issue_owner_repo listlengths 2 2 | issue_body stringlengths 0 261k ⌀ | issue_title stringlengths 1 925 | issue_comments_url stringlengths 56 81 | issue_comments_count int64 0 2.5k | issue_created_at stringlengths 20 20 | issue_updated_at stringlengths 20 20 | issue_html_url stringlengths 37 62 | issue_github_id int64 387k 2.46B | issue_number int64 1 127k |
|---|---|---|---|---|---|---|---|---|---|
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-17 11:35:20 +0200
From: @grobian
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2008-10-15 03:22:46 +0200
## Comment 13435
Date: 2005-06-17 23:35:20 +0200
From: @grobian
It would be nice if adding and dropping documents could
be done over an xquery connection instead of via a
separate MIL connection. It would also be nice if the
server could be provided with a document (given over
the connection) to store in the database, such that
connectivity APIs can easily add documents.
## Comment 13436
Date: 2008-10-15 15:22:46 +0200
From: @drstmane
We now (for quite a while) have pf:add-doc() and pf:del-doc().
Cf.,
http://monetdb.cwi.nl/projects/monetdb/XQuery/QuickTour/DOCMGT/index.html
http://monetdb.cwi.nl/projects/monetdb/XQuery/Documentation/Adding-and-Deleting-Documents.htmlAdding-and-Deleting-Documents
Hence, I assume these fulfill the requested feature.
## Comment 13437
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1222921 at http://sourceforge.net/support/tracker.php?aid=1222921
| XQuery: convenience functions for shredding and dropping doc | https://api.github.com/repos/MonetDB/MonetDB/issues/2410/comments | 0 | 2020-11-30T10:42:21Z | 2024-06-28T13:43:24Z | https://github.com/MonetDB/MonetDB/issues/2410 | 753,364,591 | 2,410 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-10 04:39:36 +0200
From: Wouter Alink <<alink>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2009-12-04 04:29:30 +0100
## Comment 13432
Date: 2005-06-10 16:39:36 +0200
From: Wouter Alink <<alink>>
To make the XQuery demo's more attractive I would like
to add more weigth to the implementation of the
Processing Instruction feature:
processing-instruction { $a } { $b }
Reason is, that this would allow an xquery to be
written with an 'XSLT' style instruction, and then we
could pretty print the results as html.
If someone can give me a clue on how to implement it, i
will try it myself. (which parts need modification?)
## Comment 13433
Date: 2009-12-04 16:29:30 +0100
From: @mlkersten
This is not likely to be realised in the M4/XQ/PF step anymore.
## Comment 13434
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1218276 at http://sourceforge.net/support/tracker.php?aid=1218276
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQ: Processing Instructions | https://api.github.com/repos/MonetDB/MonetDB/issues/2409/comments | 0 | 2020-11-30T10:42:17Z | 2024-06-28T13:43:24Z | https://github.com/MonetDB/MonetDB/issues/2409 | 753,364,545 | 2,409 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-07 06:49:57 +0200
From: @arjenpdevries
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2011-04-28 16:38:57 +0200
## Comment 13426
Date: 2005-06-07 18:49:57 +0200
From: @arjenpdevries
It would be nice to be able to access data in MonetDB
without first going through dumping the data as XML and
then shredding it.
The idea could be implemented by introducing, e.g., a
monetdb:// URI, that could refer to a sequence of items
stored in a [void,T] bat.
The preferred way is probably to use
fn:collection(URI), followed by fn:data(); however, I
am not sufficiently familiar with XQuery to judge if
that is indeed the intend of these functions.
A more hacked solution could be to change the shredder
such that it recognises such monetdb:// URIs, and
always wraps the table (without actual processing) in a
<bat><item>value</item>...</bat> document.
Cheers,
Arjen
## Comment 13427
Date: 2006-11-22 23:02:16 +0100
From: @Teggy
Logged In: YES
user_id=642632
Originator: NO
Hi,
this sounds like a reasonable feature to me and fn:collection() indeed seems a possible hook here. fn:data() would play no role in this context, though, as I see it (fn:data() would return a sequence of atomic values and what you want is a piece of XML).
We could even go all the way and return the BAT contents in the format that is used by other RDBMS for publishing tables in XML format (the 'FOR XML' feature in DB2, Sybase, or SQL Server):
-- Element normal form:
<table>
<row><head>42</head><tail>foo</tail></row>
<row><head>35</head><tail>bar</tail></row>
...
<table>
-- Attribute normal form:
<table>
<row head="42" tail="foo"/>
<row head="35" tail="bar"/>
...
</table>
Cheers,
--Teggy
## Comment 13428
Date: 2006-11-23 11:05:10 +0100
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
Don't we need to include the type of the columms?
Hence:
-- Element normal form 1:
<table>
<row><head type="integer">42</head><tail type="string">foo</tail></row>
<row><head type="integer">35</head><tail type="string">bar</tail></row>
...
<table>
-- Element normal form 2:
<table>
<row htype="integer" ttype="string"><head>42</head><tail>foo</tail></row>
<row htype="integer" ttype="string"><head>35</head><tail>bar</tail></row>
...
<table>
-- Attribute normal form:
<table>
<row htype="integer" ttype="string" head="42" tail="foo"/>
<row htype="integer" ttype="string" head="35" tail="bar"/>
...
</table>
But now, the Attribute normal form looks messy...
Regards,
Jennie
## Comment 13429
Date: 2006-11-23 11:16:37 +0100
From: @Teggy
Logged In: YES
user_id=642632
Originator: NO
If we do include types (sounds reasonable -- the SQL 'FOR XML' constructs omit types IIRC) then these could be properties of the table, not the individual rows (at least in an attribute normal form), I'd say:
<table htype="xs:integer" ttype="xs:string">
<row head="..." tail="..."/>
</table>
In the element normal form, though, having the type annotations attached to the individual column values could benefit the XML validation process which evaluates xsi:type annotations, e.g.:
<table>
<row>
<head xsi:type="xs:integer">42</head>
<tail xsi:type="xs:string">foo</tail>
</row>
</table>
Cheers,
--Torsten
## Comment 13430
Date: 2006-11-23 11:54:44 +0100
From: @yzchang
Logged In: YES
user_id=341633
Originator: NO
> <table htype="xs:integer" ttype="xs:string">
> <row head="..." tail="..."/>
> </table>
Teggy, you are so right. This form looks much better. Properties of a BAT can all be added as attributes of <table>, if necessary. We can actually define an XML schema to specify the mapping from a BAT to an XML document.
But I don't understand why the XML validation process can benefit from the element normal form, but not from this attribute normal form.
> <table>
> <row>
> <head xsi:type="xs:integer">42</head>
> <tail xsi:type="xs:string">foo</tail>
> </row>
> </table>
If we have polymorphic BAT, it is definitely necessary to specify the type of each value. Otherwise, I don't see the need to have the type annotations attached to the individual values. Do we have polymorphic BAT?
Regards,
Jennie
## Comment 13431
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1216531 at http://sourceforge.net/support/tracker.php?aid=1216531
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 15767
Date: 2011-04-28 16:38:57 +0200
From: @sjoerdmullender
Development on MonetDB4 and MonetDB/XQuery has ceased. It is unlikely that this bug will be fixed.
The last release of MonetDB/XQuery and MonetDB4 is Mar2011.
| XQ: RFE: Access to MonetDB data | https://api.github.com/repos/MonetDB/MonetDB/issues/2408/comments | 0 | 2020-11-30T10:42:14Z | 2024-06-28T13:43:23Z | https://github.com/MonetDB/MonetDB/issues/2408 | 753,364,507 | 2,408 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-05-29 04:21:33 +0200
From: @grobian
To: Merovingian devs <<bugs-merovingian>>
Version: -- development
CC: alexo15, ashishk, @hannesmuehleisen, @PedroTadim, shady.shen.sh
Last updated: 2019-07-29 15:05:13 +0200
## Comment 13421
Date: 2005-05-29 16:21:33 +0200
From: @grobian
It would be nice if it were possible to let Mapi, SQL
server and XQuery server bind to another address than
127.0.0.1. In many cases 127.0.0.1 is too restrictive
to be of any use, and the other extreme (0.0.0.0) far
too wide. It would be nice if these connections could
be bound to an interface or ipaddress.
## Comment 13422
Date: 2006-07-23 11:16:04 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Is this feature request still relevant?
## Comment 13423
Date: 2006-07-23 16:15:33 +0200
From: @grobian
Logged In: YES
user_id=963970
yes, because we only do local or any. We should have a
config option that defines to with interface or IP to bind.
## Comment 13424
Date: 2007-11-10 17:40:47 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
In fact this is still an issue. Merovingian just suffers from the same problem, either all or just local.
## Comment 13425
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1210805 at http://sourceforge.net/support/tracker.php?aid=1210805
## Comment 13987
Date: 2010-05-04 13:05:17 +0200
From: @grobian
Since M5 is wrapped by Merovingian, this really only is an issue for Merovingian.
## Comment 17771
Date: 2012-10-09 14:03:35 +0200
From: @grobian
*** Bug #3156 has been marked as a duplicate of this bug. ***
## Comment 17776
Date: 2012-10-09 17:39:25 +0200
From: Ashish Kumar Singh <<ashishk>>
is there any way as of now to make MonetDB listen on localhost only on linux platforms ? from description of this ticket I am guessing there might be a way??
## Comment 17777
Date: 2012-10-09 17:42:28 +0200
From: @grobian
No, monetdbd can only bind to all interfaces. Unless you hack its sources (should be relatively easy, see connections.c). It needs code/options to define how to listen on what.
From your bug description, I think perhaps a simpler route, which switches on an ip/hostname (and thus allows only binding to 1 IP or all via 0.0.0.0) would already provide a start.
## Comment 17778
Date: 2012-10-09 17:45:40 +0200
From: Ashish Kumar Singh <<ashishk>>
Thanks for reply , I will have a look.
But it would be nice to have it configurable OTB by monetdb, what do you think?
Regards,
Ashish
## Comment 17779
Date: 2012-10-09 18:27:01 +0200
From: @grobian
it would be a monetdbd property, and a nice to have, indeed
## Comment 21753
Date: 2016-01-28 09:44:31 +0100
From: @hannesmuehleisen
Indeed this is still an issue. Monetdbd by defaults listens on all interfaces. The default should be that it only listens on localhost, with an option to listen to another or all interfaces. Given the well-known default admin user/pass combination, this is a potentially security risk and should be avoided.
## Comment 21755
Date: 2016-01-28 12:36:11 +0100
From: MonetDB Mercurial Repository <<hg>>
Changeset [0299a8888992](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=0299a8888992) made by Hannes Muehleisen <hannes@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=0299a8888992](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=0299a8888992)
Changeset description:
monetdbd binds to localhost by default, new settable property listenaddr for dbfarm
Bug #2407
## Comment 22159
Date: 2016-05-23 10:57:29 +0200
From: @sjoerdmullender
Feature has been implemented in the Jun2016 branch.
## Comment 26054
Date: 2018-01-02 06:39:51 +0100
From: shady.shen.sh
In windows environment, server is not started with monetdbd and thus this fix is not valid. listenaddr parameter doesn't take effect and the server always listen to 127.0.0.1, which doesn't work in most cases.
Please help fix the issue for windows environment.
Thanks and Best Regards
Shady
## Comment 26656
Date: 2018-10-25 18:31:09 +0200
From: Alex Ossipov <<alexo15>>
This is also an issue on Linux configured with ipv6 only.
Per documentation:
listenaddr
This property specifies an address that is allowed to connect to the server. The user can specify one IP(v6) address, or use the 0.0.0.0 notation to allow connections from everywhere. Defaults to localhost (127.0.0.1).
Attempt to set to ipv6 fails with address not recognized.
## Comment 27191
Date: 2019-07-29 15:05:13 +0200
From: @PedroTadim
In default branch is now possible to bind to IPv6 addresses. However binding to a specified interface is not possible yet.
| Merovingian: allow binds to given ip/interface | https://api.github.com/repos/MonetDB/MonetDB/issues/2407/comments | 0 | 2020-11-30T10:42:10Z | 2024-06-28T13:43:22Z | https://github.com/MonetDB/MonetDB/issues/2407 | 753,364,460 | 2,407 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-05-29 02:30:42 +0200
From: @grobian
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2005-06-05 02:14:15 +0200
## Comment 13418
Date: 2005-05-29 14:30:42 +0200
From: @grobian
To be more inline with SQL and Mapi, pfstart() should
be pf_start().
To start SQL and Mapi, one does sql_server_start() and
mapi_start() respectively. One would expect
pathfinder_server_start(), or something similar, with
at least an underscore.
## Comment 13419
Date: 2005-06-05 14:14:15 +0200
From: @grobian
Logged In: YES
user_id=963970
added xquery_server_start() and xquery_server_start(bit
global) to completely comply with the SQL module.
## Comment 13420
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1210760 at http://sourceforge.net/support/tracker.php?aid=1210760
| XQuery: pstart() should be pf_start() | https://api.github.com/repos/MonetDB/MonetDB/issues/2406/comments | 0 | 2020-11-30T10:42:07Z | 2024-06-28T13:43:21Z | https://github.com/MonetDB/MonetDB/issues/2406 | 753,364,416 | 2,406 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-05-17 11:36:55 +0200
From: Ger Teunis <<gteunis>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-04 11:51:56 +0100
## Comment 13414
Date: 2005-05-17 11:36:55 +0200
From: Ger Teunis <<gteunis>>
I really like the MonetDB project, good and fast database.
I'd really like to use it in my C projects, are there
any plans in implementing a ado.net driver?
## Comment 13415
Date: 2005-05-17 11:49:06 +0200
From: @grobian
Logged In: YES
user_id=963970
Hi,
there is an ODBC driver available. Is it possible to use
this driver with ADO.net? I found a page which suggests
something like that, but I'm not sure, maybe it makes sense
to you:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpguide/html/cpconconnectingtoodbcdatasourceusingadonet.asp
## Comment 13416
Date: 2009-12-04 11:51:56 +0100
From: @mlkersten
There are no plans for the ADO.net implementation in the foreseable future.
## Comment 13417
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1203423 at http://sourceforge.net/support/tracker.php?aid=1203423
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| ADO.net driver | https://api.github.com/repos/MonetDB/MonetDB/issues/2405/comments | 0 | 2020-11-30T10:42:03Z | 2024-06-28T07:25:48Z | https://github.com/MonetDB/MonetDB/issues/2405 | 753,364,374 | 2,405 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-24 05:15:10 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2007-11-10 11:40:55 +0100
## Comment 13410
Date: 2005-02-24 17:15:10 +0100
From: Wouter Alink <<alink>>
When requesting declared variables (using vars(); ),
ALL variables are returned. This is a long list with
all system- and module-related variables.
In my opinion a user is most of the time only
interested in a section of these variables. I would
therefore propose to implement the following MIL-functions:
vars_all(); - would return all declared variables
vars(); - would return only locally declared variables
vars("system"); - would return all variables declared
by MonetDB itself
vars("<<module_name>>"); - would return all variables
declared in a given module
This would make the list a lot more comprehensive.
## Comment 13411
Date: 2007-11-10 11:40:55 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
MIL specific feature requests are dropped.
No intension to support it anywhere in the future.
## Comment 13412
Date: 2007-11-10 11:52:40 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Does MAL by any chance provide such a feature?
## Comment 13413
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1151150 at http://sourceforge.net/support/tracker.php?aid=1151150
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| comprehensive list of variables | https://api.github.com/repos/MonetDB/MonetDB/issues/2404/comments | 0 | 2020-11-30T10:42:00Z | 2024-06-28T07:25:48Z | https://github.com/MonetDB/MonetDB/issues/2404 | 753,364,342 | 2,404 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-22 10:07:17 +0100
From: @sjoerdmullender
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2019-04-30 12:36:05 +0200
## Comment 13405
Date: 2005-02-22 10:07:17 +0100
From: @sjoerdmullender
It might be nice to have the ability to open and read
remote files through the HTTP protocol. We should take
a look at curl to see if that can be integrated into
stream.mx.
The reason for wanting to do this is so that we can
make demos easily usable without having to distribute
them with the binaries.
## Comment 13406
Date: 2005-02-22 10:38:20 +0100
From: @grobian
Logged In: YES
user_id=963970
shouldn't this functionality be just in the client, instead
of the server?
## Comment 13407
Date: 2005-02-22 10:48:26 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Yes. This functionality is needed at several levels.
1) mm objects should be represented in the DBMS with a
handle and could be represented by urls to their origin. A
function getContent(() could then read the information into
a string/BAT/.... for further processing in the context of
the session.
(Additional functions could forward it to a browser)
2) the functionality of MonetDB would greatly improve if it
would be
possible to access read-only bats using a URL. Actually, the
access could also trigger a coercion to internal format. For
example, accessing a remote jpg file using the http protocol
delivers a BAT with the color representations (combined with
getWidth() and getHeight() on this file type you have an
image at your disposal.
## Comment 13408
Date: 2005-02-22 10:59:21 +0100
From: @grobian
Logged In: YES
user_id=963970
I don't see why for internal operations the stream library
would be used.
For demos, the customer just wants to execute a remote file:
a zip, a tar, a tar/gzip, a gzip, a plain text file. You
can do that by defining (non standard) handles within the
language you are demoing in, or just by keeping it simple
and letting the client fetch and feed to the Mserver.
For a demo, everything is already there, in fact it's just a
matter of "MapiClient | curl" or wget, or links, or lynx, etc.
## Comment 13409
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1146064 at http://sourceforge.net/support/tracker.php?aid=1146064
## Comment 26814
Date: 2019-01-21 14:45:08 +0100
From: @sjoerdmullender
In the next feature release there will be support for COPY INTO table FROM 'some:/url' ON CLIENT, if mclient is started with the --allow-remote option.
| stream: http read support in streams | https://api.github.com/repos/MonetDB/MonetDB/issues/2403/comments | 0 | 2020-11-30T10:41:56Z | 2024-06-27T11:45:13Z | https://github.com/MonetDB/MonetDB/issues/2403 | 753,364,301 | 2,403 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-21 03:49:42 +0100
From: @yzchang
To: MonetDB4 devs <<bugs-monetdb4>>
Version: -- development
Last updated: 2006-07-23 11:14:48 +0200
## Comment 13398
Date: 2005-02-21 15:49:42 +0100
From: @yzchang
If a MapiClient sends a (sql) request, after it is
already killed at the server side, Mserver will crash
with the error message: Segmentation fault.
I would expect that Mserver will display a message like
"unknown client" or just do nothing.
## Comment 13399
Date: 2005-02-21 16:22:54 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
I don't quite understand what you're doing.
Is it something like:
window 1: Mserver --dbinit "module(sql_server);
sql_server_start();"
window 2: MapiClient -lsql
window 1: quit();
window 2: select ...;
?
If I try this, MapiClient just exits.
## Comment 13400
Date: 2005-02-21 16:49:24 +0100
From: @yzchang
Logged In: YES
user_id=341633
I did the following:
window 1: Mserver --dbinit "module(sql_server);
sql_server_start();"
window 2: MapiClient -lsql
window1:
MonetDB>clients();
-----------------------------------------------------------------------------------------------------------------
client name login mil
name
int str str str
type
-----------------------------------------------------------------------------------------------------------------[
0, "adm", "Mon Feb 21 15:41:00 2005", "clients()"
]
[ 1, "adm", "Mon Feb 21 15:41:00 2005",
"sqllogmanager()"
]
[ 2, "adm", "Mon Feb 21 15:41:00 2005",
"mapi_listen(int(find(<VAR>,\"sql_port\")),5,\"SQLclient\")"
][ 3, "adm", "Mon Feb 21 15:41:00 2005",
"mapi_listen(int(find(<VAR>,\"mapi_port\")),5,\"\")"
][ 5, "adm", "Mon Feb 21 15:41:00 2005",
"sqlclient(<VAR>,<VAR>)"
]
MonetDB>kill(5);
window 2: select ...;
window 1:
MonetDB>Segmentation fault
## Comment 13401
Date: 2005-02-22 08:27:49 +0100
From: @njnes
Logged In: YES
user_id=43556
the client kill statement in mil is indeed broken. Mostly
because we
use pthread_kill which isn't very reliable. We probably need
to remove
the kill mil interface.
## Comment 13402
Date: 2005-03-23 16:07:38 +0100
From: @njnes
Logged In: YES
user_id=43556
Changed to a feature request. It is a long standing wish to
cleanup the client/thread implementation. This cleanup is or
will be done in the upcoming 5.0 version.
## Comment 13403
Date: 2006-07-23 11:14:48 +0200
From: @mlkersten
Logged In: YES
user_id=490798
In M5 each client has its own process, which means
that when it ceases to exist the client program
gets a connection error.
It won't be solved in M4 anymore.
## Comment 13404
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1145484 at http://sourceforge.net/support/tracker.php?aid=1145484
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-monetdb4@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Mserver crashes if killed (sql) client sends a request | https://api.github.com/repos/MonetDB/MonetDB/issues/2402/comments | 0 | 2020-11-30T10:41:53Z | 2024-06-28T07:25:48Z | https://github.com/MonetDB/MonetDB/issues/2402 | 753,364,264 | 2,402 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-17 03:15:26 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2006-11-22 10:22:34 +0100
## Comment 13394
Date: 2005-02-17 15:15:26 +0100
From: Wouter Alink <<alink>>
The topic says it all: the support for 8-bit bytes
(instead of CHR's which only support 7-bits) would be
nice ;)
## Comment 13395
Date: 2005-02-17 16:01:13 +0100
From: @yzchang
Logged In: YES
user_id=341633
unsigned char is 8 bits, and is supported. Could it be a
quit solution?
## Comment 13396
Date: 2006-11-22 22:22:34 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
The bte type was introduced, which replaced the "broken" chr, more or less. Of course the bte type is, like all other integer types in MonetDB signed, though. Additionally there is a wrd type of size size_t.
## Comment 13397
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1124737 at http://sourceforge.net/support/tracker.php?aid=1124737
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| BYTE | https://api.github.com/repos/MonetDB/MonetDB/issues/2401/comments | 0 | 2020-11-30T10:41:50Z | 2024-06-28T07:25:47Z | https://github.com/MonetDB/MonetDB/issues/2401 | 753,364,233 | 2,401 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-17 03:12:59 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-04 02:30:25 +0100
## Comment 13391
Date: 2005-02-17 15:12:59 +0100
From: Wouter Alink <<alink>>
The documentation ('/doc'-folder) is not included in
the Monet (4.6) source distribution. It might be a
good idea to create a separate .rpm for the documentation.
## Comment 13392
Date: 2009-12-04 14:30:25 +0100
From: @mlkersten
Out of date and obsolete.
## Comment 13393
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1124735 at http://sourceforge.net/support/tracker.php?aid=1124735
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Documentation Package | https://api.github.com/repos/MonetDB/MonetDB/issues/2400/comments | 0 | 2020-11-30T10:41:47Z | 2024-06-28T07:25:47Z | https://github.com/MonetDB/MonetDB/issues/2400 | 753,364,198 | 2,400 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-14 01:19:42 +0100
From: @drstmane
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-04 02:33:25 +0100
## Comment 13385
Date: 2005-02-14 13:19:42 +0100
From: @drstmane
To reduce maintenance overhead, we should remove all
html extraction/generation features from Mx.
Instead, we should limit Mx to latex & texi
extraction/generation; html can then be generated using
standard tools like latex2html.
## Comment 13386
Date: 2005-02-14 13:24:38 +0100
From: @grobian
Logged In: YES
user_id=963970
before doing that I'd suggest first to figure out how to
stop latex2html from making those annoying
one-page-per-chapter outputs.
## Comment 13387
Date: 2005-02-17 14:51:33 +0100
From: @drstmane
Logged In: YES
user_id=572415
Maybe, we should go even further, and also abandon the
[la]tex extraction, fooussing on texi output, only.
Other formats like [la]tex and html can then be generated
from the texi output.
Martin already started the work on the texi extraction, at
least for src/gdk/ .
Unfortunately, the literal texi derectives that had [to]
be[en] intorduced in the gdk*.mx file now break the tex &
html extraction, and hence break our website; see (e.g.)
http://monetdb.cwi.nl/TechDocs/Core/gdk/index.html
and bug report 1124595 "Kernel Guide not rendered correctly"
(http://sourceforge.net/tracker/index.php?func=detail&aid=1124595&group_id=56967&atid=482468).
## Comment 13388
Date: 2009-12-01 11:56:02 +0100
From: @grobian
What's the impact of removing that? Is there still something being HTML generated?
## Comment 13389
Date: 2009-12-04 14:33:25 +0100
From: @mlkersten
Obsolete. Mx remains the internal tool. User documentation will be re-organised in a CMS.
## Comment 13390
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1122341 at http://sourceforge.net/support/tracker.php?aid=1122341
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Mx: remove all HTML extraction/generation feature | https://api.github.com/repos/MonetDB/MonetDB/issues/2399/comments | 0 | 2020-11-30T10:41:44Z | 2024-06-28T07:25:47Z | https://github.com/MonetDB/MonetDB/issues/2399 | 753,364,153 | 2,399 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 03:11:50 +0100
From: @swingbit
To: MonetDB4 devs <<bugs-monetdb4>>
Version: -- development
Last updated: 2009-12-04 02:31:36 +0100
## Comment 13379
Date: 2005-02-11 15:11:50 +0100
From: @swingbit
As far as I know, only equality and range selects are
efficiently implemented.
A non-equality select (that I find a very common
operation) can be performed as:
[ifelse]([=](b,value),b);
This is very inefficient though.
the result of [=](b,value) is materialized before the
actual selection, which can be very costly if b is a
large bat.
I'd find very useful a native implementation of
non-equality selections.
## Comment 13380
Date: 2005-02-11 15:26:02 +0100
From: @drstmane
Logged In: YES
user_id=572415
good point, though I don't know when this feature will be
added...
for now,
union(
select(b,nil,value,FALSE.FALSE),
select(b,value,nil,FALSE.FALSE)
);
or
select(b,nil,value,FALSE.FALSE).access(BAT_APPEND).insert(select(b,value,nil,FALSE.FALSE).access(BAT_READ);
might be some (more efficient?) alternatives...
## Comment 13381
Date: 2005-02-11 15:33:41 +0100
From: @swingbit
Logged In: YES
user_id=883451
No problem for the delay.
Unfortunately these two alternatives might be even more
expensive, if the original bat was ordered on the head and
you want to preserve this property (which is often the
case): in that case a sort() would be necessary
afterwards.... :(
## Comment 13382
Date: 2009-12-01 11:57:38 +0100
From: @grobian
M4 is pretty deprecated, so this request may as well be closed, perhaps
## Comment 13383
Date: 2009-12-04 14:31:36 +0100
From: @mlkersten
Correct. this will not be implemented anymore
## Comment 13384
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120789 at http://sourceforge.net/support/tracker.php?aid=1120789
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-monetdb4@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| M4: non-equality select | https://api.github.com/repos/MonetDB/MonetDB/issues/2398/comments | 0 | 2020-11-30T10:41:41Z | 2024-06-28T07:25:46Z | https://github.com/MonetDB/MonetDB/issues/2398 | 753,364,112 | 2,398 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-11 01:22:52 +0100
From: @yzchang
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-01 11:59:05 +0100
## Comment 13374
Date: 2005-02-11 13:22:52 +0100
From: @yzchang
Would it be handy to write the debug output of the
MapiClient to a (user specified) file, instead of on
the screen?
## Comment 13375
Date: 2005-02-11 13:52:26 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Which debug output are you referring to. I am not aware of
any in the normal use of MapiClient.
## Comment 13376
Date: 2005-02-11 14:17:19 +0100
From: @grobian
Logged In: YES
user_id=963970
I thought about the -d option (--debug)...
## Comment 13377
Date: 2009-12-01 11:59:05 +0100
From: @grobian
it's not obvious what's being meant here
## Comment 13378
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1120735 at http://sourceforge.net/support/tracker.php?aid=1120735
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Write debug output of MapiClient to file | https://api.github.com/repos/MonetDB/MonetDB/issues/2397/comments | 0 | 2020-11-30T10:41:37Z | 2024-06-28T07:25:46Z | https://github.com/MonetDB/MonetDB/issues/2397 | 753,364,074 | 2,397 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-09 10:50:46 +0100
From: @grobian
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2006-11-22 10:25:52 +0100
## Comment 13369
Date: 2005-02-09 10:50:46 +0100
From: @grobian
Currently Mtest by default uses the dbfarm directory of
the installed Mserver. This results in a 'pollution'
of the dbfarm directory with mTest databases. I think
it would be great if mTest either cleaned up its own
mess in the dbfarm, or (better) used a different dbfarm
directory (for instance in the mTest output root) so it
doesn't interfere at all with existing databases and
cannot possibly mess up anything there.
## Comment 13370
Date: 2005-02-09 14:34:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
Agree,
...
but (;-)
- Mtest has a --dbfarm option (see Mtest.py --help)
- Mtest needs to be sure that the dbfarm is writeable;
suggestions for an alternative defult are welcome, though ;-)
- Mtest does not clean it's "mess" because in case of errors,
the dbfarm might be handy to analyse the problem(s)
- we might consider make the clean-up default behaviour,
and add a --keep (or alike) switch for debugging...
Stefan
## Comment 13371
Date: 2005-02-09 14:40:16 +0100
From: @grobian
Logged In: YES
user_id=963970
ok, I had not noticed the --dbfarm option
I think a --keep option is not necessary, as I do agree that
the dbfarm might be handy to analyse problems. I think it's
sufficient to place the dbfarm by default in the mTest
output directory (as primarily suggested) as that directory
needs to be writable in any case (otherwise mTest can't
report and such).
## Comment 13372
Date: 2006-11-22 22:25:52 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
this can be achieved by setting a few flags pointing to another location then the usual dbfarm, it is only a problem for tests which start an mserver outside the mtest wrapper
## Comment 13373
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1119187 at http://sourceforge.net/support/tracker.php?aid=1119187
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| A cleaner mTest | https://api.github.com/repos/MonetDB/MonetDB/issues/2396/comments | 0 | 2020-11-30T10:41:34Z | 2024-06-28T07:25:46Z | https://github.com/MonetDB/MonetDB/issues/2396 | 753,364,040 | 2,396 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-02-08 06:10:23 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-06-10 02:50:24 +0200
## Comment 13367
Date: 2005-02-08 18:10:23 +0100
From: Wouter Alink <<alink>>
In the documentation only starting MapiClient with
sql_server is explained. This should (in my opinion) be
explained for mil-only first.
something like:
module(mapi);
mapi_start();
where mapi_start() would do:
fork(mapi_server(50000,1)); or something similar.
## Comment 13368
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1118758 at http://sourceforge.net/support/tracker.php?aid=1118758
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| documentation for starting mapi | https://api.github.com/repos/MonetDB/MonetDB/issues/2395/comments | 0 | 2020-11-30T10:41:31Z | 2024-06-28T07:25:45Z | https://github.com/MonetDB/MonetDB/issues/2395 | 753,364,005 | 2,395 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-01-15 12:19:39 +0100
From: @mlkersten
To: @sjoerdmullender
Version: -- development
Last updated: 2005-11-15 10:59:23 +0100
## Comment 13355
Date: 2005-01-15 00:19:39 +0100
From: @mlkersten
Timing commands is helpful for comparison.
However, it does not provide the timing information when
you load a script, e.g.
sql> <Tapestry-1K-2
[requires addition of a few more timerEnd() calls]
Also, the command line options are handy as toggle
during a session as well, e.g.
sql> \T toggles the timer
sql> \t toggles interaction trace
sql> \d 5 sets debugging more
....
sql> \dump to dump the database to a file
sql> \restore .... equivalent to <dumpfile + caretaking
sql> \s silent mode, to ignore all 'rows affected'
messages
## Comment 13356
Date: 2005-01-17 22:27:34 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Also, it would be handy on windows cmd line to just run the
script without entering into a session. Proposal:
MapiClient -i <file> initializes the session with <file>
MapiClient <file> just executes file and stops
## Comment 13357
Date: 2005-01-18 11:19:50 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
There were already commands \t - start timer, \T - stop
timer, but they were not documented in the help message, so
I changed \T to toggle the timer.
What exactly should \t - toggle interaction trace do? In
other words, what does interaction trace do?
I made \Dtable dump table. You can dump to file by
preceding this with >file.
What is the caretaking bit in \restore supposed to do?
Currently there is a --debug flag, but I don't think it does
much. It certainly doesn't communicate anything with the
server. What should it do (and what shoud \d do)?
Should \s just skip printing all lines that start with or
should it be more sophisticated?
I have implemented -i/--interactive. If a file is given on
the commandline, it is executed (as before), but now
MapiClient will exit afterwards unless -i is also passed on
the commandline.
I haven't checked this in yet.
## Comment 13358
Date: 2005-01-18 17:30:58 +0100
From: @mlkersten
Logged In: YES
user_id=490798
The -t (trace) option for MapiClient was used for dumping
the actual interaction between Mapi and the server.
Activating the trace in the Mapi library. This is handy when
you have to debug hand-shaking applications.
## Comment 13359
Date: 2005-01-18 17:34:52 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
There were more questions to be answered than just the one
about tracing...
## Comment 13360
Date: 2005-01-18 17:59:03 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Sorry,
actually --debug would be more in line to act with the rest.
i.e. the default could be to activate the trace. Refinements
to think about are one-way/two-way trace, debugging the
Mapi.mx code itself, passing the debug request as part of
the session initialization with the back-end...
For now simple --debug == -t would be more than sufficient
Thinking more about \D, this is backend dependent and can be
issued with minimal knowledge. It is therefor questionable
functionality
The \s is new to me. Trimming all comments might be a little
too much. For, comments are used in two places, to notify
the user on state of affairs (e.g.debugging info) and as
part of the answer header. It is certainly useful for
suppressing the table header.
What I missed in the analysis is something like
--log=<file>, to capture the interaction to be played back
at a later stage. Typical this is an afterthought, 'gee, if
I just had a log of this session I might reproduce the
error'. So perhaps we should offer one as a default.
## Comment 13361
Date: 2005-01-18 18:10:32 +0100
From: @mlkersten
Logged In: YES
user_id=490798
About logging. There are actually two versions
1) a log that keeps the requests sent (available by default)
2) a session log, with timing etc (should be set explicitly)
select count(*)
from tables
issued: 2005-01-18 17:59:00.000
[ 23]
replied:2005-01-18 17:59:00.034
## Comment 13362
Date: 2005-02-09 14:01:07 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Everything that is going to be implemented has been implemented.
It is unclear what \s should do, so when there are clear
ideas, they can be put in a new feature request.
There is no caretaking to be done after a restore, so there
is no need for \restore.
Setting debugger flags in the server is too language
specific, and setting the Mapi debugger flag can be done
with \t.
A logging feature really belong in a new feature request and
won't be done before the upcoming release.
## Comment 13363
Date: 2005-11-09 13:51:03 +0100
From: @njnes
Logged In: YES
user_id=43556
BugDay_2005-11-09, Niels Nes: NO TEST / FEATURE REQUEST
## Comment 13364
Date: 2005-11-09 14:36:28 +0100
From: @njnes
Logged In: YES
user_id=43556
restore proper order of bugs for the bug day
## Comment 13365
Date: 2005-11-15 10:59:23 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
This is a feature request and so I moved it there (again :-).
Also, since everything has been implemented that I was going
to implement, I also close this request. If any more
features need to be implemented, open a new feature request.
## Comment 13366
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1102702 at http://sourceforge.net/support/tracker.php?aid=1102702
| Feature completion | https://api.github.com/repos/MonetDB/MonetDB/issues/2394/comments | 0 | 2020-11-30T10:41:27Z | 2024-06-28T07:25:45Z | https://github.com/MonetDB/MonetDB/issues/2394 | 753,363,948 | 2,394 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-12-14 02:52:17 +0100
From: @MarcinZukowski
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2007-11-10 11:38:56 +0100
## Comment 13349
Date: 2004-12-14 14:52:17 +0100
From: @MarcinZukowski
Currently there's no way of quickly checking if BAT exists.
User needs to process view_bbp_name (undocumented) or
play nasty tricks like isnil(CATCH(isnil(bat(name)))).
Would be nice if either:
- bat(str) returned nil on error
- there would be a specialized function for that.
## Comment 13350
Date: 2005-03-01 21:54:28 +0100
From: @MarcinZukowski
Logged In: YES
user_id=607094
I really miss it.
## Comment 13351
Date: 2005-03-03 15:55:59 +0100
From: @drstmane
Logged In: YES
user_id=572415
Here's my comment from
http://sourceforge.net/tracker/index.php?func=detail&aid=1146347&group_id=56967&atid=482468
- The bat(<str>) cast correctly detects that no BAT with the
given name exists,
and hence, it correctly returns "bat_nil".
