb | 435 | 68.26 |
Memory Allocation Without the GC
Unity!
To recap, C# has reference types and value types. Reference types are class instances and value types are primitives like
int and structs. Reference types are created on the heap, reference counted, and garbage collected. Value types are created on the stack unless they’re fields of classes.
So it’s good practice in Unity to avoid using a lot of reference types because they ultimately create garbage that the garbage collector has to collect, which fragments memory and creates frame spikes. So we turn to structs instead of classes because they don’t create garbage when we use them as local variables. Think of
Vector3,
Quaternion, or
Color.
We can use these structs as local variables and pass them as parameters with or without the
ref keyword. But those are temporary operations. Sometimes we’d like to keep them longer, so we add a struct field to a class or we create an array or
List of structs. Unfortunately, that class or array or
List is itself going to become garbage someday. And so it seems we’re painted into a corner and must accept the inevitability of the GC.
Thankfully, we have a way out! There is a way to allocate memory without using the
new operator and without using any classes, including arrays. The way to do it is to call
System.Runtime.InteropServices.Marshal.AllocHGlobal. You pass it the number of bytes you’d like to allocate and it allocates them on the heap and returns you an
IntPtr struct. When you’re done, call
Marshal.FreeHGlobal and pass it the
IntPtr you got back from
AllocHGlobal.
In the meantime, you can do whatever you want with that
IntPtr. You’ll probably want to turn on “unsafe” code by putting
-unsafe in your Asset’s directory’s
gmcs.rsp,
smcs.rsp, and
mcs.rsp. Then you can cast the
IntPtr to a pointer and access the memory however you want. Here’s a little example where we store a struct on the heap:
using System; using System.Runtime.InteropServices; using UnityEngine; struct MyStruct { public int Int; public bool Bool; public long Long; } unsafe void Foo() { // Allocate enough room for one MyStruct instance // Cast the IntPtr to MyStruct* so we can treat the memory as a MyStruct var pStruct = (MyStruct*)Marshal.AllocHGlobal(sizeof(MyStruct)); // Store a struct on the heap! *pStruct = new MyStruct { Int = 123, Bool = true, Long = 456 }; // Read the struct from the heap Debug.Log(pStruct->Int + ", " + pStruct->Bool + ", " + pStruct->Long); // Free the heap memory when we're done with it Marshal.FreeHGlobal((IntPtr)pStruct); }
Other than the
Debug.Log, this code doesn’t create any garbage. We can store the
MyStruct* long-term just like a reference to a class. Copying a pointer is cheap, too. A
MyStruct* is just 4 or 8 bytes regardless of how big
MyStruct gets.
Of course we can do whatever else we want with the heap memory we get back from
AllocHGlobal. Want to replace arrays? Easy!
using System; using System.Runtime.InteropServices; using UnityEngine; /// <summary> /// An array stored in the unmanaged heap /// /// </summary> unsafe struct UnmanagedArray { /// <summary> /// Number of elements in the array /// </summary> public int Length; /// <summary> /// The size of one element of the array in bytes /// </summary> public int ElementSize; /// <summary> /// Pointer to the unmanaged heap memory the array is stored in /// </summary> public void* Memory; /// <summary> /// Create the array. Its elements are initially undefined. /// </summary> /// <param name="length">Number of elements in the array</param> /// <param name="elementSize">The size of one element of the array in bytes</param> public UnmanagedArray(int length, int elementSize) { Memory = (void*)Marshal.AllocHGlobal(length * elementSize); Length = length; ElementSize = elementSize; } /// <summary> /// Get a pointer to an element in the array /// </summary> /// <param name="index">Index of the element to get a pointer to</param> public void* this[int index] { get { return ((byte*)Memory) + ElementSize * index; } } /// <summary> /// Free the unmanaged heap memory where the array is stored, set <see cref="Memory"/> to null, /// and <see cref="Length"/> to zero. /// </summary> public void Destroy() { Marshal.FreeHGlobal((IntPtr)Memory); Memory = null; Length = 0; } } unsafe void Foo() { // Create an array of 5 MyStruct instances var array = new UnmanagedArray(5, sizeof(MyStruct)); // Fill the array for (var i = 0; i < array.Length; ++i) { *((MyStruct*)array[i]) = new MyStruct { Int = i, Bool = i%2==0, Long = i*10 }; } // Read from the array for (var i = 0; i < array.Length; ++i) { var pStruct = (MyStruct*)array[i]; Debug.Log(pStruct->Int + ", " + pStruct->Bool + ", " + pStruct->Long); } // Free the array's memory when we're done with it array.Destroy();
One downside of this approach is that we need to make sure to call
FreeHGlobal or the memory will never be released. The OS will free it all for you when the app exits. One issue crops up when running in the editor because the app is the editor, not your game. So clicking the Play button to stop the game means you might leave 100 MB of memory un-freed. Do that ten times and the editor will be using an extra gig of RAM! You could just reboot the editor, but there’s a cleaner way so you don’t even need to do that.