- However, "bat_nil" cannot not be handled by any MIL
command, as already the interpreter checks for "valid" BATs,
considers "bat_nil" as invalid, and issues the above error.
- Thus, not even `isnil(bat("frboierfwfr"));' can work ...
Two questions arise:
- Should the "bat(<str>)" cast should (in addition to
returning "bat_nil") also echo a warning?
This is not handy, in case the "bat(<str>)" cast is used
internally, and "bat_nil" return are handled elsewhere.
- Should we make "isnil(bat(nil))" work, which basically
means patching the interpreter somehow?
IMHO, both solutions are not good, but I don't have a better
one, yet ...
## Comment 13352
Date: 2007-11-10 11:38:56 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
MIL is deprecated and we are not going to support this feature.
## Comment 13353
Date: 2007-11-10 11:52:00 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Does MAL by any chance provide such a feature?
## Comment 13354
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1085123 at http://sourceforge.net/support/tracker.php?aid=1085123
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Easy way of checking if bat exists | https://api.github.com/repos/MonetDB/MonetDB/issues/2393/comments | 0 | 2020-11-30T10:41:24Z | 2024-06-28T07:25:44Z | https://github.com/MonetDB/MonetDB/issues/2393 | 753,363,900 | 2,393 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-10-20 12:48:14 +0200
From: Agustin Schapira <<schapirama>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-02-22 10:10:21 +0100
## Comment 13345
Date: 2004-10-20 00:48:14 +0200
From: Agustin Schapira <<schapirama>>
We have observed that a commit crashes Monet if it tries to access
the DELETE_ME directory on Windows while the Microsoft Indexing
Service process is running (the files are momentarily owned by the
Administrator? and Monet doesn't have the right access
permissions to delete them?). Maybe the temp_ files should be
stored in a temp/ directory that Windows knows about and
therefore doesn't lock?
## Comment 13346
Date: 2004-10-20 10:13:35 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
In the original e-mail report it was mentioned that putting
the files in the application data directory might also help.
Could you try invoking Mserver with the argument
--dbfarm=<path to application directory>\MonetDB
Perhaps we can figure out this path at install time and add
it to the invocation of Mserver (if it solves the problem).
## Comment 13347
Date: 2005-02-22 10:10:21 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
The current Mserver on Windows, if started through the
provided batch file (accessible from the Start menu) puts
the database in the Application folder.
I'm assuming this is enough of a fix so that I can close
this report.
## Comment 13348
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1050357 at http://sourceforge.net/support/tracker.php?aid=1050357
| Location of temp_ files on Windows | https://api.github.com/repos/MonetDB/MonetDB/issues/2392/comments | 0 | 2020-11-30T10:41:21Z | 2024-06-28T07:25:44Z | https://github.com/MonetDB/MonetDB/issues/2392 | 753,363,861 | 2,392 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-10-19 11:20:43 +0200
From: @grobian
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2006-11-22 10:24:01 +0100
## Comment 13342
Date: 2004-10-19 11:20:43 +0200
From: @grobian
In many real-life(tm) applications, a form of sequences
is onontbeerlijk. For lots of applications, it is just
very important that some unique id can be generated
(the SQL-alike way) in a certain column, which can be
referenced from other tables. This is the typical
primary key that does auto increment upon inserts.
Sequences guarantee that a certain value of the
sequence is only returned once. This allows concurrent
threads/connections to execute SELECT
next_value(some_sequence); and lateron using this
returned value in an INSERT statement. Because the
value returned by next_value() is guaranteed to be
unique no key violations can and will occur in such
situations.
## Comment 13343
Date: 2006-11-22 22:23:51 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
This has been properly implemented, including all kinds of dialect support, such as SERIAL, AUTO_INCREMENT, you name it.
## Comment 13344
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1049870 at http://sourceforge.net/support/tracker.php?aid=1049870
| SQL: support for sequences | https://api.github.com/repos/MonetDB/MonetDB/issues/2391/comments | 0 | 2020-11-30T10:41:18Z | 2024-06-28T07:25:44Z | https://github.com/MonetDB/MonetDB/issues/2391 | 753,363,824 | 2,391 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-10-09 04:18:58 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2009-12-01 12:01:27 +0100
## Comment 13334
Date: 2004-10-09 16:18:58 +0200
From: @grobian
For some tests (e.g. marcin2) in the SQL repo
significant changes are to be seen, as the MapiClient
utility shows different error warnings than it used to.
The actual change is visible in a query like:
-- this is some explanation
SELECT * FROM WRONG QUERY;
MapiClient used to bail about an error, reporting the
wrong SQL query. However, nowadays, it bails but shows
the comment instead of the query.
Either some tests have to be reapproved, or the
original behaviour of MapiClient should be restored.
It might be evident that that behaviour is much more
useful in most cases.
## Comment 13335
Date: 2004-10-09 17:00:31 +0200
From: @drstmane
Logged In: YES
user_id=572415
AFAIK, the current behaviour of MapiClient is actually the
"original" one: in case of an error, the first line after
the last (successful) statement/query is returned as the
query that cause the error. In case there is a comment
between the last succesful statement/query and the failing
statement/query, the comment is not ignored but returned as
the query casuing the error.
If I recall correctly from what Sjoerd mentioned, this is
not easily to be fixed...
The stable output is not the "original" one, but rather the
"desired" one.
The tests used to show "major" differences, until I found
this too annoying, and made Mfilter threat them as "minor"
differences just before releasing 4.4.0/2.4.0.
I'd vote for keeping the tests and stable output as is, to
remind us of the problem --- but since I'm no longer in
charge of testing, other opinions/solutions are welcome, too ;-)
## Comment 13336
Date: 2004-10-09 17:04:09 +0200
From: @grobian
Logged In: YES
user_id=963970
I'm aware of it difficult to solve, but a test should give a
"green" mark. If that is not possible, then I should say,
disable the whole test. This way we keep getting massive
email traffic.
## Comment 13337
Date: 2004-10-09 17:19:28 +0200
From: @drstmane
Logged In: YES
user_id=572415
Just for info:
Emails are only sent, if the test results differ between two
subsuquent days,
i.e., if a test that used to be work (be green), now shows
differences (orange/red), or vice versa. If all tests show
the same behaviour as the day before, no emails are sent ---
even if all tests (still) "fail significantly"!
We might consider disabling the respective tests in the
release branch, but I'd higly recomment to keep them running
in the main trunk, to keep us reminded of the current
situation/status.
(But as I said, my role concerning testing has been
"reduced" to at most been an advisor ... ;-)
## Comment 13338
Date: 2004-10-10 19:48:59 +0200
From: @njnes
Logged In: YES
user_id=43556
MapiClient shouldn't print the query. The sql server should
include the
query in the error result.
## Comment 13339
Date: 2004-10-16 09:43:58 +0200
From: @njnes
Logged In: YES
user_id=43556
This bug is releated to not having a proper 'sql'
communication protocol.
So I'll move this feature request to the proper section.
## Comment 13340
Date: 2009-12-01 12:01:27 +0100
From: @grobian
The intention of this request has been rereported by skinkie recently since this behaviour even got worse for some situations.
## Comment 13341
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1043573 at http://sourceforge.net/support/tracker.php?aid=1043573
| Testweb: SQL tests show minor differences due to MapiClient | https://api.github.com/repos/MonetDB/MonetDB/issues/2390/comments | 0 | 2020-11-30T10:41:15Z | 2024-06-28T07:25:43Z | https://github.com/MonetDB/MonetDB/issues/2390 | 753,363,785 | 2,390 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-08-08 08:16:02 +0200
From: @grobian
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2004-08-09 10:40:46 +0200
## Comment 13330
Date: 2004-08-08 20:16:02 +0200
From: @grobian
upon start-up:
Monet Database Server V4.3.19
Copyright (c) 1993-2004, CWI. All rights reserved.
Compiled for i686-pc-linux-gnu/32bit; dynamically linked.
Visit http://monetdb.cwi.nl for further information.
monet>module(sql_server);
wall-clock: 2 ms
wall-clock: 1 ms
wall-clock: 1 ms
wall-clock: 2 ms
wall-clock: 1 ms
wall-clock: 1 ms
wall-clock: 1 ms
wall-clock: 1 ms
wall-clock: 1 ms
wall-clock: 1 ms
wall-clock: 2 ms
wall-clock: 2 ms
monet>
I'd prefer if the debug output would be suppressed.
## Comment 13331
Date: 2004-08-08 20:20:43 +0200
From: @grobian
Logged In: YES
user_id=963970
Actually it's a bug, because it also sends this output in
the client/server intermezzo, causing weird errors and
unexpected output (trans log snippet):
<< monetdb:monetdb
>> [ "version", "4.3" ]
>> [ "language", "sql" ]
>> [ "schema", "sys" ]
>> ^A^A
<< SSET auto_commit = true;
>> wall-clock: 0 ms
>> ^A^A
<< SSET reply_size = 250;
>> wall-clock: 0 ms
>> ^A^A
<< sSELECT * FROM env;
>> name, value name
>> varchar, varchar type
>> 28, tuplecount
>> 0, id
>> [ "gdk_arch", "32biti686-pc-linux-gnu" ]
>> [ "gdk_version", "4.3.19" ]
>> [ "monet_pid", "26399" ]
>> [ "prefix", "/nfs/poseidon/Temp/monet/program" ]
>> [ "exec_prefix", "/nfs/poseidon/Temp/monet/program"
]
>> [ "gdk_dbname", "demo" ]
>> [ "gdk_dbfarm",
"/nfs/poseidon/Temp/monet/program/var/MonetDB/dbfarm"
:
## Comment 13332
Date: 2004-08-09 10:40:46 +0200
From: @drstmane
Logged In: YES
user_id=572415
done:
do wall-clock performance measeuring only if a
timing-related debug-mask (2,1024,2048) is set.
## Comment 13333
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1005577 at http://sourceforge.net/support/tracker.php?aid=1005577
| SQL: remove debug output | https://api.github.com/repos/MonetDB/MonetDB/issues/2389/comments | 0 | 2020-11-30T10:41:12Z | 2024-06-28T07:25:43Z | https://github.com/MonetDB/MonetDB/issues/2389 | 753,363,749 | 2,389 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-07-01 04:13:15 +0200
From: @peterboncz
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2007-11-10 11:37:06 +0100
## Comment 13324
Date: 2004-07-01 16:13:15 +0200
From: @peterboncz
currently the MIL module loading is split into two
categories:
- physical load (like dll loading/dropping)
- logical load (adding commands + types to fcn tables)
the MIL script initialization is now done only when a
module was not loaded before (physical load)
however, it should be done always (logical load).
now, global variables, and procs introduced by a module
are not seen if a second client loads a module.
## Comment 13325
Date: 2004-09-12 15:00:25 +0200
From: @njnes
Logged In: YES
user_id=43556
The module loading is indeed split in 2 parts
physical and logical loading
For any new module (ie combination of dll and mil part)
we go through 2 steps
## Comment 13326
Date: 2004-09-12 15:04:59 +0200
From: @njnes
Logged In: YES
user_id=43556
;-( I want vi for sf bug tracker and all other web forms)
load the mil part (this is needed to get depending
modules loaded)
during the mil load the dll/so is loaded,
using a module
statement.
As long as a module is loaded early (say before mapi) there
are no problems.
Problem is 'mapi 1' loads module en 'mapi 2' also loads module
both start with the load but one stops as the load (dll/mil)
is allready
done.
Its difficult detecting the difference between a (second)
mil load for a server/client case (ie module allready loaded
by server) and a client/client case is difficult.
It should check if the module is loaded by one of his
parents or not!
Conclusion is not for this release, work around is also way
to simple and
should be advocated as the only use of dynamic modules.
Solution is server allway loads modules before loading the mapi.
I think the best solution for the future is to disable
module loading
for clients other then admin and stop loading after mapi is
loaded!
(moved to feature requests)
## Comment 13327
Date: 2006-11-22 22:33:04 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
isn't this more or less implemented right now?
## Comment 13328
Date: 2007-11-10 11:37:06 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
It relates to M4, which is depricated.
## Comment 13329
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 983455 at http://sourceforge.net/support/tracker.php?aid=983455
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| init script should be part of logical module load | https://api.github.com/repos/MonetDB/MonetDB/issues/2388/comments | 0 | 2020-11-30T10:41:09Z | 2024-06-28T07:25:43Z | https://github.com/MonetDB/MonetDB/issues/2388 | 753,363,712 | 2,388 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-06-17 10:31:03 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-10-17 03:56:54 +0200
## Comment 13320
Date: 2004-06-17 10:31:03 +0200
From: @grobian
Life just only starts to get fun when one can mess
around with the delights of always confusing time zones.
Therefore I would like to have the ability to use time
zones on TIMESTAMP and TIME.
The SQL looks like:
CREATE TABLE international (
hire_date TIMESTAMP WITH TIME ZONE
);
See page 194; SQL Complete, Really
## Comment 13321
Date: 2004-09-01 20:57:09 +0200
From: @njnes
Logged In: YES
user_id=43556
monettime 'daytime' type misses time zone support.
## Comment 13322
Date: 2004-10-17 15:56:54 +0200
From: @grobian
Logged In: YES
user_id=963970
this has been implemented
## Comment 13323
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 974579 at http://sourceforge.net/support/tracker.php?aid=974579
| SQL: time zone support for times | https://api.github.com/repos/MonetDB/MonetDB/issues/2387/comments | 0 | 2020-11-30T10:41:05Z | 2024-06-28T07:25:42Z | https://github.com/MonetDB/MonetDB/issues/2387 | 753,363,665 | 2,387 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-06-05 06:35:46 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-09-01 08:58:23 +0200
## Comment 13317
Date: 2004-06-05 18:35:46 +0200
From: @grobian
Needed string functions:
ASCII(string) (int representation of the leftmost char
in the given string)
CHAR(code) (ascii representation of the given number
between 0 and 255)
DIFFERENCE(string1, string2) (int indicating the
difference between the values returned by the function
SOUNDEX for string1 and string2)
INSERT(string1, start, length, string2) (a character
string formed by deleting length characters from
string1 beginning at start and inserting string2 into
string1 at start)
LCASE(string) (lowercase string)
LEFT(string, count) (the count leftmost characters from
string)
LENGTH(string) (the number of characters in string
EXCLUDING trailing blanks (strlen(rtrim(string))) <- I
guess only for non-varchar string type to be useful)
LOCATE(string1, string2[, start]) (string1, searching
from the beginning of string2; if start is supplied,
the search begins from position start. 0 is returned if
string2 does not contain string1. Position 1 is the
first character in string2)
LTRIM(string) (string with leading blanks removed)
REPEAT(string, count) (a charachter string formed by
repeating string count times)
REPLACE(string1, string2, string3) (replaces all
occurrences of string2 in string1 with string3)
RIGHT(string, count) (the count rightmost characters in
string)
RTRIM(string) (string with no trailing blanks)
SOUNDEX(string) (data-source dependant SOUNDEX code)
SPACE(count) (a string of count spaces)
UCASE(string) (uppercase string)
## Comment 13318
Date: 2004-06-07 10:29:29 +0200
From: @arjan
Logged In: YES
user_id=20087
I have implemented SOUNDEX and DIFFERENCE already, in my
local checkout.
However, I have not included them in SQL cvs since they are
not part of the standard. Maybe we should wait with (some
of) these functions until Niels has implemented UDFs.
## Comment 13319
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 967165 at http://sourceforge.net/support/tracker.php?aid=967165
| JDBC String Functions | https://api.github.com/repos/MonetDB/MonetDB/issues/2386/comments | 0 | 2020-11-30T10:41:02Z | 2024-06-28T07:25:42Z | https://github.com/MonetDB/MonetDB/issues/2386 | 753,363,622 | 2,386 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-06-05 06:04:53 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-08-18 09:03:11 +0200
## Comment 13311
Date: 2004-06-05 18:04:53 +0200
From: @grobian
The CONVERT() function currently does not make any
sense. The CAST() function does already supply the
required behaviour. This might be hooked up to each other?
monetdb-> select convert('15', int);
Error on line 5: Identifier: int unknown or ambigious
Executed query: select convert('15', int);
monetdb-> select convert('15', integer);
Error on line 6: Identifier: integer unknown or ambigious
Executed query: select convert('15', integer);
monetdb-> select cast('15' as int);
+----------
| convert_single_value
+----------
| 15
+----------
1 rows
monetdb-> select cast('15' as integer);
+----------
| convert_single_value
+----------
| 15
+----------
1 rows
monetdb->
Weirdest of all, CONVERT seems to do something with one
argument:
monetdb-> select convert(timestamp '2004-06-05
17:54:01.000');
+----------
| convert_single_value
+----------
| 17:54:01.000
+----------
1 rows
??????
## Comment 13312
Date: 2004-06-05 18:15:40 +0200
From: @MarcinZukowski
Logged In: YES
user_id=607094
I know nothing about SQL frontend but it seems CONVERT order
of parameters is opposite to the one in CAST.
Hence, try CONVERT(int, '15') or CONVERT(int '15').
## Comment 13313
Date: 2004-06-05 18:17:55 +0200
From: @grobian
Logged In: YES
user_id=963970
for JDBC I need CONVERT(value, SQLtype)
## Comment 13314
Date: 2004-06-24 09:17:00 +0200
From: @njnes
Logged In: YES
user_id=43556
seems the convert isn't standard in sql (cast is!), anyway
having
converts resemble the ms-sql-server convert functions may make
conversion to MonetDB/SQL easier, we put this on the feature
request list.
## Comment 13315
Date: 2004-08-18 21:03:11 +0200
From: @njnes
Logged In: YES
user_id=43556
implemeted as rewrite to cast! Currently convert(datatype,
expression) and not convert(expression,datatype)
## Comment 13316
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 967148 at http://sourceforge.net/support/tracker.php?aid=967148
| SQL: CONVERT function not (fully) implemented? | https://api.github.com/repos/MonetDB/MonetDB/issues/2385/comments | 0 | 2020-11-30T10:40:59Z | 2024-06-28T07:25:42Z | https://github.com/MonetDB/MonetDB/issues/2385 | 753,363,591 | 2,385 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-06-05 03:47:55 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-09-01 08:58:42 +0200
## Comment 13309
Date: 2004-06-05 15:47:55 +0200
From: @grobian
The JDBC interface requires a driver to support a few
functions. Most of the functions are straight forward
and already supported. Some however aren't.
JDBC still requires the following numeric functions:
ATAN(float1, float2) (arctangent in radians of
float2/float1)
COT(float) (cotangent of float in radians)
DEGREES(number) (degrees in number radians)
PI() (the constant pi)
POWER(number, power) (number raised to (integer) power)
RADIANS(number) (radians in number degrees)
RAND(integer) (random floating point for seed integer)
ROUND(number, places) (number rounded to places places)
SIGN(number) (-1 when number is < 0, 0 when number = 0,
1 when dumber > 0)
TRUNCATE(number, places) (number truncated to places
places)
## Comment 13310
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 967070 at http://sourceforge.net/support/tracker.php?aid=967070
| JDBC numeric functions | https://api.github.com/repos/MonetDB/MonetDB/issues/2384/comments | 0 | 2020-11-30T10:40:56Z | 2024-06-28T13:43:21Z | https://github.com/MonetDB/MonetDB/issues/2384 | 753,363,549 | 2,384 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-05-11 01:18:42 +0200
From: @arjan
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2004-08-19 11:19:53 +0200
## Comment 13306
Date: 2004-05-11 13:18:42 +0200
From: @arjan
I have written a small patch so mapiclient remembers
the query history between sessions.
TODO:
* Make this behaviour configurable (on/off)
* Make history filename configurable (through monet.conf?)
Patch file is attached.
Note: on windows, the history file is saved in
prefix\etc; but there is no guarantee that the current
user can write there. Should go in his home dir (My
Documents?) instead.
## Comment 13307
Date: 2004-08-19 11:19:53 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
It seems Arjan implemented this while I wasn't looking.
## Comment 13308
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 951859 at http://sourceforge.net/support/tracker.php?aid=951859
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| MapiClient: saving query history between sessions | https://api.github.com/repos/MonetDB/MonetDB/issues/2383/comments | 0 | 2020-11-30T10:40:52Z | 2024-06-28T13:43:20Z | https://github.com/MonetDB/MonetDB/issues/2383 | 753,363,496 | 2,383 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-05-07 10:53:24 +0200
From: @mlkersten
To: SQL devs <<bugs-sql>>
Version: -- development
Last updated: 2011-01-25 15:21:17 +0100
## Comment 13304
Date: 2004-05-07 10:53:24 +0200
From: @mlkersten
The COPY [INTO/FROM] statement of SQL should be extended to
recognize a protocol tag, e.g.
copy into employee from 'http://mybackup.cwi.nl/emp1';
This could be extended to support:
create table e( h varchar, t varchar);
copy into e from 'mil:environment'; -- gets the much
needed info
Semantics should be such that you can import/export
tables from/to the MIL environment
create table t(i oid, n varchar, m varchar)'
copy into t(i,n) from 'mil:bat1';
copy into t(i,m) from 'mil:bat2';
## Comment 13305
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 949727 at http://sourceforge.net/support/tracker.php?aid=949727
## Comment 15428
Date: 2011-01-25 15:21:17 +0100
From: @grobian
The MIL environment is deprecated. Copy into from remote resources can easily be done using e.g. wget or curl in a pipe chain with mclient.
| SQL: protocol based import/export | https://api.github.com/repos/MonetDB/MonetDB/issues/2382/comments | 0 | 2020-11-30T10:40:50Z | 2024-06-28T13:43:20Z | https://github.com/MonetDB/MonetDB/issues/2382 | 753,363,464 | 2,382 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-04-17 07:47:19 +0200
From: @mlkersten
To: @njnes
Version: -- development
Last updated: 2004-08-14 01:58:38 +0200
## Comment 13301
Date: 2004-04-17 07:47:19 +0200
From: @mlkersten
The database directory should contain a status file,
which should include e.g. the latest version of Mserver
used.
A newer version can then check if the database needs to
be rebuild from a backup.
This may avoid unexpected effects like, segfault...
## Comment 13302
Date: 2004-08-14 13:58:38 +0200
From: @njnes
Logged In: YES
user_id=43556
implemeted a sql catalog version number
## Comment 13303
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 936782 at http://sourceforge.net/support/tracker.php?aid=936782
| Version compatability test | https://api.github.com/repos/MonetDB/MonetDB/issues/2381/comments | 0 | 2020-11-30T10:40:47Z | 2024-06-28T13:43:19Z | https://github.com/MonetDB/MonetDB/issues/2381 | 753,363,434 | 2,381 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-04-11 10:32:13 +0200
From: @grobian
To: SQL devs <<bugs-sql>>
Version: -- development
Last updated: 2006-11-22 10:31:45 +0100
## Comment 13297
Date: 2004-04-11 22:32:13 +0200
From: @grobian
When using COPY x RECORDS INTO ... the strings in the
data columns don't get checked for their lengths.
A column which is defined to hold at max one char can
hold as many as one likes by using the COPY INTO statement.
See attached file for a simple test which passes, but
should fail in order to obey the given table constraints.
## Comment 13298
Date: 2004-04-14 21:11:03 +0200
From: @njnes
Logged In: YES
user_id=43556
moved to 'feature requests'. The copy into is for bulkloading of
(dumped) ie correct data only.
## Comment 13299
Date: 2006-11-22 22:31:45 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
won't fix
## Comment 13300
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 933370 at http://sourceforge.net/support/tracker.php?aid=933370
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-sql@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| SQL: COPY INTO does not check string lengths | https://api.github.com/repos/MonetDB/MonetDB/issues/2380/comments | 0 | 2020-11-30T10:40:45Z | 2024-06-28T13:43:18Z | https://github.com/MonetDB/MonetDB/issues/2380 | 753,363,405 | 2,380 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-04-10 01:09:31 +0200
From: @grobian
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2004-05-07 09:17:43 +0200
## Comment 13291
Date: 2004-04-10 13:09:31 +0200
From: @grobian
For some applications it would be nice to be able to
send a special query (or a query on a virtual system
table) for retrieving some information about the server.
Some things that I would like to have some information
about (in the form of a regular result set would be nice!):
monet major version
monet minor version
monet version string
current 'database'
current logged in user
## Comment 13292
Date: 2004-04-10 15:02:47 +0200
From: @mlkersten
Logged In: YES
user_id=490798
This information is available using the command env();
This table has not been mapped to an SQL view and should
await a 'general' solution to import/export SQL<->MIL
A temporary patch to JDBC is to side-step the SQL
connection and ask for this table using a MIL connection.
## Comment 13293
Date: 2004-04-10 16:08:42 +0200
From: @grobian
Logged In: YES
user_id=963970
I don't have to, as there is an option to send M commands. I
just made a little hack in sql_session.mx to enable M
commands and it works. I only think there exist a way
without the hack to be allowed to send M commands...
## Comment 13294
Date: 2004-04-12 09:59:32 +0200
From: @mlkersten
Logged In: YES
user_id=490798
The COPY [INTO/FROM] statement of SQL should be extended to
recognize a protocol tag, e.g.
copy into employee from 'http://mybackup.cwi.nl/emp1';
This could be extended to support:
create table e( h varchar, t varchar);
copy into e from 'mil:environment'; -- gets the much needed info
Semantics should be such that you can import/export
tables from/to the MIL environment
create table t(i oid, n varchar, m varchar)'
copy into t(i,n) from 'mil:bat1';
copy into t(i,m) from 'mil:bat2';
## Comment 13295
Date: 2004-05-07 09:17:43 +0200
From: @grobian
Logged In: YES
user_id=963970
there exists now an env table in SQL, such that select *
from env; gives what I need.
## Comment 13296
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 932759 at http://sourceforge.net/support/tracker.php?aid=932759
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Mapi: ability to acquire some data about the server | https://api.github.com/repos/MonetDB/MonetDB/issues/2379/comments | 0 | 2020-11-30T10:40:41Z | 2024-06-28T13:43:18Z | https://github.com/MonetDB/MonetDB/issues/2379 | 753,363,352 | 2,379 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-04-10 12:37:48 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-08-19 12:15:55 +0200
## Comment 13289
Date: 2004-04-10 12:37:48 +0200
From: @grobian
In the current Mapi protocol, queries which don't
return a result set return simply nothing. For such
queries it is however a good thing to know the number
of rows which were affected by that particular query.
I feel this is a huge miss in the current implementation.
A possible reply from the server on an update query
could be:
affected_rows name
mediumint type
1 tuplecount
-1 id
[ 2 ]
please note that the negative id here indicates it is
an affected rows count.
## Comment 13290
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 932750 at http://sourceforge.net/support/tracker.php?aid=932750
| Mapi: return affected rows on resultsetless queries | https://api.github.com/repos/MonetDB/MonetDB/issues/2378/comments | 0 | 2020-11-30T10:40:37Z | 2024-06-28T13:43:59Z | https://github.com/MonetDB/MonetDB/issues/2378 | 753,363,310 | 2,378 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-31 12:55:22 +0200
From: Johan List <<biglebowski>>
To: @njnes
Version: -- development
Last updated: 2007-11-10 11:05:31 +0100
## Comment 13286
Date: 2004-03-31 12:55:22 +0200
From: Johan List <<biglebowski>>
It would be nice if it was possible to specify the
swapping directory for Monet. Now it uses the BAT
directory as swapping directory, but in some cases the
swapping hinders performance significantly.
For example, when working on the new cluster machine, I
would want to specify a local (on-disk) swapping
directory, instead of going to the RAID for swapping.
## Comment 13287
Date: 2007-11-10 23:05:31 +0100
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
this request about swapping is unclear. Swapping if needed should be done on the fasted device. The dbfarm with the bats can be in any directory (use the dbfarm env variable). Closing this request, it can be reopened if needed.
## Comment 13288
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 926601 at http://sourceforge.net/support/tracker.php?aid=926601
| Possibility for specifying swapping directory | https://api.github.com/repos/MonetDB/MonetDB/issues/2377/comments | 0 | 2020-11-30T10:40:34Z | 2024-06-28T13:43:58Z | https://github.com/MonetDB/MonetDB/issues/2377 | 753,363,277 | 2,377 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-31 12:51:43 +0200
From: @sjoerdmullender
To: @sjoerdmullender
Version: -- development
Last updated: 2004-04-02 04:25:27 +0200
## Comment 13283
Date: 2004-03-31 12:51:43 +0200
From: @sjoerdmullender
mapi_query and mapi_query_done use answer_lookahead to
look for errors.
Currently, answer_lookahead reads until the next prompt
and caches the result for later use. This is needed
since it is possible to send a new query before the old
query is completely dealt with (see
sql/src/odbc/samples/odbcsample1.c for an example using
ODBC on Mapi). However, if the reply is large, this
requires significant overhead.
This problem can be resolved in the following manner.
answer_lookahead reads until the first tuple or the
prompt, whichever comes first. When the application
then calls mapi_quick_response, there is only very
little data in the cache and the rest can be dumped to
file without going through the cache. When the
application starts another query, the library sees
there is still unread data. It can then read the data
into the cache, the way answer_lookahead does
currently, and then send the new query to the server.
For the simple use case of MapiClient reading from
standard input (using mapi_query_{prep,part,done}
followed by mapi_quick_response) this will have
significant speed benefits. More complex applications
will continue to work.
## Comment 13284
Date: 2004-04-02 16:25:27 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Implemented in revision 1.96 of Mapi.mx.
## Comment 13285
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 926597 at http://sourceforge.net/support/tracker.php?aid=926597
| Mapi interaction improvement | https://api.github.com/repos/MonetDB/MonetDB/issues/2376/comments | 0 | 2020-11-30T10:40:31Z | 2024-06-28T13:43:57Z | https://github.com/MonetDB/MonetDB/issues/2376 | 753,363,240 | 2,376 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-31 11:02:04 +0200
From: @sjoerdmullender
To: @njnes
Version: -- development
Last updated: 2005-04-05 01:54:13 +0200
## Comment 13280
Date: 2004-03-31 11:02:04 +0200
From: @sjoerdmullender
Strings in a string BAT are stored in a heap. A hash
is calculated for the string and this gives an index to
the head of a linked list of strings that hash to the
same value.
Currently the hash function that is used is strOldHash
which only looks at a few of the characters in a
string, and so it is very likely to produce the same
hash value for similar strings. There is also a
function strHash which makes a much better hash value,
but this can't be used for backward compatibility
reasons (persistent BATs using the old hash function
cannot be updated when the new hash function is being
used).
There is a way out, allowing the use of the new hash
function whilst staying compatible with old BATs.
When reading in a BAT (or when first updating, or
whenever--the best time would have to be determined),
check the first string in the heap. Try both the old
and the new hash function, and for both try to find the
string in the usual way. At least one of these methods
should succeed since the BAT was created with one or
the other hash function (so if it doesn't succeed, you
can scream loudly). If both hash functions succeed in
locating the string, try the second, third, etc.
strings. Eventually you find a string for which the
hash functions make a difference, and this determines
which hash function to use in the future for this
particular BAT. (If you run out of strings to check,
or if there are none, it doesn't matter which hash
function you use, so you can use the new one.)
Which hash function is used for the BAT must be stored
somewhere so that it is easy to choose the hash
function when it is needed. This could be a single bit
somewhere, or a pointer to the hash function. The
information is per-heap. There is no need to save it
on disk, since it is easy to determine (as above).
Using this technique makes it possible to read old
heaps with a new server. New heaps can't be read with
an old server.
## Comment 13281
Date: 2005-04-05 13:54:13 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
The old string hash has now been removed. It was decided
that backward compatibility isn't such a concern anymore,
and therefore the scheme described here has not been
implemented.
## Comment 13282
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 926546 at http://sourceforge.net/support/tracker.php?aid=926546
| string heaps should use better (new) hash function | https://api.github.com/repos/MonetDB/MonetDB/issues/2375/comments | 0 | 2020-11-30T10:40:28Z | 2024-06-28T13:43:57Z | https://github.com/MonetDB/MonetDB/issues/2375 | 753,363,204 | 2,375 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-26 10:17:25 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2004-10-17 03:58:44 +0200
## Comment 13275
Date: 2004-03-26 10:17:25 +0100
From: @grobian
>insert into test values(1, '"', 1.0);
>select * from test;
id, var, number name
int, varchar, real type
4, , tuplecount
1, , id
[ 1, "\"", 1 ]
>alter table test add anonumb int default 12;
>select * from test;
id, var, number, anonumb name
int, varchar, real, int type
4, , , tuplecount
2, , , id
>
Altering a table adding a column (with a default)
clears the table?!?
without the default:
>select * from test;
id, var, number, anonumb name
int, varchar, real, int type
5, , , tuplecount
4, , , id
[ 1, "\"", 1, 12 ]
>alter table test add bogus varchar(12);
>select * from test;
id, var, number, anonumb, bogus name
int, varchar, real, int, varchar type
5, , , , tuplecount
5, , , , id
further:
>alter table test drop anonumb;
MAPI = monetdb@localhost:45123
ACTION= mapi_get_row
QUERY = alter table test drop anonumb;
ERROR = !Unknown table element (140718144)->token = unknown
Drop is not implemented yet?
## Comment 13276
Date: 2004-03-26 11:07:31 +0100
From: @grobian
Logged In: YES
user_id=963970
Please note the (incrementing) tuple count, while no rows
are returned.
## Comment 13277
Date: 2004-03-26 15:58:55 +0100
From: @njnes
Logged In: YES
user_id=43556
A know fact that ALTER table is not complete! So asking for a
more complete ALTER table implementation is clearly a feature
request
## Comment 13278
Date: 2004-10-17 15:58:44 +0200
From: @grobian
Logged In: YES
user_id=963970
the latest (and greatest) monetdb/sql does handle this
situation correctly.
## Comment 13279
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 923761 at http://sourceforge.net/support/tracker.php?aid=923761
| ALTER TABLE ADD column type clears table | https://api.github.com/repos/MonetDB/MonetDB/issues/2374/comments | 0 | 2020-11-30T10:40:24Z | 2024-06-28T13:43:56Z | https://github.com/MonetDB/MonetDB/issues/2374 | 753,363,161 | 2,374 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-25 04:53:21 +0100
From: @MarcinZukowski
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-04 04:26:55 +0100
## Comment 13264
Date: 2004-03-25 16:53:21 +0100
From: @MarcinZukowski
We should at least:
- get rid of line statements
- have comments inside .h
## Comment 13265
Date: 2004-03-25 18:20:46 +0100
From: @drstmane
Logged In: YES
user_id=572415
- get rid of line statements:
What about using "Mx -l"? ;-)
18:19:35 manegold@draco:~/Monet/4.3.16/MonetDB $ Mx
Usage: Mx <flags> <file>.mx.
-t Produce LaTeX document (default).
-c Extract code.
-R <dir> Set target directory to <dir>).
-S <style> Set LaTeX documentstyle to 'style'.
-s Produce nroff -ms document.
-1 Single column (default) .
-2 Double column.
-H <n> Set hide level to 'n' (-H0 default).
-d Produce a draft document.
-x <extension> Extract <extension> labelled code.
-w Extract HTML code.
-D <id> Define macro 'id'.
-T <string> Define default hide text <string>.
-l No line and alike statements.
-n Non changed files won't be touched.
-+ Treat @c (C code) as @C (C++ code).
18:19:52 manegold@draco:~/Monet/4.3.16/MonetDB $
## Comment 13266
Date: 2004-03-25 18:59:43 +0100
From: @MarcinZukowski
Logged In: YES
user_id=607094
Indeed, -l would be just enough, but I don't how to add the switch only for
creation of this file.
## Comment 13267
Date: 2004-03-25 19:54:30 +0100
From: @mlkersten
Logged In: YES
user_id=490798
The line statements are needed to keep track of the
location when you are running with a debugger/profiler.
Droppping line could mean there is not enough documentation.
In the rare cases line was a stand in the way I took the c-file
and deleted the lines with a simple edit command
## Comment 13268
Date: 2004-03-25 20:05:02 +0100
From: @MarcinZukowski
Logged In: YES
user_id=607094
I only want line removed from .h files provided for external users, e.g.
Mapi.h.
I'm pretty sure one day there will be a need for separate distribution of
libMapi and libMapi-dev (e.g. to compile MonetDB PHP client ;)), and then it
would be handy.
It's not urgent though :)
## Comment 13269
Date: 2004-03-25 20:25:34 +0100
From: @drstmane
Logged In: YES
user_id=572415
Just no to forget this:
Both, Martin and Marcin are right (which is not too
supprisingly, since there names differ by only one character
... ;-):
Yes, line statments are needed for debugging as long as the
originaly .mx file is still available, i.e., during "make"
in the build directory.
After "make install" (or "make rpm") however, the .mx file
is gone, and then the line statements are actually "wrong"
and "hinder" convenient debugging.
Hence, the core of this feature request boils down to
- remove line statements from header files during "make
install"
- make sure that there is (some) documentation left in .h
files after they got extracted from a .mx file.
[Quite busy here for a low-priority feature request ... ;-)]
## Comment 13270
Date: 2004-03-25 21:10:24 +0100
From: @mlkersten
Logged In: YES
user_id=490798
The line statements are needed to keep track of the
location when you are running with a debugger/profiler.
Droppping line could mean there is not enough documentation.
In the rare cases line was a stand in the way I took the c-file
and deleted the lines with a simple edit command
## Comment 13271
Date: 2004-08-19 11:26:10 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
We should not confuse matters. The original request is to
remove line directives from Mapi.h. More generally,
removing them from all .h files should be no problem. Since
there is (or certainly: should) not be any executable code
in .h files, line directives don't do anything for the
debugger.
As to comments in Mapi.h, definitely! The whole Mapi.mx
documentation should be redone.
## Comment 13272
Date: 2004-09-15 16:27:53 +0200
From: @MarcinZukowski
Logged In: YES
user_id=607094
It seems that nobody really noticed adding -l switch to .h
creation rule.
Good :)
Still, currently the rule system makes everything recompile
even if I only touch .mx file containing .h.
The reason is, .h rule doesn't contain -n switch (don't
touch unchanged files) for Mx.
I suggest adding -n to .h and possibly also other rules (.c
.y and so on).
## Comment 13273
Date: 2009-12-04 16:26:54 +0100
From: @mlkersten
No action to be taken on this one anymore.
## Comment 13274
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 923204 at http://sourceforge.net/support/tracker.php?aid=923204
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Making Mapi.h more C friendly | https://api.github.com/repos/MonetDB/MonetDB/issues/2373/comments | 0 | 2020-11-30T10:40:21Z | 2024-06-28T13:43:55Z | https://github.com/MonetDB/MonetDB/issues/2373 | 753,363,120 | 2,373 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-23 08:47:58 +0100
From: @drstmane
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2016-04-11 11:43:15 +0200
## Comment 13262
Date: 2004-03-23 08:47:58 +0100
From: @drstmane
[ converted from http://sourceforge.net/support/tracker.php?aid=894704 originally submitted by
Martin Kersten - mlkersten ]
The selectivity estimates for tables with >100K
elements and a selectivity factor > 20% are severely
hampered by the table realloc strategy.
Solution. For all (larger) tables take the time to get
some basic statistics, e.g. the min/max or possibly a
coarse histogram. Use this information to obtain a
better estimate of the result size.
requires patches to gdk_batop/GDK_select*
[source: Cracker experiments]
## Comment 13263
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 921586 at http://sourceforge.net/support/tracker.php?aid=921586
| calculate selectivity estimates for large tables | https://api.github.com/repos/MonetDB/MonetDB/issues/2372/comments | 0 | 2020-11-30T10:40:18Z | 2024-06-28T13:43:55Z | https://github.com/MonetDB/MonetDB/issues/2372 | 753,363,088 | 2,372 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-18 04:33:40 +0100
From: @menzowindhouwer
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-01 12:05:40 +0100
## Comment 13259
Date: 2004-03-18 16:33:40 +0100
From: @menzowindhouwer
File types and their extensions are hardcoded in both
Mx and autogen. This prevents application developers to
add their own filetypes in a transparent way, i.e. they
have to hack Mx and codegen.py in a dozen places. These
tools should just contain a small (MonetDB related)
hardcoded set of file types, while additional types can
be added on the command line or by a local
configuration file, i.e. external of the code.
## Comment 13260
Date: 2009-12-01 12:05:40 +0100
From: @grobian
there are no intentions (I hope) to use Mx and autogen outside of MonetDB
## Comment 13261
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 918834 at http://sourceforge.net/support/tracker.php?aid=918834
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Move file types out of the codebase | https://api.github.com/repos/MonetDB/MonetDB/issues/2371/comments | 0 | 2020-11-30T10:40:15Z | 2024-06-28T13:43:54Z | https://github.com/MonetDB/MonetDB/issues/2371 | 753,363,039 | 2,371 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-18 04:29:54 +0100
From: @menzowindhouwer
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2007-11-10 11:34:44 +0100
## Comment 13255
Date: 2004-03-18 16:29:54 +0100
From: @menzowindhouwer
Modules/scripts developed independently can use the
same variable/function names, which leads to binding
conflicts or improper overloading when they are used
together. For example mapi and tcpip both (used to)
have a listen command, saxophone and the xml module
both had a prefix command/variable. Currently there is
no way to resolve these clashes, or even give the
interpreter a hint, but explicitly renaming. By adding
support for the use of implicit (module) or explicit
namespaces these situations can be resolved more elegant.
## Comment 13256
Date: 2007-11-10 11:34:44 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
MIL is deprecated and this feature will never be supported.
## Comment 13257
Date: 2007-11-10 11:50:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
Just for the records:
(AFAIK), MAL enforces namespaces for functions; (AFAIK) not for variables, though.
## Comment 13258
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 918831 at http://sourceforge.net/support/tracker.php?aid=918831
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| MIL namespaces | https://api.github.com/repos/MonetDB/MonetDB/issues/2370/comments | 0 | 2020-11-30T10:40:12Z | 2024-06-28T13:43:54Z | https://github.com/MonetDB/MonetDB/issues/2370 | 753,362,993 | 2,370 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-18 04:23:04 +0100
From: @menzowindhouwer
To: @mlkersten
Version: -- development
Last updated: 2004-04-18 11:48:34 +0200
## Comment 13252
Date: 2004-03-18 16:23:04 +0100
From: @menzowindhouwer
The current help command gives only information if you
already know the exact name of the command. If not you
have to remember/guess in which module this command
resides, so you can get a list of command names using
the sigs command. Adding an apropos command would
improve this situation. This command would look for
matching substrings, i.e. like, in both command names
and help messages for loaded modules.
## Comment 13253
Date: 2004-04-18 23:48:34 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Mknife now comes with built-in functions to search
the module signatures and names based on a single
keyword.
see MILinfoset:
Function dicationary
Function help
Function search
## Comment 13254
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 918823 at http://sourceforge.net/support/tracker.php?aid=918823
| additional help facilities | https://api.github.com/repos/MonetDB/MonetDB/issues/2369/comments | 0 | 2020-11-30T10:40:08Z | 2024-06-28T13:43:53Z | https://github.com/MonetDB/MonetDB/issues/2369 | 753,362,953 | 2,369 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-16 08:45:57 +0100
From: @njnes
To: @mlkersten
Version: -- development
Last updated: 2006-11-22 10:28:27 +0100
## Comment 13249
Date: 2004-03-16 20:45:57 +0100
From: @njnes
Current mapi implementation keeps the query around just
to give the
user an informative return in case of errors. However
keeping queries
around is sometimes close to impossible ( MapiClient
file mode and/or
multiline queries). The server knows very well which
query failed and
should therefore return more information on errrors.
## Comment 13250
Date: 2006-11-22 22:28:27 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: NO
I think all (non-deprecated) languages properly return some context with a pointer in their error messages if appropriate.
## Comment 13251
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 917549 at http://sourceforge.net/support/tracker.php?aid=917549
| improve client/server error handling | https://api.github.com/repos/MonetDB/MonetDB/issues/2368/comments | 0 | 2020-11-30T10:40:05Z | 2024-06-28T13:43:52Z | https://github.com/MonetDB/MonetDB/issues/2368 | 753,362,920 | 2,368 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-11 02:23:37 +0100
From: @mlkersten
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2009-12-04 04:11:39 +0100
## Comment 13245
Date: 2004-03-11 14:23:37 +0100
From: @mlkersten
String heaps are not cleaned. It means that when you
add 100 strings and drop 99, you still have the 99 string
values in the database for 'future' use.
Possible solution is a database cleanup upon Mserver start
## Comment 13246
Date: 2009-12-01 12:06:41 +0100
From: @grobian
sounds like a vacuum operation is necessary
## Comment 13247
Date: 2009-12-04 16:11:39 +0100
From: @mlkersten
Not going to be implemented in the current code base.
## Comment 13248
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 914118 at http://sourceforge.net/support/tracker.php?aid=914118
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| GDK: delete (cleanup) strings from bats | https://api.github.com/repos/MonetDB/MonetDB/issues/2367/comments | 0 | 2020-11-30T10:40:02Z | 2024-06-28T13:43:52Z | https://github.com/MonetDB/MonetDB/issues/2367 | 753,362,896 | 2,367 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-05 09:53:10 +0100
From: @grobian
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2006-11-22 10:27:00 +0100
## Comment 13240
Date: 2004-03-05 09:53:10 +0100
From: @grobian
In order to greatly reduce both the parsing at the
client and the network transmission costs, a more
efficient way of beaming over the results in a MAPI
session can be thought of.
A proposal for this:
1. An (SQL) query is made
2. Mserver answers with the usual header, but adds a
new header indicating the number of bits each column takes:
- for numeric types this is just the number of bits,
ie. a short is 2 bytes thus 16 bits
- for a boolean this is 1
- for char (non variable) this is the column width as
stored in the database (converted to bits of course),
ie char(10) would be 80 bits UTF-8
- for varchar the length prefix in bits is given, for
example a short, thus 16 bits (can be the as minimal as
the data set allows)
3. Mserver gives back the results in the following
manor per row:
- a bitmask indicating whether a column is null or
not (for each column in the right order is set if it's
null or not)
- for all columns that are NOT null, their values are
put with the sizes as given in the header. For varchars
the length is written first, then the string itself.
- if there are more rows to come, PROMPT2 is written,
else PROMPT1 (indicating the server is done)
- repeat for the next row
A simple example could be:
<< SSELECT * FROM a;
>> var1, var2, var3, var4 name
>> char, varchar, int, boolean type
>> 80, 8, 32, 1 size
>> 2, , , tuplecount
>> 0, , , id
(binary data represented as string)
>> 0010(10 bytes char)00000011(3 bytes char)1(PROMPT2)
>> 0000(10 bytes char)00000111(7 bytes
char)000000000000000000000000000000000(PROMPT1)
(which would represent something like:
>> [ "1234567890", "het", nil, true ]
>> [ "1234567 ", "ondanks", 726, false ]
>> \1sql\2
)
## Comment 13241
Date: 2004-03-05 10:08:14 +0100
From: @grobian
Logged In: YES
user_id=963970
One added comment:
It might be useful to consider that for instance a language
like Java can only read and write bytes. It might be
necessary to put the bitmask in chunks of byte size just
filling up with bits till the byte is full. (Java can shift
bits) This would also allow the headers to specify bytes
instead of bits. The boolean type would have to be increased
to byte size though.
## Comment 13242
Date: 2004-09-09 11:20:00 +0200
From: @grobian
Logged In: YES
user_id=963970
Due to the latest protocol changes, overhead has increased.
A more efficient batch protocol is needed in order to
submit a batch for execution.
- the batch should be executed as a whole (either completely
or not)
- the affected rows should be returned in one result-set,
instead of a complete resultset for each executed query.
## Comment 13243
Date: 2006-11-22 22:27:00 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
The mapi protocol has been improved considerably.
## Comment 13244
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 910321 at http://sourceforge.net/support/tracker.php?aid=910321
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| More efficient result transfer protocol | https://api.github.com/repos/MonetDB/MonetDB/issues/2366/comments | 0 | 2020-11-30T10:39:59Z | 2024-06-28T13:43:51Z | https://github.com/MonetDB/MonetDB/issues/2366 | 753,362,851 | 2,366 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-01-08 11:47:21 +0100
From: @njnes
To: @njnes
Version: -- development
Last updated: 2004-03-04 09:21:41 +0100
## Comment 13235
Date: 2004-01-08 11:47:21 +0100
From: @njnes
The current mapi/mil and mapi/sql communication
protocol does
way to many copies and parse phases. This can be reduced by
changing the mapi protocol. An additional prompt needs to
be introduced which indicates to the mapi-client that more
input is needed to finish the current queriy. This means
changing mapi's api (mapi_query needs to return the
'need more
input' information), sql/mil scanners should use the
streams directly
and parse/scan from strings.
Also a binary interface should be added.
## Comment 13236
Date: 2004-01-30 13:54:39 +0100
From: @grobian
Logged In: YES
user_id=963970
MAPI requests:
- Some prompt or marker AFTER the header to indicate the
header is over, and results will be next
- Give prompt on !ERROR, currently it is unknown what the
state of the server is after it gave !ERROR, also hard to
synchronize with client on it
- Usage of Xexport with multiple threads concurrently over
multiple connections causes segmentation fault in Mserver
- Xexport is SLOW. Server should cache results as they are
sent in order to be quicker when returning the next block of
results.
- A quick Xexport would make it possible to solve my first
point by simply giving the prompt after the header and let
the client ask his first block. (Is possible nowadays, but
due to slowness not cool, due to crashes not cool too)
- Resultless queries like UPDATE, DELETE and INSERT should
return a header containing at least the number of rows affected
## Comment 13237
Date: 2004-01-30 14:03:10 +0100
From: @grobian
Logged In: YES
user_id=963970
SQL query sending:
In the current MAPI implementation the server keeps on
waiting for more input if a string with non-balanced quotes
is sent. The client however has sent its data and waits for
the header. To get around this the client has to check for
unbalanced quotes itself, while this should be done by the
server since it already has to check/parse the query.
There should be a way to tell the server that all data is
sent (ie termination char/sequence) after which the server
should return an error if the query could not be parsed.
## Comment 13238
Date: 2004-03-04 21:21:41 +0100
From: @njnes
Logged In: YES
user_id=43556
With the new mapi_query_{*} functions and PROMPT2 this new
feature
is fully implemented.
## Comment 13239
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 872975 at http://sourceforge.net/support/tracker.php?aid=872975
| improved communication protocol | https://api.github.com/repos/MonetDB/MonetDB/issues/2365/comments | 0 | 2020-11-30T10:39:55Z | 2024-06-28T13:43:50Z | https://github.com/MonetDB/MonetDB/issues/2365 | 753,362,802 | 2,365 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-01-07 10:16:36 +0100
From: @mlkersten
To: SQL devs <<bugs-sql>>
Version: -- development
Last updated: 2009-12-04 04:13:45 +0100
## Comment 13231
Date: 2004-01-07 10:16:36 +0100
From: @mlkersten
The Blob module has been developed 8 years ago as a
first attempt to exercise the MEL facilities. The way
it was implemented reflected the then to time approach
to manage all data within the confines of a DBMS. For
Blobs this meant effectively mimicking a file-service.
With the ubiquitious availability of HTTP services it
makes sense to drastically redesign the blob interface
along the following lines.
Assumption 1. A URL module is already available to
manage strings
satisfying the URL conventions.
Assumption 2. Blobs are 'ordinary data items' brought
under the control of the DBMS
Assumption 3. Blobs find their origin in the local file
system or in remote web-space.
The Blobs can now be managed by the DBMS as follows.
Next to the dbfarm directory for retaining the
persistent data, a directory is created in the space
managed by e.g. an Apache server. The location of which
is handed to the Blob module upon initialization. Let's
call it .../MonetDB/BlobStore
Blobs are inserted into the DBMS by providing an url
handle, e.g.
the following should work:
b:= new(int, blob);
b.insert(1,"http://www.cwi.nl/~mk/index.html");
b.insert(2,"file:///tmp/x");
Upon transaction commit (or anytime earlier as needed)
the blob objects are obtained and assigned a location
in the BlobStore using an internally generated name.
The BlobStore could be organized as a cache, e.g.
multiple directories to spread the load.
The Blob interface maintains the mapping from the
original handle to the location in its own store.
Furthermore, a reference count should be maintained to
guarantee accessibility over multiple assigments.
The operations defined over an Blob items become:
getOrigin - returns the source handle
clrOrigin - forget the source handle
getHandle- return url to enable access
The mapping between handle and physical location can be
changed by the DBMS at will. Their is no guarantee that
the handle given to a user is likely to last forever.
This Blob mechanism can be readible extended to handle
images, video, audio, etc.. They merely require some
additional test on the content type and probably need
more space.
The expected implementation effort is considered 1-2
days, provided we can easily include the HTTP access
protocol software inside the Mserver to retrieve the
Blobs in the first place from the application environment.
## Comment 13232
Date: 2004-09-09 11:22:54 +0200
From: @grobian
Logged In: YES
user_id=963970
the Mapi protocol should handle BLOBs and TEXTs (which
could/should be very large) as special urls, which can be
retrieved by the mapi client and returned to the user as
blob. This meets requirements made by for instance JDBC.
## Comment 13233
Date: 2009-12-04 16:13:45 +0100
From: @mlkersten
This is not going to be addressed in the near future.
Functional enhancements for specific applications can
include user defined functions for it.
## Comment 13234
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 872236 at http://sourceforge.net/support/tracker.php?aid=872236
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-sql@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| SQL: better BLOB protocol support | https://api.github.com/repos/MonetDB/MonetDB/issues/2364/comments | 0 | 2020-11-30T10:39:52Z | 2024-06-28T13:43:50Z | https://github.com/MonetDB/MonetDB/issues/2364 | 753,362,768 | 2,364 |
[
"MonetDB",
"MonetDB"
] | Date: 2004-03-19 03:09:57 +0100
From: @mlkersten
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2004-03-23 09:52:51 +0100
## Comment 13228
Date: 2004-03-19 15:09:57 +0100
From: @mlkersten
>create table t(i int, s varchar);
>insert into t(s,i) values('hello',1);
>insert into t values('world',3);
>insert into t values(2.3,'e');
>select * from t;
i, s name
int, varchar type
2, tuplecount
2, id
[ 1, "hello" ]
[ nil, "3" ]
[ 2, "e" ]
>
The cast of 'world' is questionable
and cast of 3 -> '3' might be SQL requirement?
but a float to an int???
If I am correct the type of the values should
comply with the attributes.
## Comment 13229
Date: 2004-03-23 09:52:51 +0100
From: @njnes
Logged In: YES
user_id=43556
introduced special cast2{sht,int,lng,flt,dbl} needed to
check the result of
string to number convertions.
The number convertions in this example are both correct (ie
3->'3' and 2.3 ->2)
## Comment 13230
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 919531 at http://sourceforge.net/support/tracker.php?aid=919531
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Illegal casts not caught | https://api.github.com/repos/MonetDB/MonetDB/issues/2363/comments | 0 | 2020-11-30T10:39:48Z | 2024-06-27T11:44:34Z | https://github.com/MonetDB/MonetDB/issues/2363 | 753,362,725 | 2,363 |
[
"MonetDB",
"MonetDB"
] | Date: 2003-11-25 03:16:39 +0100
From: @arjan
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2003-11-25 03:39:56 +0100
## Comment 13225
Date: 2003-11-25 15:16:39 +0100
From: @arjan
The current CVS version does not compile because of 2
forgotten includes in color.mx and array.mx
(src/modules/contrib). The patch below fixes both problems.
===================================================================
RCS file:
/cvsroot/monetdb/MonetDB/src/modules/contrib/array.mx,v
retrieving revision 1.1
diff -r1.1 array.mx
22a23
> include <algebra.h>
Index: color.mx
===================================================================
RCS file:
/cvsroot/monetdb/MonetDB/src/modules/contrib/color.mx,v
retrieving revision 1.1
diff -r1.1 color.mx
72a73,74
> include <math.h>
>
cvs server: Diffing Tests
## Comment 13226
Date: 2003-11-25 15:39:56 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Thanks for the report.
This has already been fixed in CVS.
## Comment 13227
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 848974 at http://sourceforge.net/support/tracker.php?aid=848974
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Missing #includes | https://api.github.com/repos/MonetDB/MonetDB/issues/2362/comments | 0 | 2020-11-30T10:39:45Z | 2024-06-27T11:44:33Z | https://github.com/MonetDB/MonetDB/issues/2362 | 753,362,682 | 2,362 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-05-03 02:48:43 +0200
From: zeuner <<zeuner>>
To: MonetDB5 devs <<bugs-monetdb5>>
Version: -- development
CC: @mlkersten
Last updated: 2010-05-04 21:52:21 +0200
## Comment 13223
Date: 2010-05-03 02:48:43 +0200
From: zeuner <<zeuner>>
Using SQL/MonetDB5 compiled from a 20100501 checkout, I noticed that the change introduced with [1] led to sql.bind and sql.append MAL instructions being intermixed by opt_reorder. In the attached SQL test, this leads to a primary key violation because the inserted data is checked for collisions with the table after the insertion has taken place.
The attached fix prevents this from happening by having update instructions "depend" on all the sql.bind instructions that were invoked earlier.
[1] http://monetdb.cvs.sourceforge.net/viewvc/monetdb/MonetDB5/src/optimizer/opt_support.mx?r1=1.115&r2=1.116
## Comment 13224
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2995671 at http://sourceforge.net/support/tracker.php?aid=2995671
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 13992
Date: 2010-05-04 21:52:01 +0200
From: @mlkersten
Thank you for the patch and accompanying test! The test has been added to the repository.
| opt_reorder intermixing sql.bind with sql update instruction | https://api.github.com/repos/MonetDB/MonetDB/issues/2361/comments | 0 | 2020-11-30T10:39:42Z | 2024-06-27T11:44:32Z | https://github.com/MonetDB/MonetDB/issues/2361 | 753,362,645 | 2,361 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-30 05:01:49 +0200
From: zeuner <<zeuner>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2010-05-01 03:33:38 +0200
## Comment 13219
Date: 2010-04-30 05:01:49 +0200
From: zeuner <<zeuner>>
When executing the attached SQL code, the last query gets compiled to a MAL sequence containing a line like "_13 := mat.slice(0:wrd,0:wrd,_45,_46);", where _45 and _46 are both empty. This causes mat.slice to throw a RUNTIME_OBJECT_MISSING exception in line 292 of src/modules/mal/mat.mx, because i1 is not initialized to a valid value, and doesn't get set if the argument BATs are all empty.
A fix is attached, which will initialize i1 to 3, which leads to correct results when the argument BATs are empty, and gets reset if there are non-empty BAT arguments.
A more efficient fix would avoid the second loop by initializing i1 to p->argc, but this would require to give up the i1 < i2 assertion in line 276, or at least replacing it by i1 != i2. The maintainers should probably decide if the assertion is still important.
## Comment 13220
Date: 2010-04-30 11:13:18 +0200
From: @drstmane
Thanks for reporting!
I'll take care of it.
## Comment 13221
Date: 2010-05-01 15:33:38 +0200
From: @drstmane
fixed in CVS.
test added in
sql/src/test/BugTracker-2010/Tests/mat.slice_on_empty_BATs.SF-2994521.*
## Comment 13222
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2994521 at http://sourceforge.net/support/tracker.php?aid=2994521
| mat.slice unable to cope with only empty BAT arguments | https://api.github.com/repos/MonetDB/MonetDB/issues/2360/comments | 0 | 2020-11-30T10:39:38Z | 2024-06-27T11:44:31Z | https://github.com/MonetDB/MonetDB/issues/2360 | 753,362,604 | 2,360 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-29 03:08:38 +0200
From: zeuner <<zeuner>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2011-09-07 10:47:04 +0200
## Comment 13217
Date: 2010-04-29 15:08:38 +0200
From: zeuner <<zeuner>>
The described issue could be observed on a x86_64 dual core machine running CentOS 5.4. A recent checkout of MonetDB CVS Head was used, and mserver5 started with the "--set gdk_nr_threads=1" command line option (without that, at least the attached SQL code won't trigger the issue).
The optimizer pipeline used was "inline,remap,evaluate,costModel,coercions,emptySet,aliases,mitosis,mergetable,deadcode,commonTerms,joinPath,deadcode,reduce,dataflow,history,multiplex,garbageCollector" ("reorder" was removed to make the MAL output more easy to understand).
When MonetDB5 is compiled with "--enable-strict=yes --enable-assert=yes --enable-debug=no --enable-optimize=yes", the emitted MAL code appears to be correct. With "--enable-strict=yes --enable-assert=yes --enable-debug=yes --enable-optimize=no", the MAL code differs. Moreover, it is incorrect as the MAT expansion leads to mismatching combinations being instantiated, ultimately leading to BATS with different sizes being passed to batcalc.< when the iBAT of the "value" table is non-empty.
## Comment 13218
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2994159 at http://sourceforge.net/support/tracker.php?aid=2994159
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 16185
Date: 2011-09-07 10:04:31 +0200
From: @sjoerdmullender
Created attachment 67
SQL code showing the issue
Re-added attachment that was lost in the move from Sourceforge to bugzilla.
> Attached file: [mitosis-debug-fail.sql](https://github.com/MonetDB/monetdb-issues-attachments/blob/master/issue_2359_mitosis-debug-fail.sql_67) (application/octet-stream, 807 bytes)
> Description: SQL code showing the issue
## Comment 16186
Date: 2011-09-07 10:05:31 +0200
From: @sjoerdmullender
Created attachment 68
expected MAL output as created by the optimized build
Re-added attachment that was lost in the move from Sourceforge to bugzilla.
> Attached file: [nodebug.explain](https://github.com/MonetDB/monetdb-issues-attachments/blob/master/issue_2359_nodebug.explain_68) (application/octet-stream, 4902 bytes)
> Description: expected MAL output as created by the optimized build
## Comment 16187
Date: 2011-09-07 10:05:59 +0200
From: @sjoerdmullender
Created attachment 69
incorrect MAL output as created by the debug build
Re-added attachment that was lost in the move from Sourceforge to bugzilla.
> Attached file: [debug.explain](https://github.com/MonetDB/monetdb-issues-attachments/blob/master/issue_2359_debug.explain_69) (application/octet-stream, 6139 bytes)
> Description: incorrect MAL output as created by the debug build
## Comment 16188
Date: 2011-09-07 10:47:04 +0200
From: @sjoerdmullender
It looks like the bug has been fixed. In the Aug2011 branch I tried running the query both with a debug build and an optimized build and the outputs of the two runs were identical. This output differs significantly from the attached outputs.
(If I remove the option "--set gdk_nr_threads=1" the output differs slightly: a dataflow barrier end exit are added to the plan.)
| mitosis/mergetable creates incorrect code in debug mode | https://api.github.com/repos/MonetDB/MonetDB/issues/2359/comments | 0 | 2020-11-30T10:39:33Z | 2024-06-27T11:44:30Z | https://github.com/MonetDB/MonetDB/issues/2359 | 753,362,551 | 2,359 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-23 05:55:55 +0200
From: Matthieu Guamis <<mguamis>>
To: SQL devs <<bugs-sql>>
Version: 2.40.1 (Oct2010) [obsolete]
CC: matthieu.guamis, @njnes, @drstmane
Last updated: 2011-03-28 17:31:28 +0200
## Comment 13215
Date: 2010-04-23 17:55:55 +0200
From: Matthieu Guamis <<mguamis>>
COPY INTO TABLE with "too long' field fails with message
SQLException:sql:value 'abcd' from line 1 field 1 not inserted, expecting type str
SQLException:importTable:failed to import table
This can be reproduce using statements:
sql>create table t (c varchar(3));
sql>COPY 1 RECORDS INTO t FROM STDIN USING DELIMITERS '|','
more>','"' NULL AS '';"abcd"
more>
more>
SQLException:sql:value 'abcd' from line 1 field 1 not inserted, expecting type str
SQLException:importTable:failed to import table
Error message is much more relevant using insert statement:
sql>insert into t values ('abcd');
SQLException:str_cast:value too long for type (var)char(3)
Best regards
$ mserver5 --version
MonetDB server v5.18.3 (64-bit), based on kernel v1.36.3 (64-bit oids)
Copyright (c) 1993-July 2008 CWI
Copyright (c) August 2008-2010 MonetDB B.V., all rights reserved
Visit http://monetdb.cwi.nl/ for further information
Found 2.0GiB available memory, 2 available cpu cores
Configured for prefix: /usr
Libraries:
libpcre: 7.8 2008-09-05 (compiled with 7.8)
openssl: OpenSSL 0.9.8g 19 Oct 2007 (compiled with OpenSSL 0.9.8g 19 Oct 2007)
libxml2: 2.7.5 (compiled with 2.7.5)
Compiled by: root@ubuntu-karmic-64 (x86_64-pc-linux-gnu)
Compilation: gcc -O2 -Wall -O2 -O6 -fomit-frame-pointer -finline-functions -falign-loops=4 -falign-jumps=4 -falign-functions=4 -fexpensive-optimizations -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -ftree-vectorize
Linking : ld -IPA -m elf_x86_64 -Wl,-Bsymbolic-functions
## Comment 13216
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2991447 at http://sourceforge.net/support/tracker.php?aid=2991447
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 14102
Date: 2010-06-14 15:37:06 +0200
From: @grobian
Changeset [a3590f983661](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=a3590f983661) made by Fabian Groffen <fabian@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=a3590f983661](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=a3590f983661)
Changeset description:
Add script from bug #2358
## Comment 14103
Date: 2010-06-14 15:37:40 +0200
From: @grobian
bug still applies to release candidate (Jun2010)
## Comment 14104
Date: 2010-06-14 15:41:31 +0200
From: @grobian
Changeset [138abbed9155](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=138abbed9155) made by Fabian Groffen <fabian@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=138abbed9155](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=138abbed9155)
Changeset description:
Reapplying commit [a3590f983661](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=a3590f983661)
this time only the test for bug #2358
## Comment 14733
Date: 2010-08-19 16:43:46 +0200
From: @sjoerdmullender
Is this still a problem? The test seems to be working.
We do need a changelog entry...
## Comment 14742
Date: 2010-08-20 13:16:36 +0200
From: @sjoerdmullender
Changeset [7c07e1c16195](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=7c07e1c16195) made by Sjoerd Mullender <sjoerd@acm.org> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=7c07e1c16195](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=7c07e1c16195)
Changeset description:
Changed error output to more closely match the intention of the bug report.
This shows that the bug (bug #2358) is still there.
## Comment 14743
Date: 2010-08-20 13:17:15 +0200
From: @sjoerdmullender
Bug is still in Jun2010-SP1.
## Comment 14844
Date: 2010-08-30 09:23:21 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
## Comment 15063
Date: 2010-10-18 19:41:28 +0200
From: @drstmane
Appears to work in Oct2010, since the error message has been approved from
!SQLException:importTable:value 'abcd' from line 1 field 1 not inserted, value too long for type (var)char(3)
to
!SQLException:importTable:value 'abcd' from line 1 field 1 not inserted, expecting type varchar(3)
in changeset http://dev.monetdb.org/hg/MonetDB/rev/bd70cf18b649
Is this as intended?
## Comment 15624
Date: 2011-03-28 17:31:28 +0200
From: @sjoerdmullender
The Mar2011 version has been released.
| Unrelevant error message when copying bad length field | https://api.github.com/repos/MonetDB/MonetDB/issues/2358/comments | 0 | 2020-11-30T10:39:30Z | 2024-06-27T11:44:29Z | https://github.com/MonetDB/MonetDB/issues/2358 | 753,362,509 | 2,358 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-23 04:26:20 +0200
From: Matthieu Guamis <<mguamis>>
To: @njnes
Version: -- development
Last updated: 2010-04-25 01:52:00 +0200
## Comment 13211
Date: 2010-04-23 16:26:20 +0200
From: Matthieu Guamis <<mguamis>>
COPY INTO TABLE with embedded record separator in quoted fields fails with message
SQLException:sql:End of string (") missing in "
" at line 2 field 0
SQLException:importTable:failed to import table
This can be reproduce using statements:
sql>create table t (c varchar(3));
sql>COPY 1 RECORDS INTO t FROM STDIN USING DELIMITERS '|','
more>','"' NULL AS '';"a
more>c"
more>
SQLException:sql:End of string (") missing in "
" at line 2 field 0
SQLException:importTable:failed to import table
It's the same with other records separators.
It works with embedded *field* seperator.
Best regards
## Comment 13212
Date: 2010-04-23 16:27:01 +0200
From: Matthieu Guamis <<mguamis>>
$ mserver5 --version
MonetDB server v5.18.3 (64-bit), based on kernel v1.36.3 (64-bit oids)
Copyright (c) 1993-July 2008 CWI
Copyright (c) August 2008-2010 MonetDB B.V., all rights reserved
Visit http://monetdb.cwi.nl/ for further information
Found 2.0GiB available memory, 2 available cpu cores
Configured for prefix: /usr
Libraries:
libpcre: 7.8 2008-09-05 (compiled with 7.8)
openssl: OpenSSL 0.9.8g 19 Oct 2007 (compiled with OpenSSL 0.9.8g 19 Oct 2007)
libxml2: 2.7.5 (compiled with 2.7.5)
Compiled by: root@ubuntu-karmic-64 (x86_64-pc-linux-gnu)
Compilation: gcc -O2 -Wall -O2 -O6 -fomit-frame-pointer -finline-functions -falign-loops=4 -falign-jumps=4 -falign-functions=4 -fexpensive-optimizations -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -ftree-vectorize
Linking : ld -IPA -m elf_x86_64 -Wl,-Bsymbolic-functions
## Comment 13213
Date: 2010-04-25 13:52:00 +0200
From: @njnes
fixed in stable. Added test to sql/src/test/BugTracker-2010/embedded...
Problem was indeed in quote handling
## Comment 13214
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2991400 at http://sourceforge.net/support/tracker.php?aid=2991400
| Embedded record separator in quoted fields | https://api.github.com/repos/MonetDB/MonetDB/issues/2357/comments | 0 | 2020-11-30T10:39:26Z | 2024-06-27T11:44:28Z | https://github.com/MonetDB/MonetDB/issues/2357 | 753,362,471 | 2,357 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-23 03:02:33 +0200
From: @gijzelaerr
To: clients devs <<bugs-clients>>
Version: -- development
CC: @drstmane
Last updated: 2010-11-15 09:39:06 +0100
## Comment 13209
Date: 2010-04-23 15:02:33 +0200
From: @gijzelaerr
The monetdb python package for ubuntu is always installed in /usr/lib/python2.5/site-packages. This will be fixed when we switch to the new repository, the python package will get a separated hg repo and setuptools will be used for package generation. Ubuntu has CDBS support for pycentral to manage different python verisons.
## Comment 13210
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2991363 at http://sourceforge.net/support/tracker.php?aid=2991363
## Comment 14025
Date: 2010-05-13 15:00:14 +0200
From: @gijzelaerr
I moved the python db API out of the client package. This way it can be packaged separately.
Things that still need to be done:
* The nightly tests need to be modified to run python/test/runtests.py. The
old python tests are depricated.
* Ubuntu packaging needs to be modified to use the new packaging style.
Debian/ubuntu files are already quite finished (description and version
number need to be changed)
* I'm working on a specs file at the moment
* Tarballs can be created by running 'python setup.py sdist' from the
python folder, see:
http://docs.python.org/distutils/sourcedist.html
## Comment 14026
Date: 2010-05-13 15:23:59 +0200
From: @gijzelaerr
actually, RPM's can be created with:
python setup.py bdist_rpm
I don't know if we need more features, but the package seems to work on fedora 12
## Comment 14027
Date: 2010-05-13 15:28:36 +0200
From: @gijzelaerr
and a windows installer (that we don't have at the moment) can be made with:
* python setup.py bdist_wininst
## Comment 15062
Date: 2010-10-18 19:32:27 +0200
From: @drstmane
The new python package(s) will be distributed as of the upcoming Oct2010 release.
| python-monetdb ubuntu package installed in the wrong location | https://api.github.com/repos/MonetDB/MonetDB/issues/2356/comments | 0 | 2020-11-30T10:39:23Z | 2024-06-27T11:44:27Z | https://github.com/MonetDB/MonetDB/issues/2356 | 753,362,425 | 2,356 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-19 11:50:50 +0200
From: Matthew W. Jones <<matburt>>
To: Gijs Molenaar <<gijs>>
Version: -- development
Last updated: 2010-05-29 10:43:56 +0200
## Comment 13205
Date: 2010-04-19 23:50:50 +0200
From: Matthew W. Jones <<matburt>>
When connecting to a remote Monet instance with the python client and setting the cursor's arraysize to something significant (100, as an example) then I will occasionally get the following error:
Traceback (most recent call last):
  File "afile.py", line 9, in <module>
   cur.execute("SELECT max(val), otherval"
  File "build/bdist.linux-x86_64/egg/monetdb/sql/cursors.py", line 203, in execute
  File "build/bdist.linux-x86_64/egg/monetdb/sql/cursors.py", line 445, in __store_result
  File "build/bdist.linux-x86_64/egg/monetdb/sql/cursors.py", line 529, in __parse_tuple
  File "build/bdist.linux-x86_64/egg/monetdb/sql/cursors.py", line 569, in __exception_handler
monetdb.monetdb_exceptions.InterfaceError: length of row doesn't match header
This doesn't happen when the client is used locally on the machine running the database.