Instead of calling
AllocHGlobal and
FreeHGlobal directly, you can insert a middle-man who remembers all of the allocations you’ve done. Then you can tell this middle-man to free them all when your app exits. Again, this is only necessary in the editor so it’s good to use
#if and
[Conditional] to strip out as much of the middle-man as possible from your game builds.
using System; using System.Diagnostics; using System.Runtime.InteropServices; #if UNITY_EDITOR using System.Collections.Generic; #endif /// <summary> /// Allocates and frees blocks of unmanaged memory. Tracks allocations in the Unity editor so they /// can be freed in bulk via <see cref="Cleanup"/>. /// /// </summary> public static class UnmanagedMemory { /// <summary> /// Keep track of all the allocations that haven't been freed /// </summary> #if UNITY_EDITOR private static readonly HashSet<IntPtr> allocations = new HashSet<IntPtr>(); #endif /// <summary> /// Allocate unmanaged heap memory and track it /// </summary> /// <param name="size">Number of bytes of unmanaged heap memory to allocate</param> public static IntPtr Alloc(int size) { var ptr = Marshal.AllocHGlobal(size); #if UNITY_EDITOR allocations.Add(ptr); #endif return ptr; } /// <summary> /// Free unmanaged heap memory and stop tracking it /// </summary> /// <param name="ptr">Pointer to the unmanaged heap memory to free</param> public static void Free(IntPtr ptr) { Marshal.FreeHGlobal(ptr); #if UNITY_EDITOR allocations.Remove(ptr); #endif } /// <summary> /// Free all unmanaged heap memory allocated with <see cref="Alloc"/> /// </summary> [Conditional("UNITY_EDITOR")] public static void Cleanup() { foreach (var ptr in allocations) { Marshal.FreeHGlobal(ptr); } allocations.Clear(); } }
Now just replace your calls to
Marshal.AllocHGlobal and
Marshal.FreeHGlobal with calls to
UnmanagedMemory.Alloc and
UnmanagedMemory.Free. If you want to use the
UnmanagedArray struct above, you might want to do this replacement in there as well. The final step is to put this one-liner in any
MonoBehaviour.OnApplicationQuit:
UnmanagedMemory.Cleanup();
Hopefully this technique will be a good addition to your toolbox as a Unity programmer. It can certainly come in handy when you need to avoid the GC!
#1 by Ed Earl on February 6th, 2017 · | Quote
Maybe you could ensure allocations are freed with a little RAII class?
#2 by jackson on February 6th, 2017 · | Quote
Can you provide a little example of how that might work in C#?
#3 by Ed Earl on February 8th, 2017 · | Quote
The idea is to tie the lifecycle of a resource, in this case a block of unmanaged memory, to the lifecycle of an object. In C++ that’s done with RAII; in C#, the Dispose pattern:
(untested – apologies for typos!)
Of course, each UnmanagedMemHolder object allocation will then add to GC pressure.
Hopefully there’s a final GC pass when you exit the Unity Editor/runtime – otherwise this doesn’t help much! (Should probably test that…)
#4 by jackson on February 8th, 2017 · | Quote
The GC won’t ever call your
Disposefunction. Users have to call it manually, but they do get access to the
usingblock syntax sugar:
Instead of this:
So
Disposegets called for them. That’s a handy convenience and the code looks a lot cleaner, but it’s functionally equivalent to the
Destroyand
Cleanupmethods in the article. On the downside, as you point out, it adds garbage for the
UnmanagedMemHolderclass instance. So you’d never want to use it on types that come and go often throughout the app, like
UnmanagedArray.