It prevents the .arraysize cursor property from being used remotely. Â It's default value is 1, which causes the server to take a ridiculously long time to send back large data sets.
## Comment 13206
Date: 2010-04-23 14:24:23 +0200
From: @gijzelaerr
I can't reproduce the problem at the moment. Can you run your script in the failing situation again, but with this in the top:
import logging
logging.basicConfig(level=logging.DEBUG)
logger = logging.getLogger('monetdb')
And report the debug output (only the last lines related to the exception) into this bug report? This would help me a lot.
## Comment 13207
Date: 2010-04-26 22:18:45 +0200
From: Matthew W. Jones <<matburt>>
Here you go, it almost looks as though that either MonetDB isn't sending a complete result or the python module isn't processing it correctly. This is done with the latest client and the latest release of the server.
$ python
Python 2.5.2 (r252:60911, Jul 22 2009, 15:33:10)
[GCC 4.2.4 (Ubuntu 4.2.4-1ubuntu3)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import monetdb.sql as monetdb
>>> import logging
>>> logging.basicConfig(level=logging.DEBUG)
>>> logger = logging.getLogger('monetdb')
>>> dbc = monetdb.connect(hostname = 'ahost', username='auser', password='apass', database='adatabase', autocommit=True)
DEBUG:monetdb:II: reading 32 bytes
DEBUG:monetdb:RX: void:merovingian:8:md5,plain:LIT
DEBUG:monetdb:II: sending 59 bytes, last: 1
DEBUG:monetdb:TX: BIG:auser:{plain}apassvoid:sql:adatabase:
DEBUG:monetdb:II: reading 54 bytes
DEBUG:monetdb:RX: ^mapi:merovingian://proxy?database=adatabase
DEBUG:monetdb:II: merovingian proxy, restarting authenticatiton
DEBUG:monetdb:II: reading 59 bytes
DEBUG:monetdb:RX: rJd5Ja2kjcl:mserver:9:RIPEMD160,SHA256,SHA1,MD5:LIT:SHA512:
DEBUG:monetdb:II: sending 88 bytes, last: 1
DEBUG:monetdb:TX: BIG:auser:{SHA1}6c856d7155585d6c17e492052234ecdae3d82423:sql:adatabase:
DEBUG:monetdb:II: reading 0 bytes
DEBUG:monetdb:RX:
DEBUG:monetdb:II: executing command Xauto_commit 1
DEBUG:monetdb:II: sending 14 bytes, last: 1
DEBUG:monetdb:TX: Xauto_commit 1
DEBUG:monetdb:II: reading 0 bytes
DEBUG:monetdb:RX:
>>> cur = dbc.cursor()
>>> cur.arraysize = 10000000
>>> cur.execute("select * from abigtable limit 50")
DEBUG:monetdb:II: executing command Xreply_size 10000000
DEBUG:monetdb:II: sending 20 bytes, last: 1
DEBUG:monetdb:TX: Xreply_size 10000000
DEBUG:monetdb:II: reading 0 bytes
DEBUG:monetdb:RX:
DEBUG:monetdb:II: executing command sselect * from abigtable limit 50;
DEBUG:monetdb:II: sending 46 bytes, last: 1
DEBUG:monetdb:TX: sselect * from abigtable limit 50;
DEBUG:monetdb:II: reading 4582 bytes
DEBUG:monetdb:RX: &1 0 50 15 50
% aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable, aschema.abigtable table_name
% id, entity, entityid, entitytype, parententity, parententityid, parententitytype, col1, col2, col3, col4, col4, col5, col6, col7 name
% bigint, varchar, int, int, varchar, int, int, int, int, int, int, int, int, int, int type
% 2, 27, 2, 1, 18, 2, 2, 6, 5, 4, 5, 1, 1, 4, 1 length
[ 1, "string1", 1, 1, "", -1, -1, 278647, 56242, 4933, 60384, 3, 2, 2010, NULL ]
[ 2, "string2", 2, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 3, "string3", 3, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 4, "string4", 4, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 5, "string5", 5, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 6, "string6", 6, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 7, "string7", 7, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 8, "string8", 8, 1, "", -1, -1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 9, "string9", 9, 1, "", -1, -1, 657, 6, 2, 233, 3, 2, 2010, NULL ]
[ 10, "string10", 10, 1, "", -1, -1, 4414, 198, 38, 2526, 3, 2, 2010, NULL ]
[ 11, "string11", 11, 1, "", -1, -1, 965, 39, 9, 689, 3, 2, 2010, NULL ]
[ 12, "string12", 1, 2, "anotherstring", 1, 1, 10090, 4089, 206, 4848, 3, 2, 2010, NULL ]
[ 13, "string13", 2, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 14, "string14", 3, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 15, "string15", 4, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 16, "astring", 5, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 17, "astring", 6, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 18, "astring", 7, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 19, "astring", 8, 2, "anotherstring", 1, 1, 3606, 664, 50, 1497, 3, 2, 2010, NULL ]
[ 20, "astring", 9, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 21, "astring", 10, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 22, "astring", 11, 2, "anotherstring", 1, 1, 16004, 2832, 195, 6715, 3, 2, 2010, NULL ]
[ 23, "astring", 12, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 24, "astring", 13, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 25, "astring", 14, 2, "anotherstring", 1, 1, 248485, 48641, 4477, 47382, 3, 2, 2010, NULL ]
[ 26, "astring", 15, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 27, "astring", 16, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 28, "astring", 17, 2, "anotherstring", 1, 1, 0, 0, 0, 0, 3, 2, 2010, NULL ]
[ 29, "astring", 18, 2, "anotherstring", 1, 1, 0, 0, 0, 0,
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "build/bdist.linux-i686/egg/monetdb/sql/cursors.py", line 208, in execute
File "build/bdist.linux-i686/egg/monetdb/sql/cursors.py", line 450, in __store_result
File "build/bdist.linux-i686/egg/monetdb/sql/cursors.py", line 534, in __parse_tuple
File "build/bdist.linux-i686/egg/monetdb/sql/cursors.py", line 574, in __exception_handler
monetdb.monetdb_exceptions.InterfaceError: length of row doesn't match header
## Comment 13208
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2989601 at http://sourceforge.net/support/tracker.php?aid=2989601
## Comment 14021
Date: 2010-05-13 14:00:26 +0200
From: @gijzelaerr
I think I isolated the problem.
Is it possible for you to do download this file:
http://dev.monetdb.org/hg/MonetDB/raw-file/3135505afb9c/python/monetdb/mapi2.py
and replace it with your mapi2.py file in the monetdb python dbapi installation folder?
You can also do a complete hg checkout of the monetdb repository and install from the python folder, but the tree is quite big.
## Comment 14022
Date: 2010-05-13 14:21:36 +0200
From: @gijzelaerr
sorry, it should be this file:
http://dev.monetdb.org/hg/MonetDB/file/52f391796452/python/monetdb/mapi2.py
## Comment 14058
Date: 2010-05-29 10:43:56 +0200
From: @gijzelaerr
I think this is fixed, reopen the bug if you/somebody thinks it is not
| Python Error "length of row doesn't match header" | https://api.github.com/repos/MonetDB/MonetDB/issues/2355/comments | 0 | 2020-11-30T10:39:19Z | 2024-06-27T11:44:26Z | https://github.com/MonetDB/MonetDB/issues/2355 | 753,362,386 | 2,355 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-14 11:43:42 +0200
From: @skinkie
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:18:33 +0200
## Comment 13201
Date: 2010-04-14 23:43:42 +0200
From: @skinkie
When concatinating constant strings, the maximum column length is based on the last column type. It can be reproduced with:
create table test2(a varchar(256), b int, c varchar(16));
insert into test2 values ('testaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa', 1, 'test');
select a||' '||b||' '||c from test2;
SQLException:str_cast:value too long for type (var)char(16)
## Comment 13202
Date: 2010-04-14 23:48:14 +0200
From: @skinkie
As work around:
select cast(a||' '||b as varchar(256))||' '||c from test2;
## Comment 13203
Date: 2010-04-15 13:59:00 +0200
From: @skinkie
Workaround revealed something as well. null results for the ones that would 'overflow'.
## Comment 13204
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2987398 at http://sourceforge.net/support/tracker.php?aid=2987398
## Comment 14120
Date: 2010-06-18 03:45:02 +0200
From: @skinkie
create table testmore2(test1 varchar(1), test2 int, test3 varchar(3));
insert into testmore2 values ('1', 23, '456');
select length(test1||' '||test2||' '||test3) from testmore2 ;
SQLException:str_cast:value too long for type (var)char(3)
## Comment 14650
Date: 2010-08-12 19:38:22 +0200
From: @njnes
Fixed by improving the function resolution.
## Comment 14652
Date: 2010-08-12 19:47:00 +0200
From: @njnes
Changeset [cddee0e5256a](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=cddee0e5256a) made by Niels Nes <niels@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=cddee0e5256a](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=cddee0e5256a)
Changeset description:
fixed bug #2354.
Improved function resolution for binary operations.
## Comment 14827
Date: 2010-08-30 09:18:33 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| SQL: concatination length is based on last column | https://api.github.com/repos/MonetDB/MonetDB/issues/2354/comments | 0 | 2020-11-30T10:39:15Z | 2024-06-27T11:44:25Z | https://github.com/MonetDB/MonetDB/issues/2354 | 753,362,344 | 2,354 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-11 03:53:53 +0200
From: @skinkie
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:18:30 +0200
## Comment 13199
Date: 2010-04-11 03:53:53 +0200
From: @skinkie
This one is in the series bizarre queries that monkeys with a keyboard can come up with [in the middle of the night].
create table (test serial, kvks int);
select count(kvks) from (select x2.t2 as fr, x1.t1 as t from (select row_number() over (order by k1.kvks) as rn, k1.kvks as t1, k2.kvks as t2 from test as k1, test as k2 where (k1.id + 1) = k2.id AND (k1.kvks + 1000) < k2.kvks) as x1, (select row_number() over (order by k1.kvks) as rn, k1.kvks as t1, k2.kvks as t2 from test as k1, test as k2 where (k1.id + 1) = k2.id AND (k1.kvks + 1000) < k2.kvks) as x2 where (x2.rn + 1) = x1.rn) as y, test where kvks between fr and t group by kvks;
Server console outputs:
could not find x1.t1
test.kvks
test.%TID%
test.test_id_pkey
x2.rn
x2.t2
k1.%TID%
k2.%TID%
mserver5: ../../../src/server/rel_bin.mx:535: exp_bin: Assertion `0' failed.
Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7fffe3e72710 (LWP 12047)]
0x00007ffff46fe435 in raise () from /lib/libc.so.6
(gdb) bt
0 0x00007ffff46fe435 in raise () from /lib/libc.so.6
1 0x00007ffff46ff931 in abort () from /lib/libc.so.6
2 0x00007ffff46f7690 in __assert_fail () from /lib/libc.so.6
3 0x00007fffe4563cee in exp_bin (sql=0x1315d98, e=0x1360a88, left=0x1361068, right=0x13ef698, grp=0x0, sel=0x0) at ../../../src/server/rel_bin.mx:535
4 0x00007fffe456510f in rel2bin_join (sql=0x1315d98, rel=0x13a4c88, refs=0x1360778) at ../../../src/server/rel_bin.mx:865
5 0x00007fffe456f273 in subrel_bin (sql=0x1315d98, rel=0x13a4c88, refs=0x1360778) at ../../../src/server/rel_bin.mx:3818
6 0x00007fffe4564fda in rel2bin_join (sql=0x1315d98, rel=0x13a3de8, refs=0x1360778) at ../../../src/server/rel_bin.mx:849
7 0x00007fffe456f273 in subrel_bin (sql=0x1315d98, rel=0x13a3de8, refs=0x1360778) at ../../../src/server/rel_bin.mx:3818
8 0x00007fffe4567527 in rel2bin_project (sql=0x1315d98, rel=0x13a4018, refs=0x1360778, topn=0x0) at ../../../src/server/rel_bin.mx:1547
9 0x00007fffe456f340 in subrel_bin (sql=0x1315d98, rel=0x13a4018, refs=0x1360778) at ../../../src/server/rel_bin.mx:3839
10 0x00007fffe456823e in rel2bin_groupby (sql=0x1315d98, rel=0x13a4e98, refs=0x1360778) at ../../../src/server/rel_bin.mx:1808
11 0x00007fffe456f390 in subrel_bin (sql=0x1315d98, rel=0x13a4e98, refs=0x1360778) at ../../../src/server/rel_bin.mx:3847
12 0x00007fffe4567527 in rel2bin_project (sql=0x1315d98, rel=0x13a4f68, refs=0x1360778, topn=0x0) at ../../../src/server/rel_bin.mx:1547
13 0x00007fffe456f340 in subrel_bin (sql=0x1315d98, rel=0x13a4f68, refs=0x1360778) at ../../../src/server/rel_bin.mx:3839
14 0x00007fffe456f587 in output_rel_bin (sql=0x1315d98, rel=0x13a4f68) at ../../../src/server/rel_bin.mx:3901
15 0x00007fffe44a290c in sql_symbol2stmt (c=0x1315d98, sym=0x13f6d08) at ../../../../src/backends/monet5/sql.mx:1352
16 0x00007fffe44f18ed in SQLparser (c=0x605878) at ../../../../src/backends/monet5/sql_scenario.mx:1154
17 0x00007ffff7b0502f in runPhase (c=0x605878, phase=1) at ../../../src/mal/mal_scenario.mx:602
18 0x00007ffff7b05165 in runScenarioBody (c=0x605878) at ../../../src/mal/mal_scenario.mx:643
19 0x00007ffff7b053bd in runScenario (c=0x605878) at ../../../src/mal/mal_scenario.mx:680
20 0x00007ffff7abd3d3 in MSserveClient (dummy=0x605878) at ../../../src/mal/mal_session.mx:474
21 0x00007ffff5669547 in start_thread () from /lib/libpthread.so.0
22 0x00007ffff479fc7d in clone () from /lib/libc.so.6
23 0x0000000000000000 in ?? ()
## Comment 13200
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2985237 at http://sourceforge.net/support/tracker.php?aid=2985237
## Comment 14657
Date: 2010-08-12 21:52:02 +0200
From: @njnes
added test sql/src/test/BugTracker-2010/Tests/crash_on_complex_join_exp.Bug-2353.sql
## Comment 14658
Date: 2010-08-12 22:13:25 +0200
From: @njnes
Changeset [957466674744](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=957466674744) made by Niels Nes <niels@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=957466674744](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=957466674744)
Changeset description:
fixed bug #2353
We now handle range joins in the relational optimizer properly.
## Comment 14820
Date: 2010-08-30 09:18:30 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| SQL: exp_bin: Assertion `0' | https://api.github.com/repos/MonetDB/MonetDB/issues/2353/comments | 0 | 2020-11-30T10:39:12Z | 2024-06-27T11:44:24Z | https://github.com/MonetDB/MonetDB/issues/2353 | 753,362,312 | 2,353 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-11 02:32:39 +0200
From: @skinkie
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:18:28 +0200
## Comment 13196
Date: 2010-04-11 02:32:39 +0200
From: @skinkie
create table test (test integer);
select test from test where test in (test + 1);
Backtrace:
mserver5: ../../../src/server/rel_select.mx:738: rel_lastexp: Assertion `(rel->op == op_project || rel->op == op_groupby || (rel->op == op_union || rel->op == op_inter || rel->op == op_except))' failed.
Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7fffe427d710 (LWP 11713)]
0x00007ffff46fe435 in raise () from /lib/libc.so.6
(gdb) bt
0 0x00007ffff46fe435 in raise () from /lib/libc.so.6
1 0x00007ffff46ff931 in abort () from /lib/libc.so.6
2 0x00007ffff46f7690 in __assert_fail () from /lib/libc.so.6
3 0x00007fffe4948724 in rel_lastexp (sql=0x1271808, rel=0x12a6c18) at ../../../src/server/rel_select.mx:738
4 0x00007fffe494e951 in rel_logical_exp (sql=0x1271808, rel=0x12a6c18, sc=0x13234f8, f=1) at ../../../src/server/rel_select.mx:2628
5 0x00007fffe4954a35 in rel_select_exp (sql=0x1271808, rel=0x12a6c18, outer=0x0, sn=0x13235f8, ek=...) at ../../../src/server/rel_select.mx:4311
6 0x00007fffe4955ce1 in rel_query (sql=0x1271808, rel=0x0, sq=0x13235f8, toplevel=1, ek=...) at ../../../src/server/rel_select.mx:4616
7 0x00007fffe4957019 in rel_subquery (sql=0x1271808, rel=0x0, sq=0x13235f8, ek=...) at ../../../src/server/rel_select.mx:4959
8 0x00007fffe4957090 in rel_selects (sql=0x1271808, s=0x13235f8) at ../../../src/server/rel_select.mx:4970
9 0x00007fffe49469ce in rel_semantic (sql=0x1271808, s=0x13235f8) at ../../../src/server/rel_semantic.mx:334
10 0x00007fffe48ad881 in sql_symbol2stmt (c=0x1271808, sym=0x13235f8) at ../../../../src/backends/monet5/sql.mx:1338
11 0x00007fffe48fc8ed in SQLparser (c=0x605878) at ../../../../src/backends/monet5/sql_scenario.mx:1154
12 0x00007ffff7b0502f in runPhase (c=0x605878, phase=1) at ../../../src/mal/mal_scenario.mx:602
13 0x00007ffff7b05165 in runScenarioBody (c=0x605878) at ../../../src/mal/mal_scenario.mx:643
14 0x00007ffff7b053bd in runScenario (c=0x605878) at ../../../src/mal/mal_scenario.mx:680
15 0x00007ffff7abd3d3 in MSserveClient (dummy=0x605878) at ../../../src/mal/mal_session.mx:474
16 0x00007ffff5669547 in start_thread () from /lib/libpthread.so.0
17 0x00007ffff479fc7d in clone () from /lib/libc.so.6
18 0x0000000000000000 in ?? ()
## Comment 13197
Date: 2010-04-11 02:33:47 +0200
From: @skinkie
Even the + 1 is not required.
## Comment 13198
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2985226 at http://sourceforge.net/support/tracker.php?aid=2985226
## Comment 14656
Date: 2010-08-12 20:05:20 +0200
From: @njnes
added test crash_on_in_+1.Bug-2352.sql
## Comment 14815
Date: 2010-08-30 09:18:28 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| SQL: in (test + 1) | https://api.github.com/repos/MonetDB/MonetDB/issues/2352/comments | 0 | 2020-11-30T10:39:09Z | 2024-06-27T11:44:23Z | https://github.com/MonetDB/MonetDB/issues/2352 | 753,362,277 | 2,352 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-08 03:02:06 +0200
From: @grobian
To: SQL devs <<bugs-sql>>
Version: 2.40.1 (Oct2010) [obsolete]
Last updated: 2011-03-28 17:31:32 +0200
## Comment 13194
Date: 2010-04-08 15:02:06 +0200
From: @grobian
When batching the airtraffic database (air3 file) using a prepared statement to the server, a transaction rollback happens when the server is in autocommit mode. When not in autocommit mode, the loading process finishes successfully.
Also when commits are done inbetween the batches a transaction abort failure is returned. It feels like a commit in the process makes every next exec fail.
The attached java file can be compiled using:
% javac PrepareBatchCrash.java
And then run:
% java -cp .:$INSTALL_DIR/share/MonetDB/lib/monetdb-1.14-jdbc.jar PrepareBatchCrash "jdbc:monetdb://localhost/somedb?user=monetdb&password=monetdb" /var/tmp/air3
You can find the air3 file in /var/tmp on volund
## Comment 13195
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2983861 at http://sourceforge.net/support/tracker.php?aid=2983861
## Comment 15427
Date: 2011-01-25 15:19:38 +0100
From: @grobian
This bug has to do with the cleanup of prepared handles on the server. It has been improved and fixed in the current development branch.
## Comment 15629
Date: 2011-03-28 17:31:32 +0200
From: @sjoerdmullender
The Mar2011 version has been released.
| SQL: transaction aborts on prepared statements in autocommit | https://api.github.com/repos/MonetDB/MonetDB/issues/2351/comments | 0 | 2020-11-30T10:39:06Z | 2024-06-27T11:44:23Z | https://github.com/MonetDB/MonetDB/issues/2351 | 753,362,233 | 2,351 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-08 12:03:51 +0200
From: Wouter Alink <<alink>>
To: Wouter Alink <<alink>>
Version: -- development
CC: @sjoerdmullender
Last updated: 2011-09-30 11:41:19 +0200
## Comment 13189
Date: 2010-04-08 12:03:51 +0200
From: Wouter Alink <<alink>>
Using Feb2010-HEAD (yesterday's checkout)
When using the minimal optimizer pipeline, the mserver5 instance crashes upon second connection from a client. (seems to be always reproducible)
The following describes the issue:
$ monetdb get all localdb002
name prop source value
localdb002 name - localdb002
localdb002 forward default proxy
localdb002 shared default yes
localdb002 nthreads default <unknown>
localdb002 optpipe default default_pipe
localdb002 master default no
localdb002 slave default <unknown>
$ monetdb stop localdb002
stopping database 'localdb002'... done
$ monetdb set optpipe=inline,remap,deadcode,multiplex localdb002
$ monetdb start localdb002
starting database 'localdb002'... done
$ mclient -lsql -dlocaldb002 -umonetdb -Pmonetdb
Welcome to mclient, the MonetDB/SQL interactive terminal
Database: MonetDB v5.18.4, 'localdb002'
Type \q to quit, \? for a list of available commands
auto commit mode: on
sql>^C
$ mclient -lsql -dlocaldb002 -umonetdb -Pmonetdb
Connection terminated
$ monetdb stop localdb002
stopping database 'localdb002'... FAILED
stop: database is not running: localdb002
$ monetdb set optpipe=default_pipe localdb002
$ monetdb start localdb002
starting database 'localdb002'... done
$ mclient -lsql -dlocaldb002 -umonetdb -Pmonetdb
Welcome to mclient, the MonetDB/SQL interactive terminal
Database: MonetDB v5.18.4, 'localdb002'
Type \q to quit, \? for a list of available commands
auto commit mode: on
sql>^C
$ mclient -lsql -dlocaldb002 -umonetdb -Pmonetdb
Welcome to mclient, the MonetDB/SQL interactive terminal
Database: MonetDB v5.18.4, 'localdb002'
Type \q to quit, \? for a list of available commands
auto commit mode: on
sql>^C
$
If there is a debugger attached to the mserver5 instance with the minimal optimizer pipe, then the following can be observed:
Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7f8937fff710 (LWP 29782)]
0x00000039f62326c5 in raise () from /lib64/libc.so.6
(gdb) bt
0 0x00000039f62326c5 in raise () from /lib64/libc.so.6
1 0x00000039f6233ea5 in abort () from /lib64/libc.so.6
2 0x00000039f622b7b5 in __assert_fail () from /lib64/libc.so.6
3 0x00007f89539aa43a in GDKfree_ (blk=0xf2ec60)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB/src/gdk/gdk_utils.mx:1150
4 0x00007f89539aa676 in GDKfree (blk=0xf2ec60)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB/src/gdk/gdk_utils.mx:1193
5 0x00007f893d64bf92 in global_variables (sql=0x7f89300d6708, user=0x7f893d708403 "monetdb", schema=0x7f893d7083ff "sys")
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:326
6 0x00007f893d64c586 in SQLinitClient (c=0x605b38)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:419
7 0x00007f89543fdddb in runPhase (c=0x605b38, phase=5)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:602
8 0x00007f89543fde83 in runScenarioBody (c=0x605b38)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:633
9 0x00007f89543fe181 in runScenario (c=0x605b38)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:679
10 0x00007f89543b1378 in MSserveClient (dummy=0x605b38)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_session.mx:469
11 0x00000039f6e06a3a in start_thread () from /lib64/libpthread.so.0
12 0x00000039f62de67d in clone () from /lib64/libc.so.6
13 0x0000000000000000 in ?? ()
(gdb)
## Comment 13190
Date: 2010-04-08 12:42:09 +0200
From: @drstmane
Indeed reproducible; here's a more detailed gdb log:
$ gdb --args mserver5 --dbinit='include sql;' --set sql_optimizer=inline,remap,deadcode,multiplex
GNU gdb (GDB) Fedora (7.0.1-37.fc12)
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /net/rig.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Candidate/prefix.--enable-strict_--disable-optimize_--enable-debug_--enable-assert/bin/mserver5...done.
(gdb) r
Starting program: /net/rig.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Candidate/prefix.--enable-strict_--disable-optimize_--enable-debug_--enable-assert/bin/mserver5 --dbinit=include\ sql\; --set sql_optimizer=inline,remap,deadcode,multiplex
[Thread debugging using libthread_db enabled]
[New Thread 0x7fffea987710 (LWP 15698)]
MonetDB server v5.18.2, based on kernel v1.36.3
Serving database 'demo', using 4 threads
Compiled for x86_64-unknown-linux-gnu/64bit with 64bit OIDs dynamically linked
Found 7.751 GiB available main-memory.
Copyright (c) 1993-July 2008 CWI.
Copyright (c) August 2008-2010 MonetDB B.V., all rights reserved
Visit http://monetdb.cwi.nl/ for further information
warning: please don't forget to set your vault key!
(see /ufs/manegold/_/scratch0/Monet/Testing/Candidate/prefix.--enable-strict_--disable-optimize_--enable-debug_--enable-assert/etc/monetdb5.conf)
[New Thread 0x7fffe14b4710 (LWP 15699)]
Listening for connection requests on mapi:monetdb://127.0.0.1:50000/
MonetDB/SQL module v2.36.2 loaded
[New Thread 0x7fffdfd88710 (LWP 15700)]
>
>io.print("1st client connection");
[ "1st client connection" ]
>[New Thread 0x7fffdfb87710 (LWP 15712)]
[Thread 0x7fffdfb87710 (LWP 15712) exited]
i
>io.print("1st client disconnected");
[ "1st client disconnected" ]
>
>io.print("2nd client connection");
[ "2nd client connection" ]
>[New Thread 0x7fffdf986710 (LWP 15734)]
mserver5: /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB/src/gdk/gdk_utils.mx:1150: GDKfree_: Assertion `(size & 2) == 0' failed.
Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7fffdf986710 (LWP 15734)]
0x00000033d7a326c5 in raise () from /lib64/libc.so.6
Missing separate debuginfos, use: debuginfo-install bzip2-libs-1.0.5-6.fc12.x86_64 cyrus-sasl-lib-2.1.23-8.fc12.x86_64 glibc-2.11.1-1.x86_64 keyutils-libs-1.2-6.fc12.x86_64 krb5-libs-1.7.1-6.fc12.x86_64 libcom_err-1.41.9-7.fc12.x86_64 libcurl-7.19.7-9.fc12.x86_64 libidn-1.9-5.x86_64 libselinux-2.0.90-5.fc12.x86_64 libssh2-1.2.4-1.fc12.x86_64 libuuid-2.16.2-7.fc12.x86_64 libxml2-2.7.6-1.fc12.x86_64 libxslt-1.1.26-1.fc12.x86_64 ncurses-libs-5.7-3.20090207.fc12.x86_64 nspr-4.8.4-2.fc12.x86_64 nss-3.12.6-1.2.fc12.x86_64 nss-softokn-3.12.4-15.fc12.x86_64 nss-softokn-freebl-3.12.4-15.fc12.x86_64 nss-util-3.12.6-1.fc12.x86_64 openldap-2.4.19-3.fc12.x86_64 openssl-1.0.0-0.13.beta4.fc12.x86_64 pcre-7.8-3.fc12.x86_64 raptor-1.4.18-5.fc12.x86_64 readline-6.0-3.fc12.x86_64 sqlite-3.6.20-1.fc12.x86_64 zlib-1.2.3-23.fc12.x86_64
(gdb) bt
0 0x00000033d7a326c5 in raise () from /lib64/libc.so.6
1 0x00000033d7a33ea5 in abort () from /lib64/libc.so.6
2 0x00000033d7a2b7b5 in __assert_fail () from /lib64/libc.so.6
3 0x00007ffff72d1bde in GDKfree_ (blk=0x637408) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB/src/gdk/gdk_utils.mx:1150
4 0x00007ffff72d1e1a in GDKfree (blk=0x637408) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB/src/gdk/gdk_utils.mx:1193
5 0x00007fffe0f74042 in global_variables (sql=0x7fffd00d8d28, user=0x7fffe1030563 "monetdb", schema=0x7fffe103055f "sys") at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/sql/src/backends/monet5/sql_scenario.mx:326
6 0x00007fffe0f74636 in SQLinitClient (c=0x605b38) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/sql/src/backends/monet5/sql_scenario.mx:419
7 0x00007ffff7d23d23 in runPhase (c=0x605b38, phase=5) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB5/src/mal/mal_scenario.mx:602
8 0x00007ffff7d23dcb in runScenarioBody (c=0x605b38) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB5/src/mal/mal_scenario.mx:633
9 0x00007ffff7d240c9 in runScenario (c=0x605b38) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB5/src/mal/mal_scenario.mx:679
10 0x00007ffff7cd732c in MSserveClient (dummy=0x605b38) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB5/src/mal/mal_session.mx:469
11 0x00000033d8606a3a in start_thread () from /lib64/libpthread.so.0
12 0x00000033d7ade67d in clone () from /lib64/libc.so.6
13 0x0000000000000000 in ?? ()
(gdb) up
1 0x00000033d7a33ea5 in abort () from /lib64/libc.so.6
(gdb)
2 0x00000033d7a2b7b5 in __assert_fail () from /lib64/libc.so.6
(gdb)
3 0x00007ffff72d1bde in GDKfree_ (blk=0x637408) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB/src/gdk/gdk_utils.mx:1150
1150 assert((size & 2) == 0);
(gdb) print size
$1 = 2217590464167432338
(gdb) print size&2
$2 = 2
(gdb) print blk
$3 = (void *) 0x637408
(gdb) up
4 0x00007ffff72d1e1a in GDKfree (blk=0x637408) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/MonetDB/src/gdk/gdk_utils.mx:1193
1193 GDKfree_(blk);
(gdb)
5 0x00007fffe0f74042 in global_variables (sql=0x7fffd00d8d28, user=0x7fffe1030563 "monetdb", schema=0x7fffe103055f "sys") at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/sql/src/backends/monet5/sql_scenario.mx:326
326 GDKfree(optimizer);
(gdb) li
321 SQLglobal("current_role", user);
322 optimizer= GDKgetenv("sql_optimizer");
323 if (optimizer == NULL)
324 optimizer= "default_pipe";
325 SQLglobal("optimizer", optimizer=setOptimizers(optimizer));
326 GDKfree(optimizer);
327 SQLglobal("trace","show,ticks,stmt");
328
329 typename = "sec_interval";
330 sql_find_subtype(&ctype, typename, 0, 0);
(gdb) print optimizer
$4 = (str) 0x637408 "inline,remap,deadcode,multiplex"
(gdb)
## Comment 13191
Date: 2010-04-08 12:56:05 +0200
From: @drstmane
this might be due to a double or illegal free, as setOptimizers() return a copy in all but one case, and that one case might trigger here:
========
@c
str
setOptimizers(str optimizer)
{
int top=0;
char *base, *nxt, *nme, *pipe="notdefined";
/* do nothing if the pipe line is already set */
if ( optimizerpipe && optimizer && strcmp(optimizerpipe,optimizer) == 0 )
return optimizerpipe;
[...]
========
In fact (as a side note) there are various places where setOptimizers() is called without assigning (and hence using, let alone freeing) its return value; I guess this can be considered a leak in case setOptimizers() returns a GDKstrdup() ...
## Comment 13192
Date: 2010-04-08 18:33:10 +0200
From: @drstmane
fixed in CVS (Feb2010 branch):
===================================================================
2010/04/08 - stmane: sql/src/backends/monet5/sql_optimizer.mx,1.243.2.4
sql/src/backends/monet5/sql_scenario.mx,1.372.2.3
fixing ID: 2983773 "SQL: minimal optimizer pipe unstable"
https://sourceforge.net/tracker/index.php?func=detail&aid=2983773&group_id=56967&atid=482468
In the code of and related to setOptimizers(),
make sure we use GDKstrdup() & GDKfree() only when indeed required
and thus not only prevent double free's
but also avoid memory leakage.
===================================================================
Wouter,
could you please double-check whether it now also works for you,
and let me know whether/when you could add a test for this one?
Thanks!
## Comment 13193
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2983773 at http://sourceforge.net/support/tracker.php?aid=2983773
| SQL: minimal optimizer pipe unstable | https://api.github.com/repos/MonetDB/MonetDB/issues/2350/comments | 0 | 2020-11-30T10:39:03Z | 2024-06-27T11:44:21Z | https://github.com/MonetDB/MonetDB/issues/2350 | 753,362,191 | 2,350 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-06 03:32:18 +0200
From: jods <<jods4>>
To: GDK devs <<bugs-common>>
Version: -- development
CC: alex.bodinner, holger.pirk, @drstmane
Last updated: 2012-11-27 15:19:40 +0100
## Comment 13185
Date: 2010-04-06 15:32:18 +0200
From: jods <<jods4>>
A clean, fresh install of the MonetDB5 Server (release 2010/03/22, 32 bits, i386) crashes on Windows XP when trying to start it up.
Here's the trace left inside the Windows event log:
Faulting application mserver5.exe, version 0.0.0.0, faulting module libbat.dll, version 1.36.3.1, fault address 0x00459ba7.
## Comment 13186
Date: 2010-04-15 14:47:29 +0200
From: @sjoerdmullender
Could it be that you still have old DLLs lying around?
In the past some DLLs were installed in the system folder (C:\Windows\system32\lib*.dll). If you still have those DLLs, you may well have problems starting the server.
In recent versions, we don't need those files anymore, so it should be save to remove them.
The files that we used to copy are (the latter three will give problems with the server):
libMonetODBC.dll
libMonetODBCs.dll
libMapi.dll
libstream.dll
libmutils.dll
## Comment 13187
Date: 2010-04-15 19:17:47 +0200
From: jods <<jods4>>
No, as it was the first time MonetDB was installed on that machine.
## Comment 13188
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2982639 at http://sourceforge.net/support/tracker.php?aid=2982639
## Comment 14085
Date: 2010-06-09 11:11:48 +0200
From: Alex Bodinner <<alex.bodinner>>
(In reply to comment 2)
> No, as it was the first time MonetDB was installed on that machine.
I have also run into this exact crash on a XP 32 bit machine.
(3GB on board with Physical address extension)
I have tried it with Aug09-SP1 & Feb10-SP1 and both crash.
I can't reproduce it on other machines.
Is there anything I can get from that machine to help figure this out?
## Comment 15061
Date: 2010-10-18 19:30:00 +0200
From: @drstmane
jods, Alex,
do you still experience the reported problems with Jun2010-SP2?
## Comment 18111
Date: 2012-11-27 15:19:40 +0100
From: Holger <<holger.pirk>>
Not sure if this is still an issue. If it is, please provide a server log.
| Crash during server startup | https://api.github.com/repos/MonetDB/MonetDB/issues/2349/comments | 0 | 2020-11-30T10:39:00Z | 2024-06-27T11:44:20Z | https://github.com/MonetDB/MonetDB/issues/2349 | 753,362,159 | 2,349 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-05 04:32:45 +0200
From: Stephen Benz <<sbenz>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2010-04-07 02:25:22 +0200
## Comment 13182
Date: 2010-04-05 16:32:45 +0200
From: Stephen Benz <<sbenz>>
Appears as though a large number of dbs (>60) causes the command `monetdb status` to trip up. (our particular instance is attempting to run 130 dbs)
It prints the following warning:
WARNING: failed to parse response from merovingian: IllegalArgumentException:sabaoth.deserialise:string does not contain path: /usr/local/MonetDB/var/Monet
(it then successfully lists 54 of the databases)
(we have our monetdb installed in /usr/local)
The merovingian.log doesn't show any errors, just:
2010-04-05 09:55:58 MSG control[2260]: (local): served status list
There also doesn't appear to be any odd characters in any of the db names.
OS: CentOS Linux
gcc version 4.1.2 20080704 (Red Hat 4.1.2-46)
## Comment 13183
Date: 2010-04-07 14:25:22 +0200
From: @grobian
Confirmed and fixed. A forgotten increment of an enlarged buffer cut off the results returned by merovingian.
Thanks for the report!
## Comment 13184
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2982104 at http://sourceforge.net/support/tracker.php?aid=2982104
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| monetdb status throws warning with large num of dbs | https://api.github.com/repos/MonetDB/MonetDB/issues/2348/comments | 0 | 2020-11-30T10:38:57Z | 2024-06-27T11:44:20Z | https://github.com/MonetDB/MonetDB/issues/2348 | 753,362,126 | 2,348 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-02 10:01:53 +0200
From: Matthew W. Jones <<matburt>>
To: @sjoerdmullender
Version: -- development
CC: @drstmane
Last updated: 2010-10-18 19:27:53 +0200
## Comment 13172
Date: 2010-04-02 22:01:53 +0200
From: Matthew W. Jones <<matburt>>
The Feb2010 SP1 superball fails to build... I also get the failure trying to build the server package individually
I'm attaching the logfile here's a snippet of the failure:
(...)
config.status: creating mal_config.h
config.status: executing depfiles commands
config.status: executing libtool commands
make all-recursive
make[1]: Entering directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5'
Making all in src
make[2]: Entering directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src'
make all-recursive
make[3]: Entering directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src'
Making all in mal
make[4]: Entering directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src/mal'
make[4]: Leaving directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src/mal'
make[4]: *** No rule to make target `mal.c', needed by `all'. Stop.
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src'
make[2]: Leaving directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src'
make[1]: Leaving directory `/var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5'
Please refer to the full log at /var/tmp/MonetDB-1270235519/log
If you believe this is an error in the script or the software,
file a bug and attach the logfile.
make[2]: *** [all] Error 2
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
making MonetDB5-server failed
## Comment 13173
Date: 2010-04-03 14:26:40 +0200
From: @drstmane
Works fine for me.
Could you please share with us your exact commands to
- unpack the tar ball
- call monetdb-install.sh
as well as the result of
ls -alFrt MonetDB-Feb2010-SuperBall-SP1/MonetDB5-server/MonetDB5/src/mal/mal.*
Thanks!
## Comment 13174
Date: 2010-04-03 14:35:19 +0200
From: @grobian
From the buildlog:
called with arguments: --prefix=/root/ahem --enable-sql
## Comment 13175
Date: 2010-04-03 14:46:02 +0200
From: @drstmane
and possibly also the output of
ls -alFrt /var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src/mal/mal.*
## Comment 13176
Date: 2010-04-03 14:53:38 +0200
From: @drstmane
[restored Matthew's Summary --- SF big tracker is not transaction-save]
## Comment 13177
Date: 2010-04-03 17:27:56 +0200
From: @drstmane
Basically, we need to know,
whether the file mal.c does exist
(1) in the extracted sources in /root/packages/MonetDB-Feb2010-SuperBall-SP1/MonetDB5-server/MonetDB5/src/mal/ and
(2) in the build directory in /var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src/mal/ ,
and whether it is newer than the file mal.mx in the same directories.
In my case, I get
$ ls -alFrt MonetDB5-server/MonetDB5/src/mal/mal.*
-rw-r--r-- 1 16932 2010-02-15 14:47 MonetDB5-server/MonetDB5/src/mal/mal.mx
-rw-r--r-- 1 4006 2010-03-22 11:36 MonetDB5-server/MonetDB5/src/mal/mal.c
-rw-r--r-- 1 2 2010-03-22 11:36 MonetDB5-server/MonetDB5/src/mal/mal.mal
-rw-r--r-- 1 2846 2010-03-22 11:36 MonetDB5-server/MonetDB5/src/mal/mal.h
both for the extracted source and the build directory.
## Comment 13178
Date: 2010-04-05 15:58:31 +0200
From: Matthew W. Jones <<matburt>>
It does appear to exist in the source directory:
SP1 ls -alFrt MonetDB5-server/MonetDB5/src/mal/mal.*
-rw-r--r-- 1 234 2000 16932 2010-02-15 08:47 MonetDB5-server/MonetDB5/src/mal/mal.mx
-rw-r--r-- 1 234 2000 4006 2010-03-22 06:36 MonetDB5-server/MonetDB5/src/mal/mal.c
-rw-r--r-- 1 234 2000 2 2010-03-22 06:36 MonetDB5-server/MonetDB5/src/mal/mal.mal
-rw-r--r-- 1 234 2000 2846 2010-03-22 06:36 MonetDB5-server/MonetDB5/src/mal/mal.h
but not in tmp:
ls -alFrt /var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src/mal/mal.*
-rw-r--r-- 1 234 2000 16932 2010-02-15 08:47 /var/tmp/MonetDB-1270235519/MonetDB5-server/MonetDB5/src/mal/mal.mx
## Comment 13179
Date: 2010-04-06 00:12:43 +0200
From: @drstmane
This is strange; it suggests that mal.{h,c,mal} somehow got "lost" while or after monetdb-install.sh copied the sources to the build directory; however, your log does not indicate any error during the copy:
Fri Apr 2 15:11:59 EDT 2010
called with arguments: --prefix=/root/bullshit --enable-sql
Copying /root/packages/MonetDB-Feb2010-SuperBall-SP1/MonetDB to /var/tmp/MonetDB-1270235519
Copying /root/packages/MonetDB-Feb2010-SuperBall-SP1/MonetDB-client to /var/tmp/MonetDB-1270235519
Copying /root/packages/MonetDB-Feb2010-SuperBall-SP1/MonetDB5-server to /var/tmp/MonetDB-1270235519
Copying /root/packages/MonetDB-Feb2010-SuperBall-SP1/MonetDB-SQL to /var/tmp/MonetDB-1270235519
* Processing MonetDB
Could you just try to call monetdb-install.sh, again, and report what happens?
## Comment 13180
Date: 2010-04-14 14:25:41 +0200
From: @sjoerdmullender
I have not been able to reproduce this on a clean Ubuntu 8.04 (64 bit) system.
I installed the necessary packages to compile MonetDB, downloaded and extracted the Feb2010-SP1 super source tar ball, and ran monetdb-install.sh --enable-sql, with and without an addition --prefix.
Did you, by any chance, do a "make clean" in one of the directories involved in the process?
## Comment 13181
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2981299 at http://sourceforge.net/support/tracker.php?aid=2981299
## Comment 15060
Date: 2010-10-18 19:27:53 +0200
From: @drstmane
Closing as we could not reporduce the problem and Feb2010-SP1 is no longer supported.
Please use Jun2010-SP2, instead.
Please file a new bug report, in case you encounter any problems with Jun2010-SP2.
| Feb2010-SP1 SB build failure on Ubuntu Linux 8.04 64bits | https://api.github.com/repos/MonetDB/MonetDB/issues/2347/comments | 0 | 2020-11-30T10:38:53Z | 2024-06-28T07:25:41Z | https://github.com/MonetDB/MonetDB/issues/2347 | 753,362,083 | 2,347 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-04-01 01:26:34 +0200
From: zeuner <<zeuner>>
To: SQL devs <<bugs-sql>>
Version: 11.5.3 (Aug2011-SP1) [obsolete]
CC: @njnes
Last updated: 2011-10-26 13:22:00 +0200
## Comment 13167
Date: 2010-04-01 01:26:34 +0200
From: zeuner <<zeuner>>
I am unable to trigger this issue using the CLI which is why I attach a C
program in this report.
The symptom this Mapi error message:
ERROR = !MALException:mkey.bulk_rotate_xor_hash:Operation failed
Operands not synced on head
The issue has been verified on a 2010-03-30 (versions *.*.4) checkout
on an x86_64 dual core running CentOS 5.4.
## Comment 13168
Date: 2010-04-08 21:19:48 +0200
From: @njnes
With your program I get (correctly) a primary key violation
error at line 993
MAPI = monetdb@localhost:50000
QUERY = INSERT
INTO transformed (base_value,general_attribute,compound,mode,activity,legacy_index,refilled,value)
SELECT gathered.base_value,general_attribute,compound,gathered.mode,state,legacy_index,refilled,value
FROM gathered,special_issue
WHERE activity=-1
AND gathered.mode=special_issue.mode
AND gathered.base_value=special_issue.base_value
MAPI = monetdb@localhost:50000
QUERY = INSERT
INTO transformed (base_value,general_attribute,compound,mode,activity,legacy_index,refilled,value)
SELECT gathered.base_value,general_attribute,compound,gathered.mode,state,legacy_index,refilled,value
FROM gathered,special_issue
WHERE activity=-1
AND gathered.mode=special_issue.mode
AND gathered.base_value=special_issue.base_value
ERROR = !SQLException:assert:INSERT INTO: PRIMARY KEY constraint 'transformed.transformed_compound_refilled_pkey' violated
## Comment 13169
Date: 2010-04-09 10:49:56 +0200
From: zeuner <<zeuner>>
Did you by any chance use the program from [1] instead? The one from this bug only has 452 lines.
FYI, I could also reproduce it with a 2010-04-09 checkout, both on the x86_64 dual core mentioned above and on a single-cored i386 system.
[1] https://sourceforge.net/tracker/?func=detail&aid=2942744&group_id=56967&atid=482468
## Comment 13170
Date: 2010-04-13 20:42:48 +0200
From: zeuner <<zeuner>>
Both the line number and the query SQL code show that the test was ran against the file from [1]. Therefore, the "works for me" tag applies only to [1] (which is already tagged as "fixed"), but is invalid for this bug. So I have reverted the change.
[1] http://sourceforge.net/tracker/?func=detail&aid=2942744&group_id=56967&atid=482468
## Comment 13171
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2980292 at http://sourceforge.net/support/tracker.php?aid=2980292
## Comment 16197
Date: 2011-09-08 13:12:32 +0200
From: @sjoerdmullender
Created attachment 70
C source to reproduce the described issue
Re-added attachment that was lost in the conversion from Sourceforge to bugzilla.
> Attached file: [mapi_fail.c](https://github.com/MonetDB/monetdb-issues-attachments/blob/master/issue_2346_mapi_fail.c_70) (text/plain, 9465 bytes)
> Description: C source to reproduce the described issue
## Comment 16198
Date: 2011-09-08 14:00:19 +0200
From: @sjoerdmullender
Changeset [8a8d5e7e9df6](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=8a8d5e7e9df6) made by Sjoerd Mullender <sjoerd@acm.org> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=8a8d5e7e9df6](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=8a8d5e7e9df6)
Changeset description:
Test for bug #2346.
The bug was constructed using the attachment to the bug report. It
fails as reported on the Aug2011 branch.
The stable outputs may have to be adjusted when the bug is fixed.
## Comment 16337
Date: 2011-09-28 12:11:08 +0200
From: @njnes
Changeset [bb23e82b4daf](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=bb23e82b4daf) made by Niels Nes <niels@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=bb23e82b4daf](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=bb23e82b4daf)
Changeset description:
fix bug #2346, ie return properly aligned head (ie head of input)
## Comment 16338
Date: 2011-09-28 12:11:41 +0200
From: @njnes
fixed, the mkey functions now properly return aligned heads
| mkey.bulk_rotate_xor_hash MAL exception | https://api.github.com/repos/MonetDB/MonetDB/issues/2346/comments | 0 | 2020-11-30T10:38:49Z | 2024-06-27T11:44:18Z | https://github.com/MonetDB/MonetDB/issues/2346 | 753,362,044 | 2,346 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-24 02:47:53 +0100
From: @skinkie
To: GDK devs <<bugs-common>>
Version: -- development
CC: @sjoerdmullender
Last updated: 2011-09-08 13:10:39 +0200
## Comment 13158
Date: 2010-03-24 14:47:53 +0100
From: @skinkie
I have created a table on a 32bits system having a boolean colum consisting of 268435456 true values and one false value.
create table test2 (mytest boolean);
Then I executed the following query:
select * from test2 where mytest = false;
This resulted in:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb198eb70 (LWP 514)]
0xb7135e46 in seqscan_eq_chr_void_tloc_oid_vid (b=0x8066550, bn=0x8931e58,
tl=0x88f01d4, th=0x88f01d4, str_idx=0)
at ../../../src/gdk//gdk_scanselect_defs.mx:119
119 ptr v = BUN@3(bi, p);
(gdb) bt
0 0xb7135e46 in seqscan_eq_chr_void_tloc_oid_vid (b=0x8066550, bn=0x8931e58,
tl=0x88f01d4, th=0x88f01d4, str_idx=0)
at ../../../src/gdk//gdk_scanselect_defs.mx:119
1 0xb76d4097 in BAT_scanselect (b=0x8066550, bn=0x8931e58, tl=0x88f01d4,
th=0x88f01d4, li=1 '\001', hi=1 '\001', equi=1, nequi=0, lval=1, hval=1,
nocheck=0) at ../../../src/gdk/gdk_scanselect.mx:347
2 0xb773b057 in BAT_select_ (b=0x8066550, tl=0x88f01d4, th=0x88f01d4,
li=1 '\001', hi=1 '\001', tail=0 '\000', anti=0 '\000',
preserve_order=0 '\000') at ../../../src/gdk/gdk_batop.mx:1229
3 0xb773b9d3 in BATuselect (b=0x8066550, h=0x88f01d4, t=0x0)
at ../../../src/gdk/gdk_batop.mx:1306
4 0xb2ea5a7a in CMDuselect1 (result=0xb198bdc4, b=0x8066550, value=0x88f01d4)
at ../../../../src/modules/kernel/algebra.mx:1434
5 0xb2eb1412 in ALGuselect1 (result=0x88f0254, bid=0x88f0204,
value=0x88f01d4) at ../../../../src/modules/kernel/algebra.mx:2544
6 0xb7f34a58 in DFLOWstep (t=0x88d96e8, fs=0xb1a8c7e0)
at ../../../src/mal/mal_interpreter.mx:1862
7 0xb7f3ab06 in runDFLOWworker (t=0x88d96e8)
at ../../../src/mal/mal_interpreter.mx:1094
8 0xb7bc573c in wrapper_routine (data=0xb1a8c79c)
at ../../../src/gdk/gdk_posix.mx:1013
9 0xb70728ff in start_thread () from /lib/libpthread.so.0
10 0xb6df738e in clone () from /lib/libc.so.6
## Comment 13159
Date: 2010-03-24 15:14:16 +0100
From: @drstmane
what are the values of bi & p at the time of the crash?
## Comment 13160
Date: 2010-03-24 15:17:59 +0100
From: @drstmane
likewise for b & q
## Comment 13161
Date: 2010-03-24 17:07:39 +0100
From: @skinkie
Observation;
It seems to work on 'new' start sometimes. But when I insert an extra row on the end it breaks again. Could it be an out of memory situation where a malloc fails?
[Switching to Thread 0xb198eb70 (LWP 548)]
0xb7135e46 in seqscan_eq_chr_void_tloc_oid_vid (b=0x8066550, bn=0x892a280,
tl=0x892a5bc, th=0x892a5bc, str_idx=0)
at ../../../src/gdk//gdk_scanselect_defs.mx:119
119 ptr v = BUN@3(bi, p);
(gdb) print b
$1 = (BAT *) 0x8066550
(gdb) print p
$2 = 94765057
(gdb) print *b
$3 = {batCacheid = 0, H = 0x0, T = 0x0, P = 0x0, U = 0x0}
(gdb) print bi
$4 = {b = 0x8066550, hvid = 0, tvid = 0}
## Comment 13162
Date: 2010-03-24 17:08:18 +0100
From: @skinkie
Missed q;
(gdb) print q
$6 = 268435458
## Comment 13163
Date: 2010-03-24 17:31:50 +0100
From: @drstmane
Yet another race condition: the BAT has suddenly disappeared after we have scanned 94765057 of the 268435458 BUNs ...
We'd need to know what the other threads are during at time of (or actually just before) the crash, in particular whether any is just busy with BBPtrim(), GDKvmtrim(), or MT_mmap_trim() and/or running madvise() or munmap() on (any heap of) that BAT ...
## Comment 13164
Date: 2010-03-24 17:33:33 +0100
From: @skinkie
Hellgrind it is then?
## Comment 13165
Date: 2010-04-15 14:49:16 +0200
From: @sjoerdmullender
thread apply all bt
## Comment 13166
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2975903 at http://sourceforge.net/support/tracker.php?aid=2975903
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 14964
Date: 2010-10-04 09:25:47 +0200
From: @sjoerdmullender
I can't reproduce this in the Oct2010 branch. Is this still an issue?
## Comment 16196
Date: 2011-09-08 13:10:39 +0200
From: @sjoerdmullender
Problem seems to be fixed.
| sequencial scan, segv | https://api.github.com/repos/MonetDB/MonetDB/issues/2345/comments | 0 | 2020-11-30T10:38:46Z | 2024-06-27T11:44:17Z | https://github.com/MonetDB/MonetDB/issues/2345 | 753,362,013 | 2,345 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-24 12:50:53 +0100
From: @skinkie
To: GDK devs <<bugs-common>>
Version: -- development
CC: y.kargin
Last updated: 2012-11-27 15:56:32 +0100
## Comment 13153
Date: 2010-03-24 12:50:53 +0100
From: @skinkie
Compile a MonetDB5 server from CVS, using a 32bit architecture.
Create a table containing too many values;
create table test (mytest integer);
insert into test values (1);
insert into test select * from test;
...
insert into test select * from test;
(that ought be 27 times to finally end up:)
sql>insert into test select * from test;
268435456 affected rows
sql>insert into test select * from test;
MALException:!ERROR: HEAPextend: failed to extend to 2147483648 for 07/755tail
Now do it *again*.
Server console reports back:
0xb78c5fef in ATOMcmp (t=6, l=0x0, r=0x4) at ../../../src/gdk/gdk_atoms.mx:989
989 return simple_CMP(l, r, int);
(gdb) bt
0 0xb78c5fef in ATOMcmp (t=6, l=0x0, r=0x4)
at ../../../src/gdk/gdk_atoms.mx:989
1 0xb7aa6433 in BATpropcheck (b=0x806627c, mode=3)
at ../../../src/gdk/gdk_bat.mx:3192
2 0xb7f33097 in DFLOWstep (t=0x88c7d28, fs=0xb1c8e870)
at ../../../src/mal/mal_interpreter.mx:1835
3 0xb7f3ab06 in runDFLOWworker (t=0x88c7d28)
at ../../../src/mal/mal_interpreter.mx:1094
4 0xb7bc573c in wrapper_routine (data=0xb1c8e79c)
at ../../../src/gdk/gdk_posix.mx:1013
5 0xb70728ff in start_thread () from /lib/libpthread.so.0
6 0xb6df738e in clone () from /lib/libc.so.6
## Comment 13154
Date: 2010-03-24 13:18:46 +0100
From: @skinkie
While still trying to reproduce the first segmentation fault I received before the above came up when trying to reproduce it. Some other interesting errors come up if one starts with for example 6 inserts of 1.
sql>insert into test select * from test;
201326592 affected rows
sql>insert into test select * from test;
MALException:!ERROR: HEAPalloc: Insufficient space for HEAP of 3758096384 bytes.
ERROR: BATpropcheck: BAT tmpr_755(-493)[int,void] with 402653184 tuples was incorrectly marked sorted!
ERROR: BATpropcheck: BAT tmpr_755(-493)[int,void] with 402653184 tuples remains marked radix-clustered on 32 bits; not checked!
My first experiment did the above, without swap enabled to exhaust memory for an experiment. And ended up in a segmentation fault after a 'trying to free up space'. While with swap disabled in this comment I do get:
GDKmmap(3758096384) fails, try to free up space [memory in use=8561584,virtual
memory in use=12124160]
GDKmmap(3758096384) result [mem=8402928,vm=12124160]
It doesn't segv at all;
sql>insert into test select * from test;
MALException:!ERROR: HEAPalloc: Insufficient space for HEAP of 3758096384 bytes.
sql>insert into test select * from test;
MALException:!ERROR: HEAPalloc: Insufficient space for HEAP of 3758096384 bytes.
sql>insert into test select * from test;
MALException:!ERROR: HEAPalloc: Insufficient space for HEAP of 3758096384 bytes.
sql>insert into test select * from test;
I forgot how many inserts of one I did at the first time.
## Comment 13155
Date: 2010-03-24 15:15:27 +0100
From: @drstmane
What are the values of l & r at the time of the first reported crash?
## Comment 13156
Date: 2010-03-24 17:39:57 +0100
From: @skinkie
l & r were reported above... but when trying to reproduce the error I get another one again:
mserver5: ../../../../src/backends/monet5/sql_optimizer.mx:370: SQLgetStatistics: Assertion `cnt <= (size_t) ((oid) 2147483647)' failed.
Program received signal SIGABRT, Aborted.
[Switching to Thread 0xb1b90b70 (LWP 566)]
0xb7fe1424 in __kernel_vsyscall ()
## Comment 13157
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2975836 at http://sourceforge.net/support/tracker.php?aid=2975836
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 18137
Date: 2012-11-27 15:56:32 +0100
From: Yagiz <<y.kargin>>
This one is really time- and resource consuming to reproduce. There have been a lot of changes to the code base, so it might even be already fixed.
| Too many values in table, segv | https://api.github.com/repos/MonetDB/MonetDB/issues/2344/comments | 0 | 2020-11-30T10:38:43Z | 2024-06-28T07:25:41Z | https://github.com/MonetDB/MonetDB/issues/2344 | 753,361,969 | 2,344 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-23 08:47:17 +0100
From: Stephen Benz <<sbenz>>
To: clients devs <<bugs-clients>>
Version: -- development
Last updated: 2010-07-09 13:49:53 +0200
## Comment 13149
Date: 2010-03-23 20:47:17 +0100
From: Stephen Benz <<sbenz>>
mapi_read in the native php socket client assumes socket_read is going to actually return the chunk size that's requested. While this does happen in many cases, when there's a large result set or network latency only a portion of that read may happen. This causes a limited number of rows to be returned, and other weird behavior.
Environment:
Linux, CentOS 5
PHP 5.1.6
MonetDB Nov 2009 (but verified the php code hasn't been fixed in Feb 2010)
Reproduction steps:
1. $res = monetdb_query("select * from largeTable;"); // where largeTable is a table with 500k+ rows, in our case
2. Notice the result doesn't contain all the rows... the number is variable based on the connection to the db server (we run monetdb on another server, *not localhost*)
Fix:
in php_mapi.inc (function mapi_read()):
$data = socket_read($socket, $chunk_size);
// can't assume the read is going to give you the actual size you expect
while(strlen($data) < $chunk_size)
{
$data .= socket_read($socket, $chunk_size - strlen($data));
}
and a few lines later:
if ( ($block = socket_read($socket, $chunk_size)) == FALSE) {
$last_error = socket_strerror(socket_last_error());
return FALSE;
}
while(strlen($block) < $chunk_size)
{
$block .= socket_read($socket, $chunk_size - strlen($block));
}
## Comment 13150
Date: 2010-04-15 15:04:37 +0200
From: @sjoerdmullender
Thanks for the report and the fix.
I checked in a variant of your fix. This should make it into the next release.
Can you check whether my version works?
## Comment 13151
Date: 2010-04-15 16:44:15 +0200
From: Stephen Benz <<sbenz>>
Yup! Just tested the new version from CVS and it fixes the issue. Thanks!
## Comment 13152
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2975433 at http://sourceforge.net/support/tracker.php?aid=2975433
## Comment 14313
Date: 2010-07-09 13:49:53 +0200
From: @sjoerdmullender
Not adding a test since it requires a larg(ish) database and a slow connection, neither of which is practical in the testweb setting.
| PHP Native client occasionally truncates large result sets | https://api.github.com/repos/MonetDB/MonetDB/issues/2343/comments | 0 | 2020-11-30T10:38:40Z | 2024-06-27T11:44:15Z | https://github.com/MonetDB/MonetDB/issues/2343 | 753,361,933 | 2,343 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-23 05:28:35 +0100
From: Mark Bucciarelli <<mkbucc>>
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:18:33 +0200
## Comment 13144
Date: 2010-03-23 05:28:35 +0100
From: Mark Bucciarelli <<mkbucc>>
Very similar to last bug. Slight change to test
file---I don't drop the column, just the constraint
and the index.
The database will start, but when I try to connect,
mclient says "Connection terminated" and the
merovingian log lists the assertion failure in the
bug report subject line.
Test output:
$ ./test.sh
Creating Monet database tst
create: database 'tst' already exists
starting database 'tst'... done
release: database 'tst' is not under maintenance
Connection terminated
$ tail -3 /var/log/merovingian.log
2010-03-23 00:36:42 ERR tst[27799]: mserver5: bat_storage.mx:822: dup_delta: Assertion `bat->ibid' failed.
2010-03-23 00:36:42 MSG merovingian[28274]: client localhost:55188 has disconnected from proxy
2010-03-23 00:36:42 MSG merovingian[28274]: database 'tst' (27799) was killed by signal 6
$
Test inputs:
$ cat test.sql
--
-- Setup two test tables, one that references the other.
--
CREATE TABLE "tst"."t1" (
"id" int NOT NULL,
CONSTRAINT "t1_id_pkey" PRIMARY KEY ("id")
);
CREATE TABLE "tst"."t2" (
"id" int NOT NULL,
"t1_id" int NOT NULL,
CONSTRAINT "t2_id_pkey" PRIMARY KEY ("id"),
CONSTRAINT "t2_t1_id_fkey" FOREIGN KEY ("t1_id") REFERENCES "tst"."t1" ("id")
);
CREATE INDEX "t2_t1_id" ON "tst"."t2" ("t1_id");
--
-- Drop fkey constraint and index
--
DROP INDEX "tst"."t2_t1_id";
ALTER TABLE "tst"."t2" DROP CONSTRAINT t2_t1_id_fkey;
--
-- Drop fkey column that links tables.
--
-- ALTER TABLE "tst"."t2" DROP COLUMN t1_id;
$ cat test.sh
! /bin/sh -e
Create aoflash database in MonetDB.
user=monetdb
pass=monetdb
db=tst
schema=tst
Note: doesn't abort if database already exists.
echo "Creating Monet database ${db}"
sudo monetdb create ${db} || exit 1
sudo monetdb start ${db} || exit 1
sudo monetdb release ${db}
export DOTMONETDBFILE="./.testmonetdb"
cat > ${DOTMONETDBFILE} << EOF
user=${user}
password=${pass}
language=sql
EOF
sql="CREATE SCHEMA "${schema}" AUTHORIZATION "${user}";"
mclient -d ${db} -s "${sql}"
mclient -d ${db} < test.sql
sudo monetdb stop tst
sudo monetdb start tst
mclient -d tst
sudo monetdb stop tst && sudo monetdb destroy tst
$
## Comment 13145
Date: 2010-03-23 05:31:23 +0100
From: Mark Bucciarelli <<mkbucc>>
I didn't quite post the exact test output. (I had already created
the database.) Here is the test output after destroying the db
then running test.sh
$ ./test.sh
Creating Monet database tst
created database in maintenance mode: tst
starting database 'tst'... done
taken database out of maintenance mode: tst
stopping database 'tst'... done
starting database 'tst'... done
Connection terminated
$ tail -3 /var/log/merovingian.log
2010-03-23 00:40:04 ERR tst[27909]: mserver5: bat_storage.mx:822: dup_delta: Assertion `bat->ibid' failed.
2010-03-23 00:40:04 MSG merovingian[28274]: client localhost:34451 has disconnected from proxy
2010-03-23 00:40:04 MSG merovingian[28274]: database 'tst' (27909) was killed by signal 6
$
## Comment 13146
Date: 2010-03-23 05:32:53 +0100
From: Mark Bucciarelli <<mkbucc>>
Note:
After commenting out the line
ALTER TABLE "tst"."t2" DROP CONSTRAINT t2_t1_id_fkey;
in the SQL, I can now connect.
## Comment 13147
Date: 2010-04-15 16:17:40 +0200
From: @sjoerdmullender
I can reproduce the error.
Test created in src/BugTracker-2010/alter-table-restart-crash.SF-2975022.
## Comment 13148
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2975022 at http://sourceforge.net/support/tracker.php?aid=2975022
## Comment 14271
Date: 2010-07-09 00:26:00 +0200
From: @njnes
Tests run fine
## Comment 14829
Date: 2010-08-30 09:18:33 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| bat_storage.mx:822: dup_delta: Assertion `bat->ibid' failed. | https://api.github.com/repos/MonetDB/MonetDB/issues/2342/comments | 0 | 2020-11-30T10:38:36Z | 2024-06-27T11:44:14Z | https://github.com/MonetDB/MonetDB/issues/2342 | 753,361,897 | 2,342 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-23 05:05:14 +0100
From: Mark Bucciarelli <<mkbucc>>
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:18:31 +0200
## Comment 13139
Date: 2010-03-23 05:05:14 +0100
From: Mark Bucciarelli <<mkbucc>>
The attached very simple sequence of SQL commands
generates an assertion failure in load_idxcolumn.
This is with Feb2010, 32bit i386 Ubuntu (4.3.3 gcc).
Test output:
$ ./test.sh
Creating Monet database tst
created database in maintenance mode: tst
starting database 'tst'... done
taken database out of maintenance mode: tst
stopping database 'tst'... done
starting database 'tst'... FAILED
start: starting 'tst' failed: database 'tst' has crashed after starting, manual intervention needed, check merovingian's logfile for details
$ tail -3 /var/log/merovingian.log
2010-03-23 00:09:46 ERR tst[27143]: mserver5: store.mx:373: load_idxcolumn: Assertion `kc->c' failed.
2010-03-23 00:09:46 MSG merovingian[28274]: database 'tst' (27143) was killed by signal 6
2010-03-23 00:09:56 ERR control[28274]: (local): failed to fork mserver: database 'tst' has crashed after starting, manual intervention needed, check merovingian's logfile for details
$
Test files:
$ cat test.sql
--
-- Setup two test tables, one that references the other.
--
CREATE TABLE "tst"."t1" (
"id" int NOT NULL,
CONSTRAINT "t1_id_pkey" PRIMARY KEY ("id")
);
CREATE TABLE "tst"."t2" (
"id" int NOT NULL,
"t1_id" int NOT NULL,
CONSTRAINT "t2_id_pkey" PRIMARY KEY ("id"),
CONSTRAINT "t2_t1_id_fkey" FOREIGN KEY ("t1_id") REFERENCES "tst"."t1" ("id")
);
CREATE INDEX "t2_t1_id" ON "tst"."t2" ("t1_id");
--
-- Drop fkey constraint and index
--
DROP INDEX "tst"."t2_t1_id";
ALTER TABLE "tst"."t2" DROP CONSTRAINT t2_t1_id_fkey;
--
-- Drop fkey column that links tables.
--
ALTER TABLE "tst"."t2" DROP COLUMN t1_id;
$ cat test.sh
! /bin/sh -e
Create aoflash database in MonetDB.
user=monetdb
pass=monetdb
db=tst
schema=tst
Note: doesn't abort if database already exists.
echo "Creating Monet database ${db}"
sudo monetdb create ${db} || exit 1
sudo monetdb start ${db} || exit 1
sudo monetdb release ${db}
export DOTMONETDBFILE="./.testmonetdb"
cat > ${DOTMONETDBFILE} << EOF
user=${user}
password=${pass}
language=sql
EOF
sql="CREATE SCHEMA "${schema}" AUTHORIZATION "${user}";"
mclient -d ${db} -s "${sql}"
mclient -d ${db} < test.sql
sudo monetdb stop tst
sudo monetdb start tst
sudo monetdb stop tst && sudo monetdb destroy tst
$
## Comment 13140
Date: 2010-03-23 05:08:52 +0100
From: Mark Bucciarelli <<mkbucc>>
Similar issue reported in 2004 (949164) with
no test to reproduce.
On my setup, the above scripts reproduce the
problem every time.
## Comment 13141
Date: 2010-04-15 15:49:45 +0200
From: @sjoerdmullender
I can reproduce the failure.
Test added: src/test/BugTracker-2010 alter-table-restart-crash.SF-2975018-{1,2}
the first file sets up the crash, the second can't be executed since the server crashes during initialization.
## Comment 13142
Date: 2010-04-15 16:18:51 +0200
From: @sjoerdmullender
I moved the test to a separate directory: src/test/BugTracker-2010/alter-table-restart-crash.SF-2975018.
## Comment 13143
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2975018 at http://sourceforge.net/support/tracker.php?aid=2975018
## Comment 14272
Date: 2010-07-09 00:26:52 +0200
From: @njnes
the test runs fine
## Comment 14822
Date: 2010-08-30 09:18:31 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| load_idxcolumn: Assertion `kc->c' failed | https://api.github.com/repos/MonetDB/MonetDB/issues/2341/comments | 0 | 2020-11-30T10:38:33Z | 2024-06-27T11:44:13Z | https://github.com/MonetDB/MonetDB/issues/2341 | 753,361,855 | 2,341 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-22 08:11:32 +0100
From: bill xin <<oursql>>
To: GDK devs <<bugs-common>>
Version: -- development
CC: charnik
Last updated: 2011-05-20 11:36:35 +0200
## Comment 13137
Date: 2010-03-22 08:11:32 +0100
From: bill xin <<oursql>>
There is a bug in monetdb, you can run the following script in two clients(mclient.exe), and before test, please disable autocommit option.
user A: USER B:
CREATE TABLE TEST(C1 INT PRIMARY KEY, C2 INT);
INSERT INTO TEST VALUES(1,1);
COMMIT;
SELECT C2 FROM TEST WHERE C1=1;
SELECT C2 FROM TEST WHERE C1=1;
UPDATE TEST SET C2=10 WHERE C1=1;
UPDATE TEST SET C2=30 WHERE C1=1;
COMMIT;
COMMIT;--CONFLICT, NEED TO ROLLBACK;
ROLLBACK;
UPDATE TEST SET C2=30 WHERE C1=1;
SELECT C2 FROM TEST WHERE C1=1;
--THE RESULT IS 10, NOT 30. TOO BAD!
## Comment 13138
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2974381 at http://sourceforge.net/support/tracker.php?aid=2974381
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 15827
Date: 2011-05-20 11:33:46 +0200
From: Babis <<charnik>>
Maybe this bug has to be closed. Tested on Mar2011 and the problem is not valid anymore.
## Comment 15828
Date: 2011-05-20 11:36:35 +0200
From: @grobian
ok, thanks
| concurrency control bug | https://api.github.com/repos/MonetDB/MonetDB/issues/2340/comments | 0 | 2020-11-30T10:38:30Z | 2024-06-28T07:25:40Z | https://github.com/MonetDB/MonetDB/issues/2340 | 753,361,817 | 2,340 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-17 09:04:43 +0100
From: @drstmane
To: SQL devs <<bugs-sql>>
Version: -- development
Last updated: 2010-10-18 19:20:25 +0200
## Comment 13134
Date: 2010-03-17 21:04:43 +0100
From: @drstmane
Since Feb 25 or Feb 26 2010, i.e., after checkins on Feb 24 and/or Feb 25 2010, some of the Via-M tests produce different output than before; cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/event_s04.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/event_s07.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/event_s08.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/event_s10.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/event_s11.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/event_s12.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_Via-m/file_s16.out.00.html
Could you please check, whether these are crucial differences, or possible acceptable differences in (e.g.) "internal" IDs (due to implementation changes) that are should hence be approved?
Thanks!
## Comment 13135
Date: 2010-03-18 10:51:58 +0100
From: @arjenderijke
Deja vu: See also SQL: various VIA-M tests fail - ID: 2893859 The problem seems to be that the way a failed insert in a fuction is handled has changed again. So the the differences are crucial.
We should discuss this offline.
## Comment 13136
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2972162 at http://sourceforge.net/support/tracker.php?aid=2972162
## Comment 15059
Date: 2010-10-18 19:20:25 +0200
From: @drstmane
All Via-M test seem to work now (for quite a while) in both Jun2010 & Oct2010;
hence, I consider this issue fixed (for now... ?).
| SQL: Via-M tests fail | https://api.github.com/repos/MonetDB/MonetDB/issues/2339/comments | 0 | 2020-11-30T10:38:27Z | 2024-06-27T11:44:11Z | https://github.com/MonetDB/MonetDB/issues/2339 | 753,361,784 | 2,339 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-16 09:22:13 +0100
From: Mark Bucciarelli <<mkbucc>>
To: @mlkersten
Version: -- development
Last updated: 2010-04-18 05:49:05 +0200
## Comment 13127
Date: 2010-03-16 21:22:13 +0100
From: Mark Bucciarelli <<mkbucc>>
Feb2010 CVS.
OpenBSD 4.6
gcc 3.3.5 (propolice)
Trying to start database, the mserver5 dumps core.
At line 92 of mal_atom.c, code tries to dereference null pointer:
case 'n':
if( idcmp("null",name)==0 && pci->argc==1){
-----> ptr atmnull = ((ptr (*)(void)) pci->fcn) ();
## Comment 13128
Date: 2010-03-16 23:44:57 +0100
From: Mark Bucciarelli <<mkbucc>>
Looks like a loader path issue.