One change I could see would be for the
UnmanagedMemoryclass to be made non-static and implement a destructor:
The destructor would be called when the
UnmanagedMemoryinstance is garbage-collected, which would hopefully only be when the application exits. Therefore, it seems like the usage would be to create one in a
MonoBehaviourthat’s active in the first scene, store it as a field, never destroy or deactivate that
MonoBehaviour, then set the field to
nullin
OnApplicationQuit. You’d then either need to make
UnmanagedMemorya singleton or pass instances of it around to anyone who needs to use it. It’s hard to see how either way provides any benefits over the static class. Perhaps without the singleton it’s easier to unit test.
You’re definitely right about C++ RAII. It’s destructors are totally different from the destructors in C#. They run at well-defined times and the classes you put them in don’t create any garbage because C++ isn’t a garbage-collected language. We don’t have anything like that for C# though. We don’t even have access to destructors in structs. So we just remember to call our
Disposefunctions.
#5 by Ed Earl on February 8th, 2017 · | Quote
You’re right, Dispose isn’t called automatically. Thanks for clearing that up. Finalize is, though, so it might be possible to free unmanaged memory in an implementation of Finalize.
The thing I like about the idea of tying the lifecycle of unmanaged memory to the lifecycle of a managed object is that it confers some of the advantages of a managed memory approach, such as not having to remember to make an explicit call in order to free memory. Have to write a working implementation, though, which I failed to do ;)
#6 by jackson on February 8th, 2017 · | Quote
C# destructors are essentially finalizers. You could use them to free the unmanaged memory. Unfortunately that would mean that you need to use a class to get garbage collected, which kind of defeats the point in the case of types like
UnmanagedArray.
If you like the C++ RAII approach there are two ways you can get it. First, you could use C++/CLI to produce .NET DLLs. I’m not so sure how well that would work with Unity and IL2CPP, but it might. Also, “pure” (no native) DLL support is deprecated in the C++/CLI compiler so you’ll probably run out of support soon. Second, you could call into native code via P/Invoke (e.g.
[DllImport]) and then write your own real, native C++ there. There’s potentially a lot of work to the “glue” layer between C# and C++, but it’s an option if you really want to escape .NET.
#7 by Ed Earl on February 9th, 2017 · | Quote
Yep, current projects are using a lot of native plugins.
#8 by ms on February 22nd, 2017 · | Quote
damn this some heavy stuff (including the comments by ed), super dope — thanks for the schoolin’
#9 by ms on February 22nd, 2017 · | Quote
How would one go about tracking (I assume) managed allocations? Would this be anytime we use the ‘new’ keyword? Does it even make sense to? Can we track the lifetime of a resource? I apologize if any of these questions are naive or redundant.
Thanks,
m
#10 by jackson on February 22nd, 2017 · | Quote
Questions of all levels are welcome! :-D
Managed allocations (e.g.
new MyClass()) are tracked by the garbage collector. It keeps track of how many references there are to an instance of an object. Sometime after there are no more references to an object the garbage collector will reclaim the memory those objects were using. Currently, Unity’s garbage collector is very slow, runs on the main thread, collects all the garbage at once, and causes memory fragmentation. This is the reason why I have written so many articles with strategies to avoid creating any garbage.
On the contrary,
AllocHGlobalallocates unmanaged memory. The garbage collector doesn’t know about this memory. It doesn’t keep a reference count for it and it never reclaims it. It’s entirely your responsibility to call
FreeHGlobalwhen you’re done with the memory you asked for with
AllocHGlobal. That can be very error prone but also very efficient, so the article is there to inform you of that option.
#11 by m on March 3rd, 2017 · | Quote
Thanks for the response!
My follow up question, albeit perhaps naive, is: can we get a count of the GC managed references at will? For example, say simply for printing the count to the console on demand.
A second question, semi-related to this article and topic: do we need to work in unmanaged territory to reap the benefits of SoA (struct of array) design? Or are we guaranteed contiguous (parallel?) blocks of memory when defining several arrays as fields in a c# class? Or is that only a guarantee in a c# struct? Or no guarantee at all?
For reference see: ‘managing data relationships’ by noel llopis. His article (as well as this talk at gdc 2017 by tim ford) has inspired me as of late and all this talk of memory management I feel is related. I am curious because traditionally we only read about DOD from a c++ perspective, not from a c# perspective.