But shouldn't the server fail earlier and harder?
Can the server ever function normally if the mdb (or blob, or calc, or ...)
library is not found?
2010-03-16 17:45:29 MSG test[3349]: MonetDB server v5.18.2, based on kernel v1.36.2
2010-03-16 17:45:29 MSG test[3349]: Serving database 'test', using 1 thread
2010-03-16 17:45:29 MSG test[3349]: Compiled for i386-unknown-openbsd4.6/32bit with 32bit OIDs dynamically linke
d
2010-03-16 17:45:29 MSG test[3349]: Found 191.590 MiB available main-memory.
2010-03-16 17:45:29 MSG test[3349]: Copyright (c) 1993-July 2008 CWI.
2010-03-16 17:45:29 MSG test[3349]: Copyright (c) August 2008-2010 MonetDB B.V., all rights reserved
2010-03-16 17:45:29 MSG test[3349]: Visit http://monetdb.cwi.nl/ for further information
2010-03-16 17:45:29 MSG test[3349]: WARNING: LoaderException:loadLibrary:File not found:mdb
2010-03-16 17:45:29 MSG test[3349]: WARNING: LoaderException:loadLibrary:File not found:unix
etc.
## Comment 13129
Date: 2010-03-17 00:57:17 +0100
From: @mlkersten
Indeed, the server should come to a graceful stop. More defensive code is added.
## Comment 13130
Date: 2010-04-15 16:20:41 +0200
From: @sjoerdmullender
Is this still open?
Do we need a test?
## Comment 13131
Date: 2010-04-15 20:37:29 +0200
From: Mark Bucciarelli <<mkbucc>>
short answer: yes this is stilll open; I'm not
sure if it needs a test.
The library loading code first calls a locate file
routine that uses open() with a specific
filename, which is more restrictive than a
dlopen() with the same file name.
the patch depends on if you consider this
an os-specific case or not. (There is already
ifdef AIX logic in this code path, so it looks
like something similar has come up before.)
My current thought is to delete the open() code
path, and just use dlopen(). I can't think of any
value the open() call adds. I plan to build a patch
for this.
## Comment 13132
Date: 2010-04-18 17:49:05 +0200
From: @mlkersten
no test needed. It is part of general defensive code.
## Comment 13133
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2971516 at http://sourceforge.net/support/tracker.php?aid=2971516
| pci->fcn = 0 in malAtomProperty: segfault | https://api.github.com/repos/MonetDB/MonetDB/issues/2338/comments | 0 | 2020-11-30T10:38:24Z | 2024-06-27T11:44:10Z | https://github.com/MonetDB/MonetDB/issues/2338 | 753,361,755 | 2,338 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-15 07:35:05 +0100
From: Hering Cheng <<heringcheng>>
To: GDK devs <<bugs-common>>
Version: -- development
CC: @mlkersten, sellam
Last updated: 2015-11-11 09:19:57 +0100
## Comment 13120
Date: 2010-03-15 19:35:05 +0100
From: Hering Cheng <<heringcheng>>
Can MonetDB handle terabytes of data? I dimly understand that MonetDB maps data in each column into memory, which is partly responsible for its performance. However, what if the table is so large that all the data in a column do not fit in RAM + swap on the server? I am currently loading such large data set into MonetDB and I can see that the mserver5 process is nearing the physical limit on my server. Here is from "top" command on my SPARC/Solaris:
PID USERNAME LWP PRI NICE SIZE RES STATE TIME CPU COMMAND
26004 chenher 39 60 0 219G 28G sleep 648:22 0.32% mserver5
Here, total size is 219 GB and resident size is 28 GB. Will it run out of memory? Is there a way to limit how much memory mserver5 consumes?
## Comment 13121
Date: 2010-03-16 15:14:21 +0100
From: Hering Cheng <<heringcheng>>
So mserver5 eventually core dumped. Unfortunately, I did not build my SPARC binary with debugging info:
t@17 (l@17) terminated by signal SEGV (no mapping at the fault address)
0xffffffff7ec5401c: BAThashjoin+0x19c4: stx %o0, [%l0]
(dbx) where
current thread: t@17
=>[1] BAThashjoin(0xffffffff698f9dc0, 0x123099bc8, 0xffffffff7ef0a570, 0x123099a10, 0x112bff9c8, 0xffffffff7ede1f78), at 0xffffffff7ec5401c
[2] BATjoin(0x112bff9c8, 0x123099a10, 0x1000000, 0x0, 0x2636bc, 0xffffffff76014c68), at 0xffffffff7ec7e8fc
[3] ALGjoinestimate(0x1194f48e0, 0xffffffff7582caf8, 0x1194f48d0, 0x8000000000000000, 0x10e924, 0x112bff9c8), at 0xffffffff7581e97c
[4] ALGjoin(0x8000000000000000, 0xffffffff7ede1f70, 0x40, 0x0, 0x10e7ac, 0xffffffff7592d1f8), at 0xffffffff7581ea80
[5] DFLOWstep(0x1194f48e0, 0xffffffff698fa0d0, 0xffffffff698fa130, 0xffffffff7f372ac8, 0x115, 0x1001069c8), at 0xffffffff7f230344
[6] runDFLOWworker(0x1012ceb98, 0xffffffff32af7198, 0x10123fae8, 0x0, 0x1, 0xffffffff7f26d6a8), at 0xffffffff7f2338a4
Worse yet, I now cannot even restart it. How do I delete a database?
2010-03-16 07:02:33 ERR control[1316]: rdcuxsrv220:58360: failed to fork mserver: database 'dent' has inconsistent state (running but dead), review merovingian''s logfile for any peculiarities
2010-03-16 07:02:40 MSG merovingian[1316]: database 'dent' (16384) has crashed (dumped core)
t@1 (l@1) terminated by signal SEGV (no mapping at the fault address)
0xffffffff6c2c2220: load_key+0x02a0: ld [%o0], %o1
(dbx) where
current thread: t@1
=>[1] load_key(0x10024fc28, 0x100d58158, 0x100258328, 0x100c30aa8, 0x100d5e208, 0xffffffff6c4096d0), at 0xffffffff6c2c2220
[2] load_table(0x10024fc28, 0x100d0af78, 0x100d581c8, 0xffffffff7ede1f78, 0x100d58164, 0x1002582c8), at 0xffffffff6c2c32dc
[3] load_schema(0x10024fc28, 0x100d0af84, 0x100d0af98, 0xffffffff6c455610, 0x100d0af78, 0x100250f68), at 0xffffffff6c2c4100
[4] load_trans(0x10024fc28, 0x4ba, 0x10024fc50, 0x100274598, 0xffffffff6c2f78d8, 0x8000000000000000), at 0xffffffff6c2c443c
[5] store_init(0x0, 0x0, 0xffffffff6c4121c0, 0x4ba, 0x400, 0x100105560), at 0xffffffff6c2c6708
[6] mvc_init(0x0, 0x0, 0x0, 0x10016a360, 0x17d4ec, 0xffffffff6c2f4668), at 0xffffffff6c28c26c
[7] SQLinit(0x0, 0xffffffff6c455658, 0xffffffff6c4120a8, 0xffffffff6c4096d0, 0xffffffff7ef0a4d8, 0x0), at 0xffffffff6c252254
[8] SQLprelude(0x0, 0xffffffff6c2ec4a0, 0x290, 0xffffffff6c4120b0, 0xffffffff7f376b80, 0xffffffff6c2ec4b0), at 0xffffffff6c251f24
[9] runMALsequence(0x1001062c0, 0x100190dc8, 0x6974, 0xffffffff7f26d5b0, 0xffffffff7fffd4a0, 0x10024f758), at 0xffffffff7f223ae0
[10] runMAL(0x1001062c0, 0x2, 0x10, 0xffffffff7fffd518, 0x0, 0x1), at 0xffffffff7f21c4b8
[11] MALengine(0x0, 0x1001062c0, 0x100190dc8, 0x0, 0xffffffff7f372ac8, 0x100117178), at 0xffffffff7f21b74c
[12] callString(0x1001062c0, 0xffffffff7f26cf88, 0xffffffff7f21b4b4, 0x0, 0x1008b14e8, 0xffffffff7f372ac8), at 0xffffffff7f219fa0
[13] main(0x100002000, 0x0, 0x10010fb7e, 0x100105660, 0x10010f000, 0x100105e70), at 0x100003090
## Comment 13122
Date: 2010-03-16 17:56:55 +0100
From: @grobian
> Worse yet, I now cannot even restart it. How do I delete a database?
monetdb destroy <database>
## Comment 13123
Date: 2010-03-17 14:38:42 +0100
From: @sjoerdmullender
MonetDB can definitely handle terabytes of data. However, since all data that a query uses is memory mapped, you need a coorespondingly large address space. In other words, a 64 bit architecture and a 64 bit build of MonetDB are essential.
MonetDB memory maps large files. This means that under certain conditions MonetDB does not use your swap space. However, as soon as you start querying, intermediate results are created, and they do occupy real memory and real swap space. Because MonetDB does everything in bulk, those intermediate results may be large as well.
In other words, to work effectively with such large databases, you need enough memory and swap space (the memory is needed to avoid thrashing).
## Comment 13124
Date: 2010-03-17 16:37:04 +0100
From: Hering Cheng <<heringcheng>>
Thank you for the explanation.
In my case, I am using a 64-bit build on 64-bit architecture (Solaris SPARC). From the sound of it, the combined RAM and swap space should be large enough to accommodate the _entire_ data set _plus_ intermediate result. Is my understanding correct?
After re-creating the database, I am finally able to load the data in without errors and perform complex queries. The trick to avoid the crash seems to be in adding constraints on the tables after the load.
After that I shut it down with "monetdb stop mydb" and tried to restart it with "monetdb start mydb". Unfortunately, mserver5 core dumps. Did I use the correct shutdown/restart command?
## Comment 13125
Date: 2010-03-19 17:04:44 +0100
From: @drstmane
See also thread "Virtual Memory Requirement" on the monetdb-users mailing list:
http://sourceforge.net/mailarchive/forum.php?thread_name=1e12cdc91003190718h3b27b3f9r899fbb493f17ae69%40mail.gmail.com&forum_name=monetdb-users
## Comment 13126
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2970780 at http://sourceforge.net/support/tracker.php?aid=2970780
## Comment 21505
Date: 2015-11-11 09:19:57 +0100
From: @mlkersten
Nothing to react upon.
| Memory Usage | https://api.github.com/repos/MonetDB/MonetDB/issues/2337/comments | 0 | 2020-11-30T10:38:20Z | 2024-06-28T07:25:40Z | https://github.com/MonetDB/MonetDB/issues/2337 | 753,361,717 | 2,337 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-14 03:40:27 +0100
From: Farid Z <<faridz>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2011-09-16 15:04:41 +0200
## Comment 13108
Date: 2010-03-14 03:40:27 +0100
From: Farid Z <<faridz>>
I am using an ANSI ODBC client (not Unicode) with latest version of MonetDB ODBC driver 1.36.01.01 2/24/2010 and I am getting clobbed resultset comparing demo database loaded with VOC dataset to itself by using Zidsoft CompareData app.
Does the ODBC driver not support ANSI clients?
## Comment 13109
Date: 2010-03-15 18:03:09 +0100
From: @sjoerdmullender
The MonetDB ODBC driver only support Unicode and UTF-8. That is to say, the wide character calls are expected to use UCS-16, and the narrow character calls are expected to use UTF-8. The driver does not support any other character sets (well, ASCII since it is a proper subset of UTF-8).
## Comment 13110
Date: 2010-03-15 18:28:28 +0100
From: Farid Z <<faridz>>
My app is ANSI so I am using narrow calls. So that's not the cause of this issue. I see the cause of the problem: the driver is returning incorrect information for ::SQLTables character resultset columns:
rc = SQLColAttribute(hstmt,
nCol,
SQL_DESC_OCTET_LENGTH,
NULL,
0,
NULL,
&pRow->buffer_length);
rc = SQLColAttribute(hstmt,
nCol,
SQL_DESC_LENGTH,
NULL,
0,
NULL,
(SQLLEN*) &pRow->column_size);
The driver is returning zero for SQL_DESC_OCTET_LENGTH and SQL_DESC_LENGTH for ::SQLTables character resultsetset columns instead of actual values.
Same thing for ::SQLPrimaryKeys, ::SQLStatistics and ::SQLForeignKeys resultsets character columns
## Comment 13111
Date: 2010-04-15 21:55:07 +0200
From: @sjoerdmullender
If by ANSI you mean one of the ISO-8859 versions (e.g. ISO-8859-1 aka "latin 1") then that is definitely a problem. Those are *not* compatible with UTF-8.
What are the types of pRow->buffer_length and pRow->column_size?
Is your application 32 bit or 64 bit?
The MonetDB ODBC driver may not be completely compliant with the 64-bit definition of the ODBC interface. I'm hoping that can be fixed in the next feature release (not in a bug fix release since it's an API change).
## Comment 13112
Date: 2010-04-15 22:42:37 +0200
From: Farid Z <<faridz>>
I am running the 32-bit version of my app on Windows XP SP3 32-bit and using Windows characters set 1252 (US-EN).
I am using ODBC types for column_size and buffer_length which in 32-bit version of my app are SQUINTEGER and SQLINTEGER.
SQLULEN column_size;
SQLLEN buffer_length;
My app is tested successfully with 30-40 different DBMS ODBC drivers without encountering this issue http://www.zidsoft.com/dbmsnotes.html
The driver is returning zero for SQL_DESC_OCTET_LENGTH and SQL_DESC_LENGTH
for ::SQLTables character resultsetset columns instead of actual values.
Same thing for ::SQLPrimaryKeys, ::SQLStatistics and ::SQLForeignKeys
resultsets character columns
## Comment 13113
Date: 2010-04-16 10:07:58 +0200
From: @sjoerdmullender
I have been able to also get the value 0 from the call to SQLColAttribute on column 1 after a call to SQLTables.
The reason for the value 0 is that (in my case, anyway) all values in the column are NULL.
If I look at column 2 where I do get a non-NULL value, the value for SQL_DESC_OCTET_LENGTH and SQL_DESC_LENGTH is indeed greater than 0.
The ODBC speciification just says that the types of the columns are "varchar" without any size indication. So if the columns are completely filled with NULLs, the server just returns a length indication of 0.
I suppose I could arbitrarily change the length to 1 when it would otherwise have been 0.
## Comment 13114
Date: 2010-04-16 14:43:07 +0200
From: Farid Z <<faridz>>
The driver needs to return valid values (not 0 and certainly usually greater than 1) for SQL_DESC_LENGTH and SQL_DESC_OCTET_LENGTH for the resultset columns as per ODBC descriptions for these.
http://msdn.microsoft.com/en-us/library/ms713558(VS.85).aspx
Quote:
SQL_DESC_LENGTH: A numeric value that is either the maximum or actual character length of a character string or binary data type. It is the maximum character length for a fixed-length data type, or the actual character length for a variable-length data type. Its value always excludes the null-termination byte that ends the character string
SQL_DESC_OCTET_LENGTH :
The length, in bytes, of a character string or binary data type. For fixed-length character or binary types, this is the actual length in bytes. For variable-length character or binary types, this is the maximum length in bytes. This value does not include the null terminator.
So for SQLTables resultset table names these would be for example, 128, 128 assuming that the maximum length for a table name in MonetDB is 128 character. Similar info should be returned by the driver for SQLStatistics, SQLPrimaryKeys and SQLForeignKesy, SQLGetTypeInfo and all other ODBC catalog functions for these attributes.
## Comment 13115
Date: 2010-04-16 15:46:14 +0200
From: @sjoerdmullender
I read that too, but I'm not sure I can come to the same conclusions. I'm struggling to understand what the specification means.
The columns SQLTables returns are all varchar, without mention how large they should be. The first column is the catalog name. The values here can be NULL if the data source does not support catalogs. Since MonetDB indeed doesn't support catalogs, the value returned is indeed NULL.
The description in SQLTables doesn't say what size varchar column to return, but presumably one that is large enough to hold the data. In this case, varchar(1) is plenty long enough (for this particular column).
For SQLColAttribute, the description of SQL_DESC_LENGTH says that, for variable-length data (which varchar is), the size returned is the "actual character length" (we can ignore the fixed-length case). What is the actual character length of a NULL value? I can argue 0.
The problem for SQL_DESC_OCTET_LENGTH is that, for variable-length character types, it returns the *maximum* length in bytes. Again, what is the maximum length of a column that is always going to be NULL? The query that SQLTables does under the hood uses cast(null as varchar(1)) for the catalog column.
There *is*, however, a problem in MonetDB. The server returns a single column length value, and for varchar columns that value is the actual length of the longest value in the column (0 if all values are NULL or if there are no rows). (The value returned is the declared length for fixed-length columns, i.e. 50 for char(50).) In other words, the client does not know the maximum allowed length of the column if the longest value is shorter. So if a column is declared varchar(50) but only contains NULLs, the value returned by the server is 0. This means that determining SQL_DESC_LENGTH is relatively easy, but determining SQL_DESC_OCTET_LENGTH is not currently possible for varchar columns.
Another problem is that SQL_DESC_LENGTH counts *characters* and SQL_DESC_OCTET_LENGTH counts *bytes*. These are not interchangeable. MonetDB uses the UTF-8 encoding, which is a variable-length encoding. A character can take from 1 to 6 bytes. Should we then multiply the value by 6? If a column is declared as e.g. varchar(10), that means that it can store up to 10 characters in MonetDB.
The problem we're having here doesn't really have anything to do with SQLTables as such. The implementation does a SQLExecDirect call under the hood, and SQLColAttribute doesn't know (or care) how the result set was created. The problem has to do with getting metainformation about the result set. In other words, your suggestion of using 128 characters (or is that bytes?) doesn't help. That information is not passed on to SQLColAttribute, and anyway, the same problems would exist for hand-created SQL queries.
I do agree that all metainformation-returning queries should return compatible information.
## Comment 13116
Date: 2010-04-16 16:16:28 +0200
From: Farid Z <<faridz>>
>The columns SQLTables returns are all varchar, without mention how large
>they should be
The returned info would represent how the column is stored at the dbms. That's why the ODBC spec does not specify exact character counts. For DB2 this a table name can be maximum of 128 characters so their drivers returns 128 for SQL_DESC_LENGTH and 128 for SQL_DESC_OCTET_LENGTH for ANS or 128*2 for UTF16.
You can ignore the catalog column of the resultset since MonetDB does not support catalogs and if the driver correctly reports that catalog are not supported by the DBMS.
>Another problem is that SQL_DESC_LENGTH counts *characters* and
>SQL_DESC_OCTET_LENGTH counts *bytes*. These are not interchangeable.
>MonetDB uses the UTF-8 encoding
Example column declared as varchar(30) the driver would return
SQL_DESC_LENGTH -> 30
SQL_DESC_OCTET_LENGTH -> 30 for ANSI version of the driver, 30*2 for Unicode version of the driver. For Windows the driver must translate from dbms characterset (UTF8) to client characterset: ANSI (single byte characterset) or Windows UTF16 (implemented by ODBC spec as wchar_t two bytes).
>your suggestion of using 128 characters (or is that bytes?) doesn't help
That was an example, not a suggestion. For MontDB the driver needs to return the actual info for the underlying catalog "views" where the information for SQLTables is stored. MonetDB must be storing the table names somewhere in its catalogs.
>There *is*, however, a problem in MonetDB. The server returns a single column length value, and for varchar >columns that value is the actual length of the longest value in the column (0 if all values are NULL or if
>there are no rows)
That would be a problem. If I create a column as varchar(30) then I should get (as explained above) whether or not there is ANY data in the table:
SQL_DESC_LENGTH -> 30
SQL_DESC_OCTET_LENGTH -> 30 for ANSI version of the driver, 30*2 for Unicode version of the driver
SQL_DESC_LENGTH and SQL_DESC_OCTET_LENGTH are used by the ODBC client to create and allocate buffers for the resultset data and without correct information for these attributes, the ODBC client can not correctly retrieve data from the database
## Comment 13117
Date: 2010-04-16 18:17:26 +0200
From: @sjoerdmullender
What does "the actual character length for a variable-length data type" mean? I thought it meant the length of the (longest) actual value in the column. Since this is from SQLColAttribute, there are actual values in the result set to talk about. From your comment I understand you have a different interpretation (namely the maximum allowed length for the column). Is there anywhere a clearer definition of these attributes?
The definition of SQL_DESC_OCTET_LENGTH is perhaps a little clearer, although I could interpret that either way as well: "For variable-length character or binary types, this is the maximum length in bytes". Which maximum? The maximum according to the column definition or the maximum in the current result set?
In addition to SQL_DESC_LENGTH and SQL_DESC_OCTET_LENGTH, there is also SQL_DESC_DISPLAY_SIZE which has a similar issue. Does that value depend on the actual content of the column (as returned by the query) or does that depend on the definition of the column?
I can see that the width of the columns returned by SQLTables could depend on the limitations of the server for the type of information (schema name length, etc.) Unfortunately, that doesn't help currently since, as mentioned, the information is not passed on.
As to ANSI vs. Unicode. We only support UTF-8 in the narrow calls and UTF-16 in the wide calls. As I understand it, the term ANSI refers to a Windows code page which is close to ISO-8859-1, but in any case *not* compatible with UTF-8.
I suppose I could change the code so that the narrow interface is CP-1252 and if you want to use characters that don't occur in that code page you're forced to use the wide interface. I could also differentiate between the calls with or without A at the end (e.g. SQLTables vs. SQLTablesA) where the calls with A use the "ANSI" code page and the calls without use UTF-8. But I don't really like this idea.
(My weekend is starting now, so I probably won't respond before Monday.)
## Comment 13118
Date: 2010-04-16 18:52:39 +0200
From: Farid Z <<faridz>>
>What does "the actual character length for a variable-length data type" mean
It means the length of the column as defined by the DDL. Example, varchar(30) -> 30
I think we need to differentiate between describing a resultset columns (resultset description as returned by SQLColAttribute and other calls) and actual data values that are only known when the resultset is actually retrieved (stored in location pointed to by StrLen_or_Ind used in SQLBindCol).
When I am getting resultset description from the driver no data is retrieved (I don't want to retrieve any data yet, the table could have 4 million rows or no rows, but client only wants to know the resultset column descriptions). This is usually done by preparing the select statement "select * from <table>"
Remember I don't want the driver to retrieve the resultset I only want to know what is resultset description is. So the database can have 1000 tables and all the tables are empty, but the client should still be able to determine the resultset description for each table by preparing "select * from <table>" for example (or actually executing the select). So for example,
create table test (col1 varchar(30))
I should be able to determine that col1 resultset column length is 30 and octet length is 30 (ANSI) even though the table is empty by calling SQLColAttribute and SQL_DESC_LENGTH, SQL_DESC_OCTET_LENGTH
Whether the table has data or not is not-relevant to the resultset description.
>I suppose I could change the code so that the narrow interface is CP-1252 and if you want to use >characters that don't occur in that code page you're forced to use the wide interface
No. The client can be using Windows codepage 1252 or some other codepage (1256 Arabic, for example), depending on the client locale, etc. You can determine the code page the client is using (ANSI) by calling Windows API GetACP(VOID)
Since the server uses UTF8 then when transferring data to an ANSI client (narrow calls) the driver needs to convert from UTF8 to the codepage returned by GetACP() call. The conversion of course could fail if the client codepage has no equivalent to the UTF8 character and in this case the driver would report a conversion error for only that piece of data.
Similarly for wide clients, the driver would convert from UTF8 to UTF16 (UCS16, actually) and report any conversion errors (less likely).
## Comment 13119
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2970087 at http://sourceforge.net/support/tracker.php?aid=2970087
## Comment 16138
Date: 2011-08-22 20:21:27 +0200
From: @sjoerdmullender
A number of recent changes in the Aug2011 branch have hopefully solved the problems described here. The server now returns the size of the underlying datatype to the client (the ODBC driver) which passes it on to the application. Also, the various lengths that can be requested from SQLColAttribute (SQL_DESC_LENGTH, SQL_DESC_OCTET_LENGTH, SQL_DESC_DISPLAY_SIZE) now result in the values as described in the ODBC documentation. In particular, the octet length for character types (char, varchar, clob) is 6 times the value of the SQL_DESC_LENGTH value since the latter is number of characters, and UTF-8 encoding can take up to 6 bytes for a Unicode code point (character).
I will call this bug resolved, although you can always reopen the bug if you don't agree with the resolution.
## Comment 16256
Date: 2011-09-16 15:04:41 +0200
From: @sjoerdmullender
The Aug2011 version has been released.
| ANSI Client clobbered data ::SQLTables | https://api.github.com/repos/MonetDB/MonetDB/issues/2336/comments | 0 | 2020-11-30T10:38:17Z | 2024-06-27T11:44:07Z | https://github.com/MonetDB/MonetDB/issues/2336 | 753,361,675 | 2,336 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-13 05:55:16 +0100
From: Farid Z <<faridz>>
To: @sjoerdmullender
Version: -- development
Last updated: 2010-03-15 06:06:18 +0100
## Comment 13103
Date: 2010-03-13 17:55:16 +0100
From: Farid Z <<faridz>>
---------------------------
mserver5.exe - Entry Point Not Found
---------------------------
The procedure entry point stream_fgetpos could not be located in the dynamic link library libstream.dll.
---------------------------
OK
---------------------------
Trying to start server on Windows XP 32-bit SP3
Start -> All Programs -> MonetDB5 -> MonetDB SQL Server.
Using the latest build installed from MonetDB5-SQL-Installer-i386-20100224.msi
## Comment 13104
Date: 2010-03-13 21:28:00 +0100
From: @sjoerdmullender
This is probably caused by an old version of libstream.dll that you happen to have lying around.
In the past we used to install files in the Windows system folder (C:\Windows\system32). Look for files matching the pattern lib*.dll there. But this was quite a few releases (and years) ago.
Another possibility is that your libstream.dll file wasn't updated correctly (assuming you did an update from a previous version). Look in C:\Program Files\MonetDB\MonetDB5\lib and look at the date of the file libstream.dll. It should be dated Feb 24, 2010.
In the former case, remove those ancient files (making sure you don't remove too much). In the latter case, a repair installation (reinstall) might help, especially if you first remove the too old file.
When removing files, it is usually a good idea to rename them first, try things out, and if things work, really remove them.
## Comment 13105
Date: 2010-03-14 03:21:20 +0100
From: Farid Z <<faridz>>
Renamed old monetdb lib*.dll file in windows\system32 and this solved the issue. thanks
## Comment 13106
Date: 2010-03-15 18:06:18 +0100
From: @sjoerdmullender
Closed as "invalid" since this was not a bug in MonetDB.
(Actually, it could have been a bug in an ancient version that perhaps did not clean up all files when it was uninstalled. Since we no longer install in the system folder, that bug would be Out of Date.)
## Comment 13107
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2969906 at http://sourceforge.net/support/tracker.php?aid=2969906
| stream_fgetpos could not be located in libstream.dll | https://api.github.com/repos/MonetDB/MonetDB/issues/2335/comments | 0 | 2020-11-30T10:38:13Z | 2024-06-28T07:25:40Z | https://github.com/MonetDB/MonetDB/issues/2335 | 753,361,621 | 2,335 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-11 09:01:52 +0100
From: Hering Cheng <<heringcheng>>
To: @sjoerdmullender
Version: -- development
Last updated: 2010-03-15 11:17:36 +0100
## Comment 13098
Date: 2010-03-11 21:01:52 +0100
From: Hering Cheng <<heringcheng>>
Does MonetDB support Unicode/UTF-8 by default? I built my own 64-bit version on Solaris SPARC but am unable to load UTF data into a VARCHAR field. What am I missing? Here is the output from mclient:
MAPI = chenher@myserver:50000
QUERY = DELETE FROM mytable; COPY 132590 RECORDS INTO mytable FROM 'mydata.raw.txt.tmp' DELIMITERS '|', '\n', '"';
ERROR = !SQLException:sql:value 'ESAVINGSSTORE.COM,â–’INC.' from line 7797 field 5 not inserted, expecting type str
!SQLException:importTable:failed to import table
I am also attaching the record containing the offending data.
## Comment 13099
Date: 2010-03-11 21:23:19 +0100
From: @sjoerdmullender
MonetDB does support UTF-8 fully. However, your data isn't UTF-8. At least when I download it from SF, I see that between the comma and the word INC there is a Latin-1 no-break space (\240). That is not UTF-8, so MonetDB complains correctly about that.
## Comment 13100
Date: 2010-03-11 21:33:20 +0100
From: Hering Cheng <<heringcheng>>
Thank you for the response.
While I understand that my data is not actually UTF-8, I was still wondering if there is anything special I need to do to enable UTF-8.
This non-UTF-8 data comes from a Sybase IQ database with VARCHAR as the data type. This must have been some garbage that somehow Sybase accepted. Is there a recommended way for me to cleanse my text file (i.e., to rid it of non-UTF data) before I load it into MonetDB?
## Comment 13101
Date: 2010-03-11 21:41:39 +0100
From: @sjoerdmullender
I suppose you could use a tool such as iconv to convert your data from just about any encoding to UTF-8, but you do need to know the encoding.
If you want to just convert no-break spaces and other weird characters to something more "normal", I suppose you could use sed.
Mclient can also convert data, but when you use the COPY INTO command and you specify an actual file name (as opposed to STDIN), then the server reads the data directly, and it does not do conversions.
## Comment 13102
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2968881 at http://sourceforge.net/support/tracker.php?aid=2968881
| Unicode Support | https://api.github.com/repos/MonetDB/MonetDB/issues/2334/comments | 0 | 2020-11-30T10:38:10Z | 2024-06-28T07:25:39Z | https://github.com/MonetDB/MonetDB/issues/2334 | 753,361,577 | 2,334 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-08 03:12:39 +0100
From: @grobian
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:18:31 +0200
## Comment 13095
Date: 2010-03-08 15:12:39 +0100
From: @grobian
% $INSTALL_DIR/bin/mclient -dhibernate
Welcome to mclient, the MonetDB/SQL interactive terminal
Database: MonetDB v5.18.2, 'hibernate'
Type \q to quit, \? for a list of available commands
auto commit mode: on
sql>select 'yes' where 3 in (1,2);
MAPI = monetdb@localhost:50000
ACTION= read_line
QUERY = select 'yes' where 3 in (1,2);
ERROR = !Connection terminated
2010-03-08 15:09:26 MSG merovingian[10450]: database 'hibernate' has crashed aft
er start on 2010-03-08 14:09:53, attempting restart, up min/avg/max: 0s/0s/0s, c
rash average: 1.00 0.10 0.03 (1-0=1)
2010-03-08 15:09:26 MSG hibernate[12286]: arguments: /ufs/fabian/scratch/monetdb
/Feb2010/program-x86_64/bin/mserver5 --config=/ufs/fabian/scratch/monetdb/Feb201
0/program-x86_64/etc/monetdb5.conf --dbname=hibernate --dbinit=include sql; --se
t merovingian_uri=mapi:monetdb://volund.ins.cwi.nl:50000/hibernate --set monet_d
aemon=yes --set mapi_open=false --set mapi_autosense=true --set mapi_port=50001
--set monet_vault_key=/ufs/fabian/scratch/monetdb/Feb2010/program-x86_64/var/Mon
etDB5/dbfarm/hibernate/.vaultkey --set sql_optimizer=default_pipe
2010-03-08 15:09:26 MSG hibernate[12286]: MonetDB server v5.18.2, based on kernel v1.36.2
2010-03-08 15:09:26 MSG hibernate[12286]: Serving database 'hibernate', using 4 threads
2010-03-08 15:09:26 MSG hibernate[12286]: Compiled for x86_64-pc-linux-gnu/64bit with 64bit OIDs dynamically linked
2010-03-08 15:09:26 MSG hibernate[12286]: Found 7.751 GiB available main-memory.
2010-03-08 15:09:26 MSG hibernate[12286]: Copyright (c) 1993-July 2008 CWI.
2010-03-08 15:09:26 MSG hibernate[12286]: Copyright (c) August 2008-2010 MonetDB B.V., all rights reserved
2010-03-08 15:09:26 MSG hibernate[12286]: Visit http://monetdb.cwi.nl/ for further information
2010-03-08 15:09:26 MSG hibernate[12286]: Listening for connection requests on mapi:monetdb://127.0.0.1:50002/
2010-03-08 15:09:26 MSG hibernate[12286]: MonetDB/SQL module v2.36.2 loaded
2010-03-08 15:09:27 MSG merovingian[10450]: proxying client localhost:39642 for database 'hibernate' to mapi:monetdb://127.0.0.1:50002/hibernate
2010-03-08 15:09:31 MSG merovingian[10450]: database 'hibernate' (12286) was killed by signal 11
2010-03-08 15:09:31 MSG merovingian[10450]: client localhost:39642 has disconnected from proxy
## Comment 13096
Date: 2010-04-09 14:25:54 +0200
From: @njnes
seems to run fine.
## Comment 13097
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2965448 at http://sourceforge.net/support/tracker.php?aid=2965448
## Comment 14651
Date: 2010-08-12 19:45:45 +0200
From: @njnes
added test to BugTracker-2010/crash_on_where_in.Bug-2333.sql
## Comment 14653
Date: 2010-08-12 19:47:00 +0200
From: @njnes
Changeset [b24b7f059dec](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=b24b7f059dec) made by Niels Nes <niels@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=b24b7f059dec](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=b24b7f059dec)
Changeset description:
added test for (long) fixed bug #2333
## Comment 14823
Date: 2010-08-30 09:18:31 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| SQL: crash on where in query | https://api.github.com/repos/MonetDB/MonetDB/issues/2333/comments | 0 | 2020-11-30T10:38:07Z | 2024-06-27T11:44:04Z | https://github.com/MonetDB/MonetDB/issues/2333 | 753,361,544 | 2,333 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-05 03:46:09 +0100
From: @grobian
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2010-03-29 02:56:49 +0200
## Comment 13090
Date: 2010-03-05 15:46:09 +0100
From: @grobian
Merovingian crashes with a buffer overrun if a remote client connects to it if this client's remote IP has 3 digits in most of its quads.
## Comment 13091
Date: 2010-03-05 15:52:31 +0100
From: @grobian
additionally the remote IP must not resolve to a hostname
## Comment 13092
Date: 2010-03-05 15:56:38 +0100
From: @grobian
Fix committed, should appear in Feb2010-SP1
## Comment 13093
Date: 2010-03-29 14:56:48 +0200
From: @grobian
Fix released in Feb2010-SP1
## Comment 13094
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2964247 at http://sourceforge.net/support/tracker.php?aid=2964247
| Merovingian: crash upon client connect from remote host | https://api.github.com/repos/MonetDB/MonetDB/issues/2332/comments | 0 | 2020-11-30T10:38:02Z | 2024-06-27T11:44:03Z | https://github.com/MonetDB/MonetDB/issues/2332 | 753,361,490 | 2,332 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-05 04:28:20 +0100
From: Mark Bucciarelli <<mkbucc>>
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2010-03-05 03:43:49 +0100
## Comment 13084
Date: 2010-03-05 04:28:20 +0100
From: Mark Bucciarelli <<mkbucc>>
Hi,
This is with Feb2010 branch. The second argument to the
snprintf() call in generatePassphraseFile() was taking sizeof
a pointer. This patch fixes this issue, as well as a couple other
issues with the routine, namely:
* fopen() return value was not checked
* strdup() return value was not checked
You may not agree with the way I handle these
errors (crash immediately), but I wasn't sure
how important this routine is. (In general I
have found it productive to crash early and
crash hard.)
A quick grep turned up another couple instances
where fopen's return value was not checked around
line 125 of monet5/merovingian/databaase.c.
That routine also has lots of strdup's with no return
value check. (Not that it matters on Linux ... where
calloc() _always_ returns non-NULL.)
## Comment 13085
Date: 2010-03-05 10:29:00 +0100
From: @grobian
I think these are bits in my code, can you report exactly which files these are in?
## Comment 13086
Date: 2010-03-05 13:53:35 +0100
From: Mark Bucciarelli <<mkbucc>>
I'm not sure what you are asking for.
I attached patch for one.
I referenced one other spot in the description with
the filename and line .
## Comment 13087
Date: 2010-03-05 14:03:55 +0100
From: @grobian
I didn't notice the attachment, sorry
## Comment 13088
Date: 2010-03-05 15:43:49 +0100
From: @grobian
The failing strdup in an error condition is happening all over the code, I won't fix that soon. The fopen return are being handled now, thanks.
## Comment 13089
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2963916 at http://sourceforge.net/support/tracker.php?aid=2963916
| sizeof a pointer passed to snprintf | https://api.github.com/repos/MonetDB/MonetDB/issues/2331/comments | 0 | 2020-11-30T10:37:58Z | 2024-06-28T07:25:39Z | https://github.com/MonetDB/MonetDB/issues/2331 | 753,361,439 | 2,331 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-03-04 11:50:26 +0100
From: Mark Bucciarelli <<mkbucc>>
To: @sjoerdmullender
Version: -- development
Last updated: 2010-03-05 01:56:32 +0100
## Comment 13081
Date: 2010-03-04 23:50:26 +0100
From: Mark Bucciarelli <<mkbucc>>
Hi,
With the attached patch, the Mx segfaults I was experiencing
go away. It's a variant of what Sjoerd posted on the dev list.
## Comment 13082
Date: 2010-03-05 13:56:31 +0100
From: @sjoerdmullender
I apllied my own version of the patch instead of yours. The effect is the same, though.
(I wan't convinced of the need for an extra variable.)
No test needed.
## Comment 13083
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2963782 at http://sourceforge.net/support/tracker.php?aid=2963782
| Patch for Mx Segfault | https://api.github.com/repos/MonetDB/MonetDB/issues/2330/comments | 0 | 2020-11-30T10:37:55Z | 2024-06-27T11:44:01Z | https://github.com/MonetDB/MonetDB/issues/2330 | 753,361,403 | 2,330 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-28 07:15:23 +0100
From: @drstmane
To: @mlkersten
Version: -- development
Last updated: 2010-03-17 08:18:54 +0100
## Comment 13075
Date: 2010-02-28 19:15:23 +0100
From: @drstmane
As far as I can see, global static variable "hottop" in MonetDB5/src/mal/mal_interpreter.mx is never initialized, nor is there any check that, in particular with the increment in updMemoryUsed(), it does not exceed MAXHOT. Either can (and does) lead to SEGFAULTs:
(while running TPCH SF-100 with a debug build from http://monetdb.cwi.nl/downloads/sources/Feb2010/MonetDB-Feb2010-SuperBall.tar.bz2)
Program terminated with signal 11, Segmentation fault.
0 0x000000000097d6ee in updMemoryUsedPart (stk=0x7f3a467e3858, pci=0x7f3f5edc3fa8, start=0, stop=1, argclaim=-1) at mal_interpreter.c:6778
6778 if ( hotpotatoes[j].bid != ABS(bid))
Missing separate debuginfos, use: debuginfo-install bzip2-libs-1.0.5-3.fc10.x86_64 cyrus-sasl-lib-2.1.22-19.fc10.x86_64 e2fsprogs-libs-1.41.4-6.fc10.x86_64 glibc-2.9-3.x86_64 keyutils-libs-1.2-3.fc9.x86_64 krb5-libs-1.6.3-18.fc10.x86_64 libcurl-7.19.6-2.fc10.x86_64 libidn-0.6.14-8.x86_64 libselinux-2.0.78-1.fc10.x86_64 libssh2-0.18-7.fc9.x86_64 libxml2-2.7.6-1.fc10.x86_64 libxslt-1.1.26-1.fc10.x86_64 ncurses-libs-5.6-20.20080927.fc10.x86_64 nspr-4.7.6-1.fc10.1.x86_64 nss-3.12.3.99.3-2.10.6.fc10.x86_64 openldap-2.4.12-1.fc10.x86_64 openssl-0.9.8g-14.fc10.x86_64 pcre-7.8-1.fc10.x86_64 raptor-1.4.18-5.fc10.x86_64 readline-5.2-13.fc9.x86_64 zlib-1.2.3-18.fc9.x86_64
(gdb) li
6773 }
6774 for ( i = start; i< stop; i++)
6775 if (stk->stk[getArg(pci,i)].vtype == TYPE_bat && (bid = stk->stk[getArg(pci,i)].val.bval) && bid)
6776 {
6777 for ( h = j= 0; j< hottop; j++)
6778 if ( hotpotatoes[j].bid != ABS(bid))
6779 hotpotatoes[h++]= hotpotatoes[j];
6780 else{
6781 ifdef DEBUG_MEMORY_CLAIM
6782 if ( hotpotatoes[j].claim){
(gdb) print hottop
$1 = 600037897
(gdb) print j
$2 = 44184
(gdb) print h
$3 = 44184
(gdb) print hotpotatoes[j]
Cannot access memory at address 0xc83000
(gdb) print hotpotatoes
$4 = {{claim = 0, bid = 0} <repeats 11 times>, {claim = 600037896, bid = 1334}, {claim = 600037896, bid = 1278}, {claim = 600037896, bid = 1359}, {claim = 600037944, bid = 1377}, {claim = 600037896, bid = 1275}}
(gdb) print hotpotatoes[0]
$5 = {claim = 0, bid = 0}
(gdb) print hotpotatoes[15]
$6 = {claim = 600037896, bid = 1275}
(gdb) print hotpotatoes[16]
$7 = {claim = 600037897, bid = 1413}
(gdb) print hotpotatoes[17]
$8 = {claim = 0, bid = 0}
(gdb) print hotpotatoes[100]
$9 = {claim = 0, bid = 0}
(gdb) print hotpotatoes[1000]
$10 = {claim = 0, bid = 0}
(gdb) print hotpotatoes[10000]
$11 = {claim = 0, bid = 0}
(gdb) print hotpotatoes[100000]
Cannot access memory at address 0xd5d080
(gdb)
## Comment 13076
Date: 2010-02-28 19:50:44 +0100
From: @mlkersten
Indeed. It missed the point that an instruction could return multiple results, which could bypass the limit without noticing.
## Comment 13077
Date: 2010-02-28 20:16:17 +0100
From: @drstmane
Thanks.
What would be a proper initial value for hottop?
0, I guess?
## Comment 13078
Date: 2010-03-01 09:28:49 +0100
From: @drstmane
I check in an initialization of hottop to 0 as follows.
Is that OK?
Index: MonetDB5/src/mal/mal_interpreter.mx
===================================================================
RCS file: /cvsroot/monetdb/MonetDB5/src/mal/mal_interpreter.mx,v
retrieving revision 1.337.2.33
retrieving revision 1.337.2.34
diff -u -r1.337.2.33 -r1.337.2.34
--- MonetDB5/src/mal/mal_interpreter.mx 28 Feb 2010 18:49:24 -0000 1.337.2.33
+++ MonetDB5/src/mal/mal_interpreter.mx 1 Mar 2010 08:27:04 -0000 1.337.2.34
@@ -89,7 +89,7 @@
lng claim; /* actual claim on memory*/
int bid;
} hotpotatoes[MAXHOT];
-static int hottop;
+static int hottop = 0;
@-
The struct alignment leads to 40% gain in simple instructions when set.
@c
## Comment 13079
Date: 2010-03-17 20:18:54 +0100
From: @drstmane
My 0-initialization of a static variable was of course "bogus". Sorry.
However, the hottop "overflow" appears to be fixed by
===================================================================
2010/02/28 - mlkersten: MonetDB5//src/mal/mal_interpreter.mx,1.337.2.33
Protect hottop against the case that an MAL instruction can produce
multiple results and then cross the MAXHOT barrier without notice.
===================================================================
Adding TPCH SF-100 as test is not feasible, hence, I close without test.
## Comment 13080
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2960781 at http://sourceforge.net/support/tracker.php?aid=2960781
| M5: potiential SEGFAULTs in mal_interpreter | https://api.github.com/repos/MonetDB/MonetDB/issues/2329/comments | 0 | 2020-11-30T10:37:51Z | 2024-06-27T11:44:00Z | https://github.com/MonetDB/MonetDB/issues/2329 | 753,361,355 | 2,329 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-27 02:31:02 +0100
From: Mark Kubacki <<hurrikane>>
To: @sjoerdmullender
Version: -- development
Last updated: 2010-04-16 04:24:56 +0200
## Comment 13068
Date: 2010-02-27 14:31:02 +0100
From: Mark Kubacki <<hurrikane>>
Given MonetDB-1.36.1 (Feb2010 testing) and installed PCL (tested with
dev-libs/pcl-1.6) following command leads to errors:
(this didn't happen in MonetDB-1.34.4)
Reproduce:
cd MonetDB-1.36.1
./configure --with-pcl
make
Expected:
compiles fine
Got:
[... configuring OK]
[... 1 line by make, then:]
/bin/sh ../../libtool --tag=CC --mode=link gcc -DLIBMUTILS -O2 -D_REENTRANT -version-number 1:36:1 -o libmutils.la -rpath /usr/local/lib libmutils_la-monet_options.lo -Lyes/lib
../../libtool: line 4984: cd: yes/lib: No such file or directory
libtool: link: cannot determine absolute directory name of `yes/lib'
[...]
make: *** [all] Error 2
Additional information:
OS: GNU/Linux-2.6.31
Dist: Gentoo 10.1
app-shells/bash: 4.0_p35
dev-lang/python: 2.6.4
dev-util/ccache: 2.4-r7
dev-util/cmake: 2.8.0-r2
sys-apps/baselayout: 1.12.13
sys-apps/sandbox: 1.6-r2
sys-devel/autoconf: 2.63-r1
sys-devel/automake: 1.9.6-r2, 1.10.2
sys-devel/binutils: 2.20.51.0.6
sys-devel/gcc: 4.3.4, 4.4.3
sys-devel/gcc-config: 1.4.1
sys-devel/libtool: 2.2.6b
virtual/os-headers: 2.6.32
## Comment 13069
Date: 2010-02-27 14:47:11 +0100
From: Mark Kubacki <<hurrikane>>
MonetDB-client and perhaps other packages are affected by this, too.
## Comment 13070
Date: 2010-03-01 17:19:26 +0100
From: @drstmane
Indeed, our configure scripts appear to be buggy in this case, i.e., they cannot handle --with-pc[=yes] correctly.
Thanks for reporting and sorry for the inconvenience.
Until we manage to fix this, --with-pcl[=yes] should be the default, i.e., pcl should be found automatically if it is available in /usr/{include,lib}, i.e., omitting --with-pcl[=yes] should work fine.
If pcl is available elsewhere, say in /usr/local/pcl/{include,lib}, you'd need to tell this to configure via --with-pcl=/usr/local/pcl .
Note, though, that only the depercated MonetDB4 back-end provides some pcl support via it's pcl module.
MonetDB5 (and hence MonetDB/SQL) does not (yet?) provide any pcl support at all; hence, omitting --with-pcl is just fine.
## Comment 13071
Date: 2010-03-17 14:26:33 +0100
From: @sjoerdmullender
I fixed configure, hopefully correctly. I have no easy access to a PCL library (there is no prepackaged version available from the Fedora repository), so I can't easily check this (except that it now properly gives an error message when you say --with-pcl when the library cannot be found).
Can you test this? This would involve trying from CVS or using tar balls from the nightly builds.
## Comment 13072
Date: 2010-03-26 17:34:29 +0100
From: Mark Kubacki <<hurrikane>>
If --with-pcl is actually not used, it should be removed from configure scripts completely.
## Comment 13073
Date: 2010-04-16 16:24:56 +0200
From: @sjoerdmullender
In order to reduce confusion I moved the PCL configuration code to MonetDB4 configure and removed it from all other configures. Only MonetDB4 has code to use the pcl library.
## Comment 13074
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2960174 at http://sourceforge.net/support/tracker.php?aid=2960174
| compiling fails with --with-pcl | https://api.github.com/repos/MonetDB/MonetDB/issues/2328/comments | 0 | 2020-11-30T10:37:48Z | 2024-06-27T11:43:59Z | https://github.com/MonetDB/MonetDB/issues/2328 | 753,361,314 | 2,328 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-27 05:20:01 +0100
From: @skinkie
To: @njnes
Version: -- development
Last updated: 2010-04-09 02:28:15 +0200
## Comment 13065
Date: 2010-02-27 05:20:01 +0100
From: @skinkie
I wonder if the following is per spec behavior, every constant value not equal to 0 gets casted to true. Is there a reason to allow arbitrary constants?
select * from tables where 12345;
## Comment 13066
Date: 2010-04-09 14:28:15 +0200
From: @njnes
The spec isn't clear on this. Auto-conversions between numbers are indeed always done. From numeric to boolean is just an added bonus. I don't consider this a bug.
## Comment 13067
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2959974 at http://sourceforge.net/support/tracker.php?aid=2959974
| SQL; where constant value | https://api.github.com/repos/MonetDB/MonetDB/issues/2327/comments | 0 | 2020-11-30T10:37:45Z | 2024-06-28T07:25:39Z | https://github.com/MonetDB/MonetDB/issues/2327 | 753,361,281 | 2,327 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-22 05:35:27 +0100
From: @grobian
To: @mlkersten
Version: 11.11.5 (Jul2012)
Last updated: 2012-07-17 14:31:00 +0200
## Comment 13059
Date: 2010-02-22 17:35:27 +0100
From: @grobian
(volund:five/src/mal) fabian% $INSTALL_DIR/bin/mclient -dsn4 -lmal
Welcome to mclient, the MonetDB interactive terminal
Type \q to quit, \? for a list of available commands
mal>function user.s0_0():void;
mal>io.print("hello");
mal>end s0_0;
mal>user.s0_0();
[ "hello" ]
(volund:five/src/mal) fabian% $INSTALL_DIR/bin/mclient -dsn4 -lmal
Welcome to mclient, the MonetDB interactive terminal
Type \q to quit, \? for a list of available commands
mal>user.s0_0();
[ "hello" ]
(volund:five/src/mal) fabian% $INSTALL_DIR/bin/mclient -dsn4 -lmal
Welcome to mclient, the MonetDB interactive terminal
Type \q to quit, \? for a list of available commands
mal>user.s0_0();
[ "hello" ]
(volund:five/src/mal) fabian%
I would have expected the function s0_0 to be destroyed upon client disconnect.
(volund:five/src/mal) fabian% $INSTALL_DIR/bin/mserver5 --version
MonetDB server v5.19.0 (64-bit), based on kernel v1.37.0 (64-bit oids)
Copyright (c) 1993-July 2008 CWI
Copyright (c) August 2008-2010 MonetDB B.V., all rights reserved
Visit http://monetdb.cwi.nl/ for further information
Found 7.8GiB available memory, 4 available cpu cores
Configured for prefix: /ufs/fabian/scratch/monetdb/current/program-x86_64
Libraries:
libpcre: 8.00 2009-10-19 (compiled with 8.00)
openssl: OpenSSL 0.9.8l 5 Nov 2009 (compiled with OpenSSL 0.9.8l 5 Nov 2009)
libxml2: 2.7.6 (compiled with 2.7.6)
Compiled by: fabian@volund.ins.cwi.nl (x86_64-pc-linux-gnu)
Compilation: x86_64-pc-linux-gnu-gcc -march=native -pipe -W -Wall -Wextra -g -Werror -Werror-implicit-function-declaration -Wpointer-arith -Wdeclaration-after-statement -Wp,-D_FORTIFY_SOURCE=2
Linking : ld -m elf_x86_64 -Wl,--as-needed
## Comment 13060
Date: 2010-02-22 17:47:23 +0100
From: @mlkersten
The quit() command can be embedded in MAL functions, easily leading into a shooting in the foot when you clean out the user module. However, we can clean it upon first re-use of the client context.
## Comment 13061
Date: 2010-02-22 17:49:24 +0100
From: @grobian
So, I can assume you fixed it like that? Or did you mean that this bug is invalid?
## Comment 13062
Date: 2010-02-22 17:54:23 +0100
From: @grobian
I see it was committed to start fresh upon connect, ok, thanks!
## Comment 13063
Date: 2010-02-23 09:48:26 +0100
From: @grobian
reopening, since msql mode seems not to be fixed by previous checkin (the one that suffers the most from the problem)
## Comment 13064
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2956664 at http://sourceforge.net/support/tracker.php?aid=2956664
## Comment 16151
Date: 2011-08-24 00:48:44 +0200
From: @mlkersten
Patched the client initialization code to get rid of old MAL plans.
It seems to work in the distributed setting of Centipede.
The example below also works.
There is still a place where instructions freed are not properly
reflected in the mal blocks.
## Comment 16152
Date: 2011-08-24 01:05:07 +0200
From: @mlkersten
Only the centipede causes problems when multiple connections are used.
The other pipe lines seem clean.
## Comment 16195
Date: 2011-09-08 12:42:55 +0200
From: @sjoerdmullender
On the Aug2011 branch this is still a problem. On the default branch, at least when using -lmal, this seems to be fixed.
## Comment 17485
Date: 2012-07-17 14:31:00 +0200
From: @grobian
fixed in Jul2012
| "user" module persistent over connections | https://api.github.com/repos/MonetDB/MonetDB/issues/2326/comments | 0 | 2020-11-30T10:37:41Z | 2024-06-27T11:43:57Z | https://github.com/MonetDB/MonetDB/issues/2326 | 753,361,249 | 2,326 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-22 02:24:01 +0100
From: @swingbit
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2010-04-15 05:51:08 +0200
## Comment 13054
Date: 2010-02-22 14:24:01 +0100
From: @swingbit
The error code of mclient --help is 255. This is unusual, as no error occurred and also not in line with mserver5 and Mserver, which return 0 on --help.
## Comment 13055
Date: 2010-02-22 16:27:09 +0100
From: @sjoerdmullender
This is not as easy to do correctly as it sounds.
The problem is, we use -? as alternative spelling for --help. This means that getopt returns '?' if you pass --help or -? as an option. But '?' is also returned for an incorrect option. Only in the former case should we exit with exit status 0, in the latter case we should exit with exit status != 0. But there doesn't seem to be a way to distinguish between the two cases.
## Comment 13056
Date: 2010-02-22 16:33:07 +0100
From: @swingbit
If you use incorrect options in mserver5, you still get the help message - which is what you want in mclient right?
Is '-? ' alternative spelling in mclient necessary?
## Comment 13057
Date: 2010-04-15 17:51:07 +0200
From: @sjoerdmullender
I implemented a fix.
No test. I don't think we look at the exit code during testing, or do we?
## Comment 13058
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2956574 at http://sourceforge.net/support/tracker.php?aid=2956574
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| 'mclients --help' returns 255 | https://api.github.com/repos/MonetDB/MonetDB/issues/2325/comments | 0 | 2020-11-30T10:37:38Z | 2024-06-27T11:43:56Z | https://github.com/MonetDB/MonetDB/issues/2325 | 753,361,215 | 2,325 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-17 01:56:36 +0100
From: @skinkie
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2010-02-18 09:30:49 +0100
## Comment 13045
Date: 2010-02-17 01:56:36 +0100
From: @skinkie
Current head, since this afternoon.
sql>select * from kvk where kvk = 1;
MALException:!ERROR: madvise(a1d4bfa0, 9572264, 0) on (&b->T->heap) BUF_TO_MMAP[tb] failed
OS: Invalid argument
## Comment 13046
Date: 2010-02-17 02:05:51 +0100
From: @drstmane
Most probably fixed in the Feb2010 branch by
===================================================================
2010/02/12 - stmane: MonetDB5/src/mal/mal_interpreter.mx,1.337.2.22
undoing the part of
"
2010/02/12 - mlkersten: MonetDB5/src/mal/mal_interpreter.mx,1.337.2.21
Clean up of the getMemoryClaim. Debugging.
"
that added a BATmadvise(),
as this seems to cause BATmadvise() errors
and segfaults in various tests
including TPCH & SSB
===================================================================
last propagation to the development trunk happened between the original checkin and the above fix;
next propagation to the development trunk will happen once time and SF allow ...
## Comment 13047
Date: 2010-02-17 03:02:46 +0100
From: @skinkie
Vielen Dank!
## Comment 13048
Date: 2010-02-18 02:26:15 +0100
From: @drstmane
Sjoerd propagated the fix.
Please check and report whether it now works for you.
## Comment 13049
Date: 2010-02-18 02:33:18 +0100
From: @drstmane
Reports can be closed once Stefan dK & this test
http://monetdb.cwi.nl/testing/projects/monetdb/Current/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_BugTracker-2009/union_leak.SF-2540169.err.00.html
confirm the fix.
## Comment 13050
Date: 2010-02-18 02:52:09 +0100
From: @skinkie
The error doesn't occur anymore. But maybe someone should look at this, especially the time taken for the where part.
sql>select max(kvk) from kvk;
+--------------+
| L6 |
+==============+
| 412657690010 |
+--------------+
1 tuple
Timer 27.874 msec 1 rows
sql>select kvk from kvk where kvk = 412657690010;
+--------------+
| kvk |
+==============+
| 412657690010 |
+--------------+
1 tuple
Timer 414.478 msec 1 rows
sql>select avg(kvk) from kvk;
+------------------------+
| L11 |
+========================+
| 180371715703.39435 |
+------------------------+
1 tuple
Timer 67.727 msec 1 rows
## Comment 13051
Date: 2010-02-18 09:32:39 +0100
From: @drstmane
Stefan, you have the enviroment (data, database) and means (MonetDB) to check yourself where the time goes:
TRACE select max(kvk) from kvk;
TRACE select kvk from kvk where kvk = 412657690010;
TRACE select avg(kvk) from kvk;
Since this is not related to the original bug report, please repost the results elsewhere, preferably on the MonetDB-developers mailing list, or a new bug report, if you indeed consider it a bug.
## Comment 13052
Date: 2010-02-18 21:30:49 +0100
From: @drstmane
Also test
http://monetdb.cwi.nl/testing/projects/monetdb/Current/sql/.mTests103/GNU.64.64.d.1-Fedora10/src_test_BugTracker-2009/union_leak.SF-2540169.err.00.html
works fine, again.
Closing.
## Comment 13053
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2953215 at http://sourceforge.net/support/tracker.php?aid=2953215
| Head results in madvise error | https://api.github.com/repos/MonetDB/MonetDB/issues/2324/comments | 0 | 2020-11-30T10:37:34Z | 2024-06-27T11:43:55Z | https://github.com/MonetDB/MonetDB/issues/2324 | 753,361,176 | 2,324 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-15 05:57:58 +0100
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2010-02-18 09:58:39 +0100
## Comment 13038
Date: 2010-02-15 17:57:58 +0100
From: Wouter Alink <<alink>>
Nov2009, the bug can be reproduced with the attached script.
Result of a merge-join operation doesn't pass all the checks.
Running the query (the last statement of the attached script)
SELECT dd.docid as docid
FROM y as sd, x as dd
WHERE sd.docid = dd.docoid;
returns:
ERROR = !MALException:!ERROR: BATpropcheck: BAT tmpr_6767(-3575) with 26584 tuples seqbase of dense oid bat is wrong! 0 != 4196
## Comment 13039
Date: 2010-02-15 18:29:58 +0100
From: @swingbit
Notice that there is probably nothing wrong with mergejoin here.
The whole point is that mergejoin should not have been selected in the first place, as dd.docoid is NOT sorted (it is until almost the end, but not till the end) and the attached bat has properties correctly set (not sorted).
sql debug session:
mdb> _20 := algebra.join(_3=<LHM_sec_doc_docID>:bat[:oid,:lng][26584],_19=<tmpr_3130>[999]);
L is [void,lng], tail is sorted
mdb>info _3
Show info for 291
-------------------------------------------------
h t name
str str type
-------------------------------------------------
[ "batId", "LHM_sec_doc_docID" ]
[ "batCacheid", "291" ]
[ "hparentid", "0" ]
[ "tparentid", "0" ]
[ "batSharecnt", "0" ]
[ "batCount", "26584" ]
[ "batCapacity", "32768" ]
[ "head", "void" ]
[ "tail", "lng" ]
[ "batPersistence", "persistent" ]
[ "batRestricted", "read-only" ]
[ "batRefcnt", "1" ]
[ "batLRefcnt", "2" ]
[ "batDirty", "clean" ]
[ "batSet", "0" ]
[ "hsorted", "65" ]
[ "hident", "h" ]
[ "hdense", "1" ]
[ "hseqbase", "0@0" ]
[ "hkey", "1" ]
[ "hvarsized", "1" ]
[ "halign", "1018390" ]
[ "hnosorted", "0" ]
[ "hnosorted_rev", "0" ]
[ "hnodense", "0" ]
[ "hnokey[0]", "0" ]
[ "hnokey[1]", "0" ]
[ "hnonil", "1" ]
[ "hnil", "0" ]
[ "tident", "t" ]
[ "tdense", "0" ]
[ "tseqbase", "0@0" ]
[ "tsorted", "65" ]
[ "tkey", "0" ]
[ "tvarsized", "0" ]
[ "talign", "1018391" ]
[ "tnosorted", "0" ]
[ "tnosorted_rev", "0" ]
[ "tnodense", "0" ]
[ "tnokey[0]", "0" ]
[ "tnokey[1]", "1" ]
[ "tnonil", "1" ]
[ "tnil", "0" ]
[ "batInserted", "26584" ]
[ "batDeleted", "0" ]
[ "batFirst", "0" ]
[ "htop", "0" ]
[ "ttop", "212672" ]
[ "batStamp", "0" ]
[ "lastUsed", "485486" ]
[ "curStamp", "210" ]
[ "batCopiedtodisk", "1" ]
[ "batDirtydesc", "clean" ]
[ "H->heap.dirty", "clean" ]
[ "T->heap.dirty", "clean" ]
[ "head.free", "0" ]
[ "head.size", "0" ]
[ "head.maxsize", "0" ]
[ "head.storage", "absent" ]
[ "head.newstorage", "memory mapped" ]
[ "head.filename", "no file" ]
[ "tail.free", "212672" ]
[ "tail.size", "262144" ]
[ "tail.maxsize", "262144" ]
[ "tail.storage", "memory mapped" ]
[ "tail.newstorage", "memory mapped" ]
[ "tail.filename", "04/443.tail" ]
[ "H->vheap->dirty", "clean" ]
[ "T->vheap->dirty", "clean" ]
mdb> _20 := algebra.join(_3=<LHM_sec_doc_docID>:bat[:oid,:lng][26584],_19=<tmpr_3130>[999]);
R is [lng,oid], head is NOT sorted
mdb>info _19
Show info for -1624
-----------------------------------------
h t name
str str type
-----------------------------------------
[ "batId", "tmpr_3130" ]
[ "batCacheid", "-1624" ]
[ "hparentid", "-998" ]
[ "tparentid", "0" ]
[ "batSharecnt", "0" ]
[ "batCount", "999" ]
[ "batCapacity", "1024" ]
[ "head", "lng" ]
[ "tail", "void" ]
[ "batPersistence", "transient" ]
[ "batRestricted", "read-only" ]
[ "batRefcnt", "1" ]
[ "batLRefcnt", "1" ]
[ "batDirty", "dirty" ]
[ "batSet", "0" ]
[ "hsorted", "0" ]
[ "hident", "t" ]
[ "hdense", "0" ]
[ "hseqbase", "0@0" ]
[ "hkey", "1" ]
[ "hvarsized", "0" ]
[ "halign", "1027928" ]
[ "hnosorted", "990" ]
[ "hnosorted_rev", "0" ]
[ "hnodense", "0" ]
[ "hnokey[0]", "0" ]
[ "hnokey[1]", "0" ]
[ "hnonil", "1" ]
[ "hnil", "0" ]
[ "tident", "h" ]
[ "tdense", "1" ]
[ "tseqbase", "0@0" ]
[ "tsorted", "65" ]
[ "tkey", "1" ]
[ "tvarsized", "1" ]
[ "talign", "1029656" ]
[ "tnosorted", "0" ]
[ "tnosorted_rev", "0" ]
[ "tnodense", "0" ]
[ "tnokey[0]", "0" ]
[ "tnokey[1]", "0" ]
[ "tnonil", "1" ]
[ "tnil", "0" ]
[ "batInserted", "999" ]
[ "batDeleted", "0" ]
[ "batFirst", "0" ]
[ "htop", "7992" ]
[ "ttop", "0" ]
[ "batStamp", "208" ]
[ "lastUsed", "485645" ]
[ "curStamp", "209" ]
[ "batCopiedtodisk", "0" ]
[ "batDirtydesc", "dirty" ]
[ "H->heap.dirty", "clean" ]
[ "T->heap.dirty", "clean" ]
[ "head.free", "7992" ]
[ "head.size", "8192" ]
[ "head.maxsize", "8192" ]
[ "head.storage", "malloced" ]
[ "head.newstorage", "malloced" ]
[ "head.filename", "17/1746.tail" ]
[ "tail.free", "0" ]
[ "tail.size", "0" ]
[ "tail.maxsize", "0" ]
[ "tail.storage", "absent" ]
[ "tail.newstorage", "malloced" ]
[ "tail.filename", "no file" ]
[ "H->vheap->dirty", "clean" ]
[ "T->vheap->dirty", "clean" ]
mdb> _20 := algebra.join(_3=<LHM_sec_doc_docID>:bat[:oid,:lng][26584],_19=<tmpr_3130>[999]);
mdb>dehttp://sourceforge.net/support/tracker.php?aid=2097152
Set debug mask to 2097152
with the head of R not sorted, a mergejoin is selected!
mdb> _20 := algebra.join(_3=<LHM_sec_doc_docID>:bat[:oid,:lng][26584],_19=<tmpr_3130>[999]);
mdb>
BATjoin: BATmergejoin(l,r,9223372036854775807);
ERROR: GDKerror:!ERROR: BATpropcheck: BAT tmpr_3132(-1626) with 26584 tuples seqbase of dense oid bat is wrong! 0 != 4196
mdb>
## Comment 13040
Date: 2010-02-15 20:01:54 +0100
From: @drstmane
(a)
it seems to work fine in the Feb2010 branch
(need to check what has changed since Nov2009 ...
(b)
in the Nov2009 branch, mergejoin --- in fact a binary lookup join between BATmirror(r) and BATmirror(l) --- is chosen here:
Breakpoint 1, mergejoin (l=0x7fffd00008c8, r=0x7fffc00008f0, bn=0x0, nil_on_miss=0x0, estimate=9223372036854775807, limit=0x0) at /ufs/manegold/_/scratch0/Monet/Testing/Stable/source/MonetDB/src/gdk/gdk_relop.mx:264
264 ptr nil = ATOMnilptr(r->htype);
Missing separate debuginfos, use: debuginfo-install bzip2-libs-1.0.5-6.fc12.x86_64 cyrus-sasl-lib-2.1.23-4.fc12.x86_64 glibc-2.11.1-1.x86_64 keyutils-libs-1.2-6.fc12.x86_64 krb5-libs-1.7-18.fc12.x86_64 libcom_err-1.41.9-5.fc12.x86_64 libcurl-7.19.7-6.fc12.x86_64 libidn-1.9-5.x86_64 libselinux-2.0.90-3.fc12.x86_64 libssh2-1.2.2-5.fc12.x86_64 libxml2-2.7.6-1.fc12.x86_64 libxslt-1.1.26-1.fc12.x86_64 ncurses-libs-5.7-3.20090207.fc12.x86_64 nspr-4.8.2-1.fc12.x86_64 nss-3.12.5-8.fc12.x86_64 nss-softokn-3.12.4-10.fc12.x86_64 nss-softokn-freebl-3.12.4-10.fc12.x86_64 nss-util-3.12.5-1.fc12.1.x86_64 openldap-2.4.19-1.fc12.x86_64 openssl-1.0.0-0.13.beta4.fc12.x86_64 pcre-7.8-3.fc12.x86_64 raptor-1.4.18-5.fc12.x86_64 readline-6.0-3.fc12.x86_64 sqlite-3.6.20-1.fc12.x86_64 zlib-1.2.3-23.fc12.x86_64
(gdb) up
1 0x00007ffff762e704 in batmergejoin (l=0x7fffc00008c8, r=0x7fffd00008f0, estimate=9223372036854775807, swap=1 '\001', limit=0x0) at /ufs/manegold/_/scratch0/Monet/Testing/Stable/source/MonetDB/src/gdk/gdk_relop.mx:410
410 BAT *bn = mergejoin(BATmirror(r), BATmirror(l), NULL, NULL, estimate, limit);
(gdb)
2 0x00007ffff766b6ee in batjoin (l=0x7fffc00008c8, r=0x7fffd00008f0, estimate=9223372036854775807, swap=1 '\001') at /ufs/manegold/_/scratch0/Monet/Testing/Stable/source/MonetDB/src/gdk/gdk_relop.mx:1230
1230 return batmergejoin(l, r, estimate, swap, NULL);
(gdb) li
1225 loop binary search (both implemented by BATmergejoin).
1226 @c
1227 if ((BATtordered(l) & BAThordered(r) & 1) || (must_hash && (((BATtordered(l) & 1) && ((lng) lcount > logl * (lng) rcount) && swap) || ((BAThordered(r) & 1) && ((lng) rcount > logr * (lng) lcount))))) {
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1228 ALGODEBUG THRprintf(GDKout, "BATjoin: BATmergejoin(l,r," BUNFMT ");\n", estimate);
1229
1230 return batmergejoin(l, r, estimate, swap, NULL);
1231 }
1232 @-
1233 hash join: the bread&butter join of monet
1234 @c
(gdb) print l->T->sorted &1
$1 = 1
(gdb) print r->H->sorted &1
$2 = 0
(gdb) print must_hash
$3 = 1
(gdb) print l->U->count
$4 = 26584
(gdb) print r->U->count
$5 = 999
(gdb) print logl
$6 = 19
(gdb) print logl * r->U->count
$7 = 18981
(gdb) print swap
$8 = 1 '\001'
(gdb) print logr
$9 = 14
(gdb)
## Comment 13041
Date: 2010-02-17 20:50:39 +0100
From: @drstmane
In a debugging session with Roberto, we located a bug in mergejoin where properties are set incorrectly.
I'll checkin a fix, once testing has confirmed it's correctness.
## Comment 13042
Date: 2010-02-18 02:38:00 +0100
From: @drstmane
fixed in CVS (Nov2009 branch) by fixing the property settings in mergejoin()
propagation to Feb2010 branch and development trunk pending
a proper test still needs to be added before closing this report
## Comment 13043
Date: 2010-02-18 09:58:39 +0100
From: @drstmane
added test in MonetDB5/tests/BugTracker/Tests/merge-join_properties.SF-2952191.*
## Comment 13044
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2952191 at http://sourceforge.net/support/tracker.php?aid=2952191
| GDK: merge-join produces invalid BAT | https://api.github.com/repos/MonetDB/MonetDB/issues/2323/comments | 0 | 2020-11-30T10:37:31Z | 2024-06-27T11:43:54Z | https://github.com/MonetDB/MonetDB/issues/2323 | 753,361,142 | 2,323 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-15 03:06:12 +0100
From: Henning Rode <<hrode>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
CC: @drstmane
Last updated: 2011-05-03 10:46:11 +0200
## Comment 13035
Date: 2010-02-15 15:06:12 +0100
From: Henning Rode <<hrode>>
a query using the * symbol to match all attribute namespaces like:
doc("/tmp/example.xml")//foo[@*:type="bar"]
causes the following parsing error:
!ERROR: Segmentation fault.
The Pathfinder compiler experienced an internal problem.
You may want to report this problem to the Pathfinder
development team (pathfinder@pathfinder-xquery.org).
When reporting problems, please attach your XQuery input,
as well as the following information:
Compilation stage: 2
We apologize for the inconvenience...
Segmentation fault (core dumped)
## Comment 13036
Date: 2010-02-15 15:47:08 +0100
From: @drstmane
Here's a quick analysis to locate the segfault.
$ cat /tmp/err.xq
doc("/tmp/example.xml")//foo[@*:type="bar"]
$ gdb --args pf /tmp/err.xq
GNU gdb (GDB) Fedora (7.0.1-31.fc12)
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /net/rig.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Candidate/prefix.--enable-strict_--disable-optimize_--enable-debug_--enable-assert/bin/pf...done.
(gdb) r
Starting program: /net/rig.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Candidate/prefix.--enable-strict_--disable-optimize_--enable-debug_--enable-assert/bin/pf /tmp/err.xq
Program received signal SIGSEGV, Segmentation fault.