Have you explored DOD in the context of unity?
Cheers!
#12 by jackson on March 3rd, 2017 · | Quote
I don’t know of a way to get the number of managed objects, but you can get the total managed memory size with
System.GC.GetTotalMemory. The
GCclass has other useful functionality, such as
Collectto force garbage collection (e.g. on a loading screen).
For SoA, you can easily make a class or struct containing arrays. These arrays are managed references to
Arrayobjects that contain the pointer to the actual memory (i.e. array of values) as well as other data such as the
Lengthof the array. It’s the C++ equivalent of a
structcontaining multiple
std::shared_ptr<std::vector<T>>, except the
shared_ptris garbage-collected instead of reference counted.
If you want to contain the actual array in your class or struct, you have two options. First, you can use a fixed array of a compile-time known number of primitives (e.g.
int) in a struct. That’s pretty restrictive, but it may work in some use cases. Second, you can use unmanaged memory as described in this article to store pointers directly in your class or struct. You can then treat those pointers as arrays as you choose. If you need them to be contiguous, simply allocate (i.e. with
AllocHGlobal) a block of memory large enough for all your arrays and then set the pointers to offset into the memory you get back. For example, if you need three arrays of contiguous floats you can do this:
Unmanaged memory is a lot more flexible, but has the downsides of needing to explicitly free it, double-frees, using freed memory, etc.
As for DOD, it’s really only discussed among C/C++/Assembly programmers because they tend to be the ones who care most about squeezing out the last bits of CPU performance. If that’s your concern, C# is a terrible option. It’s just not designed for that purpose. That said, there are certain “escape hatches” you can use to get around a lot of its overhead. This article discusses one of them: use
AllocHGlobal/
FreeHGlobalto avoid the GC entirely. Structs, “unsafe” code, and pointers are more such “escape hatches”. If you avoid most of the language and .NET library then you can pull off some semblance of DOD in C#, especially with IL2CPP.
Thanks for the reference to the article and the talk. :)
#13 by m on March 5th, 2017 · | Quote
Thanks for the response man, very insightful and interesting. Looking forward to the next article.
#14 by benjamin guihaire on September 10th, 2017 · | Quote
with Unity games, often at loading time, while for example deserializing protobuf , the need for memory heap gets really, really high… so internally Unity bump the heap size, but unfortunately never shrink it again, so we get memory allocated for nothing, for ever, and less memory for system memory (textures, audio …)
So .. using your technique, I am wondering if its going to also grow the same heap used by the Managed Memory .. in that case, maybe we could call Marshal.CoTaskMemAlloc() instead of AllocHGlobal ?
#15 by jackson on September 11th, 2017 · | Quote
“Deserializing protobuf” makes me think you’re talking about the managed heap since Unity doesn’t provide any Protocol Buffers functionality as far as I’m aware. In that case,
AllocHGlobalshould be fine since it allocates unmanaged memory. I assume it’s a pass-through to a native function like
malloc, but it might be implemented by a native memory manager. Given that the doc for
AllocCoTaskMemsays it’s for COM purposes, I’m not sure how this is implemented on non-Windows platforms like iOS and Android. It may just be a synonym for
AllocHGlobal.
#16 by benjamin guihaire on September 12th, 2017 · | Quote
it looks like allocCoTaskMem and AllocHGlobal both call malloc.
#17 by Kailang Fu on September 21st, 2018 · | Quote
This IS the solution that I’ve been looking for.
I’ll rewrite all my music synthesis stuffs in unsafe code.
Writing C in C# is so beautiful.
#18 by Stephan on October 21st, 2018 · | Quote
Great stuff!
If I’m using Unity is there a way to make sure the memory I allocate is low in the heap so I don’t badly fragment the heap? All I can think is to allocate that memory as early as possible.
#19 by jackson on October 21st, 2018 · | Quote
No, both .NET’s
Marshal.AllocHGlobaland Unity’s
UnsafeUtility.Mallocdon’t give you any control over the location of the allocated memory. In the case of
UnsafeUtility.Malloc, you can at least choose the
Allocatorso perhaps
Tempor
TempJobwill be useful to you. | https://jacksondunstan.com/articles/3740 | CC-MAIN-2020-16 | refinedweb | 3,107 | 55.64 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.