0x000000000045644a in try_rewrite (stack=0x7fffffffb5a0, depth=4, curvar=0) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/pathfinder/compiler/semantics/heuristic.c:626
626 arg = args(lit_str(req_name->sem.qname_raw.loc),
Missing separate debuginfos, use: debuginfo-install glibc-2.11.1-1.x86_64 libxml2-2.7.6-1.fc12.x86_64 zlib-1.2.3-23.fc12.x86_64
(gdb) bt
0 0x000000000045644a in try_rewrite (stack=0x7fffffffb5a0, depth=4, curvar=0) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/pathfinder/compiler/semantics/heuristic.c:626
1 0x00000000004577e2 in PFheuristic_index (root=0x7ffff7bcba40) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/pathfinder/compiler/semantics/heuristic.c:740
2 0x000000000042a825 in PFcompile (url=0x7fffffffddd1 "/tmp/err.xq", pfout=0x33d7d73780, status=0x7fffffffd780) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/pathfinder/compiler/compile.c:457
3 0x0000000000429e91 in main (argc=2, argv=0x7fffffffd908) at /ufs/manegold/_/scratch0/Monet/Testing/Candidate/source/pathfinder/compiler/main.c:1079
(gdb) li
621 loc = LL(LL(h))->sem.qname_raw.loc;
622 prefix = LL(LL(h))->sem.qname_raw.prefix;
623 }
624 if (tst == TST_ATTR) {
625 /* pf:attribute() has additional parameters wrt pf:text() */
626 arg = args(lit_str(req_name->sem.qname_raw.loc),
627 args(lit_str(req_name->sem.qname_raw.prefix[0]?"*":""),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
628 args(lit_str(loc),
629 args(lit_str(prefix[0]?"*":""), arg))));
630 }
(gdb) print req_name
$1 = (PFpnode_t *) 0x7ffff7bcc100
(gdb) print *req_name
$2 = {kind = p_req_name, sem = {num = 0, dec = 0, dbl = 0, tru = false, str = 0x0, chr = 0 '\000', qname_raw = {prefix = 0x0, loc = 0x7ffff7bcb202 "type"}, qname = 0, axis = p_ancestor, kind = p_kind_node, mode = {dir = p_asc, empty = p_greatest, coll = 0x7ffff7bcb202 "type"}, oci = p_one, copy_ns = {
preserve = false, inherit = false}, empty = greatest, var = 0x0, fun = 0x0, pragma = {qn = {qname_raw = {prefix = 0x0, loc = 0x7ffff7bcb202 "type"}, qname = 0}, content = 0x0}, revalid = revalid_strict, insert = p_first_into}, child = {0x0, 0x0}, loc = {first_row = 1, first_col = 31, last_row = 1,
last_col = 37}, core = 0x0, state_label = 113}
(gdb) print req_name->sem
$3 = {num = 0, dec = 0, dbl = 0, tru = false, str = 0x0, chr = 0 '\000', qname_raw = {prefix = 0x0, loc = 0x7ffff7bcb202 "type"}, qname = 0, axis = p_ancestor, kind = p_kind_node, mode = {dir = p_asc, empty = p_greatest, coll = 0x7ffff7bcb202 "type"}, oci = p_one, copy_ns = {preserve = false, inherit = false},
empty = greatest, var = 0x0, fun = 0x0, pragma = {qn = {qname_raw = {prefix = 0x0, loc = 0x7ffff7bcb202 "type"}, qname = 0}, content = 0x0}, revalid = revalid_strict, insert = p_first_into}
(gdb) print req_name->sem.qname_raw
$4 = {prefix = 0x0, loc = 0x7ffff7bcb202 "type"}
(gdb) print req_name->sem.qname_raw.loc
$5 = 0x7ffff7bcb202 "type"
(gdb) print req_name->sem.qname_raw.prefix <=====================
$6 = 0x0 <=====================================================
(gdb) print loc
$7 = 0x6bb713 "*"
(gdb) print prefix
$8 = 0x6bb713 "*"
## Comment 13037
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2952088 at http://sourceforge.net/support/tracker.php?aid=2952088
## Comment 15783
Date: 2011-05-03 10:46:11 +0200
From: @drstmane
Development on MonetDB4 and MonetDB/XQuery has ceased. It is unlikely that
this bug will be fixed.
The last release of MonetDB/XQuery and MonetDB4 is Mar2011.
| attribute namespace | https://api.github.com/repos/MonetDB/MonetDB/issues/2322/comments | 0 | 2020-11-30T10:37:27Z | 2024-06-28T07:25:38Z | https://github.com/MonetDB/MonetDB/issues/2322 | 753,361,101 | 2,322 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-12 02:55:29 +0100
From: @swingbit
To: @njnes
Version: -- development
CC: @sjoerdmullender
Last updated: 2010-09-14 13:19:48 +0200
## Comment 13031
Date: 2010-02-12 14:55:29 +0100
From: @swingbit
In a
SELECT .... OFFSET <N>;
statement, N is limited to a 32bit positive integer (2147483647), regardless of whether the system is compiled 64bit or not.
Moreover, there seem to be a problem with the last allowed value (2147483647):
SELECT ... OFFSET 2147483647; -- never finishes
SELECT ... OFFSET 2147483646; -- returns instantly
## Comment 13032
Date: 2010-02-25 10:02:47 +0100
From: @njnes
added test to BugTracker-2010 offset_limited_32bit
## Comment 13033
Date: 2010-04-25 17:21:13 +0200
From: @njnes
a fix for this problem exists in the head branch.
## Comment 13034
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2950579 at http://sourceforge.net/support/tracker.php?aid=2950579
## Comment 14910
Date: 2010-09-14 13:19:48 +0200
From: @sjoerdmullender
This was fixed ages ago and the fixed version is in fact already released.
| SQL:OFFSET is limited to 32bit | https://api.github.com/repos/MonetDB/MonetDB/issues/2321/comments | 0 | 2020-11-30T10:37:24Z | 2024-06-27T11:43:52Z | https://github.com/MonetDB/MonetDB/issues/2321 | 753,361,065 | 2,321 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-12 11:13:40 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2010-05-06 08:28:55 +0200
## Comment 13028
Date: 2010-02-12 11:13:40 +0100
From: Wouter Alink <<alink>>
On both Nov2009 and Feb2010
I don't know whether the following is allowed, but I assume the query "SELECT * FROM x ORDER BY score" in the small example below should be possible if "SELECT * FROM x" is allowed.
$ mclient -lsql -dtest -umonetdb -Pmonetdb
Welcome to mclient, the MonetDB/SQL interactive terminal
Database: MonetDB v5.16.5, 'test'
Type \q to quit, \? for a list of available commands
auto commit mode: on
sql>start transaction;
auto commit mode: off
sql>CREATE VIEW x AS SELECT 1.0 as score, 'a' as avalue, 'b' AS displayname;
sql>SELECT * FROM x;
+-------+--------+-------------+
| score | avalue | displayname |
+=======+========+=============+
| 1 | a | b |
+-------+--------+-------------+
1 tuple
sql>SELECT * FROM x ORDER BY score;
TypeException:user.s1_1[2]:'algebra.sortTail' undefined in: _3:any := algebra.sortTail(_1:bte)
SQLException:SQLengine:Program contains errors
sql>^C
Actually the following one-liner triggers the same error:
SELECT * FROM (SELECT 1.0 as score) AS x ORDER BY x.score;
## Comment 13029
Date: 2010-02-25 09:34:34 +0100
From: @njnes
added test to BugTracker-2010 order_by_on_single_row_view
## Comment 13030
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2950453 at http://sourceforge.net/support/tracker.php?aid=2950453
## Comment 14004
Date: 2010-05-06 08:28:55 +0200
From: @njnes
This bug has been fixed some time ago
| SQL: order by on single row view | https://api.github.com/repos/MonetDB/MonetDB/issues/2320/comments | 0 | 2020-11-30T10:37:20Z | 2024-06-27T11:43:51Z | https://github.com/MonetDB/MonetDB/issues/2320 | 753,361,024 | 2,320 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-08 02:54:47 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2010-08-30 09:17:57 +0200
## Comment 13025
Date: 2010-02-08 14:54:47 +0100
From: Wouter Alink <<alink>>
Feb2010
given a table 'x' with tuples (id1 varchar(1000), id2 int)
and a table dict with tuples (id int, x varchar)
The following raises an assertion (i guess it should have been a SQL error that the types do not match):
UPDATE dict SET id = (SELECT id2 FROM "x" WHERE id1=dict.id);
gdb says:
Program received signal SIGABRT, Aborted.
[Switching to Thread 0x7f16b1caf710 (LWP 14596)]
0x00000039f62326c5 in raise () from /lib64/libc.so.6
(gdb) bt
0 0x00000039f62326c5 in raise () from /lib64/libc.so.6
1 0x00000039f6233ea5 in abort () from /lib64/libc.so.6
2 0x00000039f622b7b5 in __assert_fail () from /lib64/libc.so.6
3 0x00007f16b30fb867 in rel_select_exp (sql=0x7f16a4005518, rel=0x7f16a4dbb748, outer=0x7f16a4e41c78, sn=0x7f16a4125158, ek=...)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_select.mx:4372
4 0x00007f16b30fc6c8 in rel_query (sql=0x7f16a4005518, rel=0x7f16a4e67248, sq=0x7f16a4125158, toplevel=0, ek=...)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_select.mx:4607
5 0x00007f16b30fdb47 in rel_subquery (sql=0x7f16a4005518, rel=0x7f16a4e67248, sq=0x7f16a4125158, ek=...)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_select.mx:4950
6 0x00007f16b30fa428 in rel_value_exp2 (sql=0x7f16a4005518, rel=0x7f16b1caea18, se=0x7f16a4125158, f=3, ek=..., is_last=0x7f16b1cae9b4)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_select.mx:4007
7 0x00007f16b30fa6c3 in rel_value_exp (sql=0x7f16a4005518, rel=0x7f16b1caea18, se=0x7f16a4125158, f=3, ek=...)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_select.mx:4071
8 0x00007f16b30ff34c in update_table (sql=0x7f16a4005518, qname=0x7f16a4124b08, assignmentlist=0x7f16a4125268, opt_where=0x0)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_updates.mx:398
9 0x00007f16b3100991 in rel_updates (sql=0x7f16a4005518, s=0x7f16a4125368)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_updates.mx:744
10 0x00007f16b30ec44e in rel_semantic (sql=0x7f16a4005518, s=0x7f16a4125368)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_semantic.mx:242
11 0x00007f16b3054cdf in sql_symbol2stmt (c=0x7f16a4005518, sym=0x7f16a4125368)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql.mx:1277
12 0x00007f16b309f1ab in SQLparser (c=0x604e70)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1153
13 0x00007f16c9fd690b in runPhase (c=0x604e70, phase=1)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:602
14 0x00007f16c9fd6a62 in runScenarioBody (c=0x604e70) at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:645
15 0x00007f16c9fd6cbe in runScenario (c=0x604e70) at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:679
16 0x00007f16c9f8b334 in MSserveClient (dummy=0x604e70)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_session.mx:474
17 0x00000039f6e06a3a in start_thread () from /lib64/libpthread.so.0
18 0x00000039f62de67d in clone () from /lib64/libc.so.6
19 0x0000000000000000 in ?? ()
(gdb) frame 3
3 0x00007f16b30fb867 in rel_select_exp (sql=0x7f16a4005518, rel=0x7f16a4dbb748, outer=0x7f16a4e41c78, sn=0x7f16a4125158, ek=...)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/server/rel_select.mx:4372
4372 assert(re->type == e_column);
(gdb) l
4367 for (n = ce->h; n; n = n->next) {
4368 sql_exp *e = n->data;
4369 sql_exp *re = e->r;
4370
4371 assert(e->type == e_cmp);
4372 assert(re->type == e_column);
4373 re = exp_column(rel_name(inner), re->r, exp_subtype(re), inner->card, has_nil(re), is_intern(re));
4374 e = exp_compare( exp_dup(e->l), re, e->flag);
4375 append(jexps, e);
4376 }
(gdb)
## Comment 13026
Date: 2010-02-25 14:12:14 +0100
From: @njnes
test added to BugTracker-2010 correlated_update_type_mismatch.SF-2947841
## Comment 13027
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2947841 at http://sourceforge.net/support/tracker.php?aid=2947841
## Comment 14654
Date: 2010-08-12 19:54:53 +0200
From: @njnes
added test BugTracker-2010/Tests/assert_on_type_mismatch.Bug-2319.sql
## Comment 14655
Date: 2010-08-12 19:54:57 +0200
From: @njnes
Changeset [d343bf31c576](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=d343bf31c576) made by Niels Nes <niels@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=d343bf31c576](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=d343bf31c576)
Changeset description:
added test for (long) fixed bug #2319.
## Comment 14784
Date: 2010-08-30 09:17:57 +0200
From: @sjoerdmullender
The Jun2010-SP2 version has been released.
| SQL: got assertion for type-mismatch | https://api.github.com/repos/MonetDB/MonetDB/issues/2319/comments | 0 | 2020-11-30T10:37:17Z | 2024-06-27T11:43:50Z | https://github.com/MonetDB/MonetDB/issues/2319 | 753,360,993 | 2,319 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-08 12:07:02 +0100
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2010-04-14 03:29:24 +0200
## Comment 13018
Date: 2010-02-08 12:07:02 +0100
From: Wouter Alink <<alink>>
Feb2010
When running a script which creates and fills a database exactly the same way each time, sometimes (but not every time) I get a segfault on a specific point in the SQL script. It is always in the following query:
INSERT INTO x_1 SELECT a.x as idstr " +
"FROM x as a LEFT JOIN x_1 as b ON a.x=b.idstr " +
"WHERE isnull(b.idstr);
The query basically inserts values into a table which are not yet in the table.
It only seems to happens when x_1 is not empty anymore. (perhaps interesting detail is that the stringheap of x_1 is probably close to 128k in size)
gdb says:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f498ce12710 (LWP 1817)]
0x00007f49a47039eb in strPut (h=0x7f49805332a8, dst=0x7f498ce0ef80, v=0x7f49740220a0 "wherea")
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB/src/gdk/gdk_atoms.mx:1986
1986 if (GDK_STRCMP(v, h->base + pos) == 0) {
(gdb) l
1981 if (pad < sizeof(stridx_t))
1982 pad += GDK_VARALIGN; /* if not, pad more */
1983 } else if (*bucket) {
1984 /* large string heap (>=64KB) -- opportunistic/probabilistic double elimination */
1985 pos = elimbase + *bucket + extralen;
1986 if (GDK_STRCMP(v, h->base + pos) == 0) {
1987 return *dst = (var_t) (pos >> GDK_VARSHIFT); /* already in heap; do not insert! */
1988 }
1989 if SIZEOF_VAR_T < SIZEOF_VOID_P
1990 pad &= (GDK_VARALIGN-1); /* a full VARALIGN pad can be omitted */
(gdb) bt
0 0x00007f49a47039eb in strPut (h=0x7f49805332a8, dst=0x7f498ce0ef80, v=0x7f49740220a0 "wherea")
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB/src/gdk/gdk_atoms.mx:1986
1 0x00007f49a44e1359 in BATins (b=0x7f498053a838, n=0x7f49740073a8, force=0 '\000')
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB/src/gdk/gdk_batop.mx:407
2 0x00007f4991f952eb in MATpackInternal (stk=0x7f49805a8e08, p=0x7f498053ec08)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/modules/mal/mat.mx:181
3 0x00007f4991f95629 in MATpack (cntxt=0x604e70, mb=0x7f4980534d78, stk=0x7f49805a8e08, p=0x7f498053ec08)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/modules/mal/mat.mx:241
4 0x00007f49a50f9e2d in runMALsequence (cntxt=0x604e70, mb=0x7f4980534d78, startpc=1, stoppc=0, stk=0x7f49805a8e08, env=0x0, pcicaller=0x0)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_interpreter.mx:1904
5 0x00007f49a50f06e2 in callMAL (cntxt=0x604e70, mb=0x7f4980534d78, env=0x7f498ce11b88, argv=0x7f498ce11bd0, debug=0 '\000')
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_interpreter.mx:400
6 0x00007f498e202b2a in SQLexecutePrepared (c=0x604e70, be=0x7f4980039728, q=0x7f4980562548)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1327
7 0x00007f498e202da9 in SQLengineIntern (c=0x604e70, be=0x7f4980039728)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1380
8 0x00007f498e2032e4 in SQLengine (c=0x604e70)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1490
9 0x00007f49a513990b in runPhase (c=0x604e70, phase=4)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:602
10 0x00007f49a5139b01 in runScenarioBody (c=0x604e70) at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:654
11 0x00007f49a5139cbe in runScenario (c=0x604e70) at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:679
12 0x00007f49a50ee334 in MSserveClient (dummy=0x604e70)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_session.mx:474
13 0x00000039f6e06a3a in start_thread () from /lib64/libpthread.so.0
14 0x00000039f62de67d in clone () from /lib64/libc.so.6
15 0x0000000000000000 in ?? ()
(gdb) p h->base
$1 = 0x7f4980631008 "\260\257"
(gdb) p pos
$2 = 303315
(gdb) p v
$3 = (str) 0x7f49740220a0 "wherea"
(gdb)
## Comment 13019
Date: 2010-02-08 13:38:10 +0100
From: Wouter Alink <<alink>>
added script and data to reproduce (seems to fail every time, but that might be timing related, because when running it from java environment the same script sometimes doesn't crash mserver)
$ sh doCrash.sh
stopping database 'testbug'... FAILED
stop: database is not running: testbug
destroyed database: testbug
created database in maintenance mode: testbug
taken database out of maintenance mode: testbug
starting database 'testbug'... done
MAPI = monetdb@localhost:51234
ACTION= read_line
QUERY = START TRANSACTION;
CREATE TABLE "sys"."temp_27_TERM" (
"termid" int NOT NULL,
"term" varchar(1000),
"prob" double,
CONSTRAINT "temp_27_TERM_termid_pkey" PRIMARY KEY ("termid")
);
COPY 2222 RECORDS INTO "sys"."temp_27_TERM" FROM stdin USING DELIMITERS '\t','\n','"';
0 "STOPWORD" 1
1 "a" 1
2 "about" 1
3 "above" 1
4 "across" 1
5 "after" 1
6 "again" 1
7 "against" 1
8 "all" 1
9 "almost" 1
10 "alone" 1
11 "along" 1
12 "already" 1
13 "also" 1
14 "although" 1
15 "always" 1
16 "among" 1
17 "an" 1
ERROR = !Connection terminated
## Comment 13020
Date: 2010-02-09 12:35:10 +0100
From: Wouter Alink <<alink>>
It seems to also occur in the Nov2009 branch. (the attached script does the job)
## Comment 13021
Date: 2010-04-14 11:41:37 +0200
From: Wouter Alink <<alink>>
Is there any news on this issue? The attached script still crashes the database with the current Feb2010 (last week's checkout).
## Comment 13022
Date: 2010-04-14 12:45:09 +0200
From: @sjoerdmullender
I was able to reproduce the crash. The values are slightly different, but the location is the same.
First impression is that h->base+pos is outside of the area allocated for h->base (i.e. pos > h->maxsize). That this gives a crash isn't strange.
In my case, I have these values:
(gdb) p pos
$3 = 283936
(gdb) p elimbase
$5 = 262144
(gdb) p *bucket
$6 = 21784
(gdb) p extralen
$7 = 8
(gdb) p *h
$8 = {maxsize = 270336, free = 265561, size = 270336,
base = 0x7f225041a3b8 "hQ", filename = 0x7f22503a8848 "07/731.theap",
storage = 0 '\000', copied = 0 '\000', hashash = 1 '\001',
newstorage = 0 '\000', dirty = 0 '\000'}
So elimbase looks good, but *bucket might be too large. I need to investigate this more.
(extralen==8 is to be expected, it means that hashes are stored in the heap.)
## Comment 13023
Date: 2010-04-14 15:29:23 +0200
From: @sjoerdmullender
Fixed by clearing the duplicate elimination hash table in gdk_batop.mx when the heap is extended when one string bat is inserted into another. So this wasn't really an SQL bug but a GDK bug.
Test added in src/test/BugTracker-2010 heap_combine_crash.SF-2947763.
## Comment 13024
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2947763 at http://sourceforge.net/support/tracker.php?aid=2947763
| SQL: insert into + select on same table | https://api.github.com/repos/MonetDB/MonetDB/issues/2318/comments | 0 | 2020-11-30T10:37:14Z | 2024-06-27T11:43:50Z | https://github.com/MonetDB/MonetDB/issues/2318 | 753,360,948 | 2,318 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-08 11:16:09 +0100
From: Wouter Alink <<alink>>
To: GDK devs <<bugs-common>>
Version: 11.11.5 (Jul2012)
CC: @mlkersten, y.kargin
Last updated: 2014-08-11 14:36:32 +0200
## Comment 13015
Date: 2010-02-08 11:16:09 +0100
From: Wouter Alink <<alink>>
Using the default_pipe with the 'constants' optimizer, the following statement crashes mserver. (Feb2010, checkout of last week):
CREATE TABLE x (id SERIAL, a VARCHAR(1000));
Leaving the 'constants' out of the pipeline, it seems to work fine.
This is what gdb tells me:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f3227a1a710 (LWP 19018)]
0x00007f3228dc069b in create_table_or_view (sql=0x7f3218005518, sname=0x7f3218054538 "sys", t=0x7f32180bd898, temp=0)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql.mx:1460
1460 sql_column *c = n->data;
(gdb) l
1455 return sql_message("CREATE TABLE: insufficient privileges for user '%s' in schema '%s'", stack_get_string(sql, "current_user"), s->base.name);
1456 }
1457
1458 /* first check default values */
1459 for (n = t->columns.set->h; n; n = n->next) {
1460 sql_column *c = n->data;
1461
1462 if (c->def) {
1463 char buf[BUFSIZ];
1464 sql_rel *r = NULL;
(gdb) p n
$1 = (node *) 0x135
(gdb) p n->data
Cannot access memory at address 0x13d
(gdb) p *n
Cannot access memory at address 0x135
(gdb) p t->columns.set
$2 = (struct list *) 0x7f321804d388
(gdb) p *t->columns.set
$3 = {sa = 0x7f3218000078, destroy = 0x20, h = 0x135, t = 0x108, cnt = 402653722}
(gdb) bt
0 0x00007f3228dc069b in create_table_or_view (sql=0x7f3218005518, sname=0x7f3218054538 "sys", t=0x7f32180bd898, temp=0)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql.mx:1460
1 0x00007f3228dc1f4a in SQLcatalog (cntxt=0x604e70, mb=0x7f321806fac8, stk=0x7f3218080d98, pci=0x7f32180818b8)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql.mx:1821
2 0x00007f323fd01e2d in runMALsequence (cntxt=0x604e70, mb=0x7f321806fac8, startpc=1, stoppc=0, stk=0x7f3218080d98, env=0x0, pcicaller=0x0)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_interpreter.mx:1904
3 0x00007f323fcf86e2 in callMAL (cntxt=0x604e70, mb=0x7f321806fac8, env=0x7f3227a19b88, argv=0x7f3227a19bd0, debug=0 '\000')
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_interpreter.mx:400
4 0x00007f3228e0ab2a in SQLexecutePrepared (c=0x604e70, be=0x7f321803fdc8, q=0x7f321804bd18)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1327
5 0x00007f3228e0ada9 in SQLengineIntern (c=0x604e70, be=0x7f321803fdc8)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1380
6 0x00007f3228e0b2e4 in SQLengine (c=0x604e70)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/sql/src/backends/monet5/sql_scenario.mx:1490
7 0x00007f323fd4190b in runPhase (c=0x604e70, phase=4)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:602
8 0x00007f323fd41b01 in runScenarioBody (c=0x604e70) at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:654
9 0x00007f323fd41cbe in runScenario (c=0x604e70) at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_scenario.mx:679
10 0x00007f323fcf6334 in MSserveClient (dummy=0x604e70)
at /export/scratch0/alink/workspace/MonetDB-installer/MonetDB-Feb2010/MonetDB5/src/mal/mal_session.mx:474
11 0x00000039f6e06a3a in start_thread () from /lib64/libpthread.so.0
12 0x00000039f62de67d in clone () from /lib64/libc.so.6
13 0x0000000000000000 in ?? ()
(gdb) q
A debugging session is active.
Inferior 1 [process 18984] will be detached.
## Comment 13016
Date: 2010-03-18 00:11:34 +0100
From: @drstmane
FYI:
===================================================================
2010/01/24 - nielsnes: MonetDB5/conf/monetdb5.conf.in,1.82.2.1
disable constant optimizer (breaks atom ptr values)
===================================================================
removed the constants optimizer from the mitosis_pipe which was back then used as default,
and
===================================================================
2010/02/19 - mlkersten: MonetDB5/conf/monetdb5.conf.in,1.82.2.2
Lets declare the new default pipeline with a fall-back position.
===================================================================
renamed the "mitosis_pipe" (without "constants" optimizer) into "default_pipe" and uses that as default,
keeping the former "default_pipe" as "nov2009_pipe" (incl. the "constants" optimizer).
## Comment 13017
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2947749 at http://sourceforge.net/support/tracker.php?aid=2947749
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
## Comment 15790
Date: 2011-05-09 16:13:56 +0200
From: @mlkersten
This problem seems to have disappeared in the default branch.
Test added as BugTracker-2010/Tests/constants-optimizer.Bug-2317.sql
## Comment 16194
Date: 2011-09-08 11:37:57 +0200
From: @sjoerdmullender
(In reply to comment 3)
> This problem seems to have disappeared in the default branch.
> Test added as BugTracker-2010/Tests/constants-optimizer.Bug-2317.sql
The problem has not disappeared. If you add the constants optimizer to the default pipe line, the server crashes:
sql>set optimizer='inline,remap,evaluate,costModel,coercions,emptySet,aliases,mitosis,mergetable,deadcode,constants,commonTerms,joinPath,reorder,deadcode,reduce,dataflow,history,multiplex,accumulators,garbageCollector';
sql>CREATE TABLE x (id SERIAL, a VARCHAR(1000));
MAPI = (monetdb) /tmp/.s.monetdb.50012
ACTION= read_line
QUERY = CREATE TABLE x (id SERIAL, a VARCHAR(1000));
ERROR = !Connection terminated
## Comment 17487
Date: 2012-07-17 14:37:04 +0200
From: @grobian
the constants optimizer is still in the accumulator_pipe, flagged as stable.
## Comment 18150
Date: 2012-11-27 16:31:52 +0100
From: @mlkersten
Expanded the test with accumulator_pipe, which contains the constant optimizer.
The result does not trigger a segfault.
## Comment 20021
Date: 2014-08-11 14:36:32 +0200
From: MonetDB Mercurial Repository <<hg>>
Changeset [f62a6432981c](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=f62a6432981c) made by Stefan Manegold <Stefan.Manegold@cwi.nl> in the MonetDB repo, refers to this bug.
For complete details, see [http//devmonetdborg/hg/MonetDB?cmd=changeset;node=f62a6432981c](https://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=http//devmonetdborg/hg/MonetDB?cmd=changeset;node=f62a6432981c)
Changeset description:
constants-optimizer.Bug-2317: clean-up (supposely ignored parts of) stable.err
as an attempt to avoid "false positive" diffs
| constants optimizer | https://api.github.com/repos/MonetDB/MonetDB/issues/2317/comments | 0 | 2020-11-30T10:37:11Z | 2024-06-27T11:43:49Z | https://github.com/MonetDB/MonetDB/issues/2317 | 753,360,902 | 2,317 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-08 11:14:10 +0100
From: Wouter Alink <<alink>>
To: @mlkersten
Version: -- development
Last updated: 2010-03-18 12:05:00 +0100
## Comment 13012
Date: 2010-02-08 11:14:10 +0100
From: Wouter Alink <<alink>>
The default_pipe optimizer pipe is not the default optimizer pipe (currently it is mitosis). This is misleading.
## Comment 13013
Date: 2010-03-18 00:04:59 +0100
From: @drstmane
mitosis has become the default and the mitosis_pipe has been renamed to default_pipe:
===================================================================
2010/02/19 - mlkersten: MonetDB5/conf/monetdb5.conf.in,1.82.2.2
Lets declare the new default pipeline with a fall-back position.
===================================================================
## Comment 13014
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2947747 at http://sourceforge.net/support/tracker.php?aid=2947747
| name default_pipe is misleading | https://api.github.com/repos/MonetDB/MonetDB/issues/2316/comments | 0 | 2020-11-30T10:37:08Z | 2024-06-27T11:43:48Z | https://github.com/MonetDB/MonetDB/issues/2316 | 753,360,864 | 2,316 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-02-02 10:33:55 +0100
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2010-02-05 04:47:38 +0100
## Comment 13008
Date: 2010-02-02 10:33:55 +0100
From: Wouter Alink <<alink>>
On Feb2010 (yesterday's CVS checkout) it seems that CREATE TABLE .... AS SELECT doesn't work properly. See example below. I would have expected two columns as output for the second SELECT clause, preferably with the columnnames "a1" and "prob".
sql>START TRANSACTION;
auto commit mode: off
sql>CREATE VIEW selectTerm_term_id AS SELECT docid as a1, prob as prob from docdict;
operation successful
sql>select * from selectTerm_term_id limit 10;
+------+------------------------+
| a1 | prob |
+======+========================+
| 0 | 1 |
| 1 | 1 |
| 2 | 1 |
| 3 | 1 |
| 4 | 1 |
| 5 | 1 |
| 6 | 1 |
| 7 | 1 |
| 8 | 1 |
| 9 | 1 |
+------+------------------------+
10 tuples
sql>CREATE TABLE selectTerm_term_id_top AS SELECT ALL a1, prob FROM selectTerm_term_id ORDER BY prob DESC LIMIT 26 WITH DATA;
sql>SELECT * FROM selectTerm_term_id_top LIMIT 10;
+------+
| v |
+======+
| 0 |
| 1 |
| 2 |
| 3 |
| 4 |
| 5 |
| 6 |
| 7 |
| 8 |
| 9 |
+------+
10 tuples
sql>
## Comment 13009
Date: 2010-02-05 16:15:30 +0100
From: @njnes
fixed (topn wasn't handled properly). Waiting for a nice test to close the bug report
## Comment 13010
Date: 2010-02-05 16:47:38 +0100
From: Wouter Alink <<alink>>
test added: sql/src/test/BugTracker-2010/Tests/Attic/tableDataLimit.SF-2944504
output is ok. closing bug.
## Comment 13011
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2944504 at http://sourceforge.net/support/tracker.php?aid=2944504
| SQL: "create table ... as select..." fails | https://api.github.com/repos/MonetDB/MonetDB/issues/2315/comments | 0 | 2020-11-30T10:37:00Z | 2024-06-27T11:43:47Z | https://github.com/MonetDB/MonetDB/issues/2315 | 753,360,755 | 2,315 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-01-31 10:37:10 +0100
From: @skinkie
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2010-01-31 10:48:34 +0100
## Comment 13005
Date: 2010-01-31 22:37:10 +0100
From: @skinkie
What is the rationale in rendering of the following:
sql>select -11.000000000000001;
+---------------------+
| single_value |
+=====================+
| -11.000000000000001 |
+---------------------+
1 tuple
sql>select 11.000000000000001;
+------------------------+
| single_value |
+========================+
| 11.000000000000002 |
+------------------------+
While using a float or double direct results in:
sql>select cast (-11.000000000000001 as float);
+------------------------+
| single_value |
+========================+
| -11.000000000000002 |
+------------------------+
1 tuple
## Comment 13006
Date: 2010-01-31 22:48:34 +0100
From: @sjoerdmullender
Floating point precision.
## Comment 13007
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2943398 at http://sourceforge.net/support/tracker.php?aid=2943398
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Rendering negative fraction | https://api.github.com/repos/MonetDB/MonetDB/issues/2314/comments | 0 | 2020-11-30T10:36:58Z | 2024-06-27T11:43:46Z | https://github.com/MonetDB/MonetDB/issues/2314 | 753,360,723 | 2,314 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-01-31 10:31:40 +0100
From: Alex Bo <<alexbod>>
To: @sjoerdmullender
Version: -- development
Last updated: 2010-02-03 01:54:29 +0100
## Comment 12999
Date: 2010-01-31 22:31:40 +0100
From: Alex Bo <<alexbod>>
Building the MonetDB5 module of the Feb2010 branch yields these errors:
.\..\..\..\..\src\modules\kernel\group.mx(1330) : warning C4267: '=' : conversion from 'size_t' to 'var_t', possible loss of data
.\..\..\..\..\src\modules\kernel\group.mx(1330) : warning C4267: '=' : conversion from 'size_t' to 'var_t', possible loss of data
## Comment 13000
Date: 2010-01-31 22:45:05 +0100
From: @sjoerdmullender
Looking at the code, a cast to (var_t) would be an easy solution, but it is *wrong*.
What I think is the right solution, also given that inside sort_flush we do a <<GDK_VARSHIFT on the value in the offp table, is to use VarHeapValRaw instead of VarHeapVal. VarHeapValRaw returns the unshifted value from the offset heap, and VarHeapVal returns the shifted value. Since the shift seems to happen in sort_flush, we should use the unshifted value here, I think.
After my checkin to gdk.mx, the types should also be correct.
## Comment 13001
Date: 2010-01-31 22:46:29 +0100
From: @sjoerdmullender
A quick note why the cast to var_t is wrong. The value may not fit into a var_t! On a 64-bit system with 32-bit OIDs, the result of VarHeapVal can be larger than what fits into a var_t.
## Comment 13002
Date: 2010-01-31 22:59:16 +0100
From: @sjoerdmullender
I checked in a fix, but I'll wait until it has been better tested before closing this report.
## Comment 13003
Date: 2010-02-03 13:54:29 +0100
From: @sjoerdmullender
Testing doesn't show worse behavior, so it seems the bug was fixed properly.
No test since this was a compilation issue.
## Comment 13004
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2943395 at http://sourceforge.net/support/tracker.php?aid=2943395
| Can't build MonetDB5 on Windows | https://api.github.com/repos/MonetDB/MonetDB/issues/2313/comments | 0 | 2020-11-30T10:36:55Z | 2024-06-27T11:43:45Z | https://github.com/MonetDB/MonetDB/issues/2313 | 753,360,681 | 2,313 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-01-30 10:18:32 +0100
From: zeuner <<zeuner>>
To: @njnes
Version: -- development
Last updated: 2010-03-23 06:40:26 +0100
## Comment 12995
Date: 2010-01-30 10:18:32 +0100
From: zeuner <<zeuner>>
I am unable to trigger this issue using the CLI which is why I attach a C program in this report.
The symptom is a Mapi error message like this (the numbers vary):
ERROR = !MALException:!ERROR: BATpropcheck: BAT tmpr_1002(-514) with 19 tuples seqbase of dense oid bat is wrong! 0 != 18
The issue has been verified on a 2010-01-26 and a 2010-01-30 checkout on an x86_64 dual core running CentOS 5.4.
## Comment 12996
Date: 2010-03-18 00:20:29 +0100
From: @drstmane
This might (have be) the same problem as reported in
ID: 2952191 "GDK: merge-join produces invalid BAT"
https://sourceforge.net/tracker/?func=detail&aid=2952191&group_id=56967&atid=482468
This has been fixed on Feb 18.
Could you please check with the Feb2010 release (or an up-to-date checkout of the Feb2010 CVS branches), whether your problem still occurs or not?
Thanks!
## Comment 12997
Date: 2010-03-23 06:31:34 +0100
From: zeuner <<zeuner>>
On a 2010-03-22 Feb2010 checkout, the issue seems to be gone.
## Comment 12998
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2942744 at http://sourceforge.net/support/tracker.php?aid=2942744
| BATpropcheck fails on C Mapi usage | https://api.github.com/repos/MonetDB/MonetDB/issues/2312/comments | 0 | 2020-11-30T10:36:51Z | 2024-06-27T11:43:43Z | https://github.com/MonetDB/MonetDB/issues/2312 | 753,360,631 | 2,312 |
[
"MonetDB",
"MonetDB"
] | Date: 2010-01-30 10:02:42 +0100
From: zeuner <<zeuner>>
To: @njnes
Version: -- development
Last updated: 2010-01-30 10:51:49 +0100
## Comment 12992
Date: 2010-01-30 10:02:42 +0100
From: zeuner <<zeuner>>
The SQL statements in the attached file create a failed assertion for me:
mserver5: bat_storage.mx:1274: tr_update_delta: Assertion `obat->ibase == cbat->ibase' failed.
The issue has been verified on a 2010-01-26 and a 2010-01-30 checkout on an x86_64 dual core running CentOS 5.4.
I did some tracing on the ibase member variable. Both obat->ibase and cbat->ibase gets initialized through new_persistent_delta (called by create_col), and both are updated through a previous tr_update_delta invocation. But while cbat gets the value passed through dup_delta (called by dup_col), it gets zeroed by clear_delta (called by clear_col) for obat. However, I am not sure how the value is supposed to get updated throughout a transaction.
## Comment 12993
Date: 2010-01-30 10:51:48 +0100
From: @njnes
incorrect assert is removed. No test needed in this case.
## Comment 12994
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 2942739 at http://sourceforge.net/support/tracker.php?aid=2942739
| ibase mismatch in tr_update_delta (bat_storage.mx) | https://api.github.com/repos/MonetDB/MonetDB/issues/2311/comments | 0 | 2020-11-30T10:36:48Z | 2024-06-27T11:43:43Z | https://github.com/MonetDB/MonetDB/issues/2311 | 753,360,579 | 2,311 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.