repo
stringclasses
1 value
pull_number
int64
20
3.81k
instance_id
stringlengths
18
20
issue_numbers
listlengths
1
2
base_commit
stringlengths
40
40
patch
stringlengths
214
153k
test_patch
stringlengths
288
262k
problem_statement
stringlengths
22
7.42k
hints_text
stringlengths
0
66.9k
created_at
stringlengths
20
20
version
stringclasses
21 values
environment_setup_commit
stringclasses
21 values
hyperium/hyper
1,674
hyperium__hyper-1674
[ "1517" ]
8bfe3c220c450d6a3d049fab4d7a591fa2ea0d82
diff --git a/src/client/connect/dns.rs b/src/client/connect/dns.rs --- a/src/client/connect/dns.rs +++ b/src/client/connect/dns.rs @@ -1,45 +1,183 @@ -use std::io; +use std::{fmt, io, vec}; use std::net::{ - Ipv4Addr, Ipv6Addr, + IpAddr, Ipv4Addr, Ipv6Addr, SocketAddr, ToSocketAddrs, SocketAddrV4, Sock...
diff --git a/src/client/connect/dns.rs b/src/client/connect/dns.rs --- a/src/client/connect/dns.rs +++ b/src/client/connect/dns.rs @@ -77,6 +215,30 @@ impl Iterator for IpAddrs { } } +// Make this Future unnameable outside of this crate. +pub(super) mod sealed { + use super::*; + // Blocking task to be ex...
Substitute DNS Resolver This relates to #1174. It would be great if we could substitute a resolver instead of using `to_socket_addrs` and `getaddrinfo`. It seems there was already a good amount of discussion around this in #1174, but I'm not sure what the conclusion was. It seems like there are three options: ...
The `ClientProto` stuff is long gone, nothing to worry about there :) I've seen others ask to replace the resolve in `HttpConnector` also, and it would be nice to be do so. The *only* thing that has held me back is that it would be another trait to need to stabilize. I don't have a good design for one, and haven't s...
2018-10-17T23:07:16Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,662
hyperium__hyper-1662
[ "1661" ]
0cfa7231b2483715e3bd886a059e2792c090e635
diff --git a/src/client/connect/mod.rs b/src/client/connect/mod.rs --- a/src/client/connect/mod.rs +++ b/src/client/connect/mod.rs @@ -8,7 +8,7 @@ use std::error::Error as StdError; use std::mem; -use bytes::{BufMut, BytesMut}; +use bytes::{BufMut, Bytes, BytesMut}; use futures::Future; use http::{uri, Uri}; use...
diff --git a/src/client/connect/mod.rs b/src/client/connect/mod.rs --- a/src/client/connect/mod.rs +++ b/src/client/connect/mod.rs @@ -305,6 +312,30 @@ mod tests { assert_eq!(dst.host(), "seanmonstar.com", "error doesn't modify dst"); assert_eq!(dst.port(), None, "error doesn't modify dst"); + ...
Calling Destination::set_host with an IPv6 string errors
2018-09-27T22:16:59Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,643
hyperium__hyper-1643
[ "1642" ]
168c7d2155952ba09f781c331fd67593b820af20
diff --git a/src/proto/h2/client.rs b/src/proto/h2/client.rs --- a/src/proto/h2/client.rs +++ b/src/proto/h2/client.rs @@ -108,7 +108,7 @@ where } let (head, body) = req.into_parts(); let mut req = ::http::Request::from_parts(head, (...
diff --git a/src/proto/h2/mod.rs b/src/proto/h2/mod.rs --- a/src/proto/h2/mod.rs +++ b/src/proto/h2/mod.rs @@ -15,15 +15,17 @@ mod server; pub(crate) use self::client::Client; pub(crate) use self::server::Server; -fn strip_connection_headers(headers: &mut HeaderMap) { +fn strip_connection_headers(headers: &mut Head...
TE "trailers" header incorrectly stripped from HTTP/2 requests According to the HTTP/2 spec (https://http2.github.io/http2-spec/#rfc.section.8.1.2.2), connection-specific headers are not supposed to be in HTTP/2 messages, but a TE header is allowed for request messages as long as it only contains the value "trailers". ...
Yep, looks like hyper is being too eager here.
2018-08-25T07:58:17Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,641
hyperium__hyper-1641
[ "1639" ]
1448e4067b10da6fe4584921314afc1f5f4e3c8d
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -753,8 +753,7 @@ pub struct Builder { keep_alive_timeout: Option<Duration>, h1_writev: bool, h1_title_case_headers: bool, - //TODO: make use of max_idle config - max_idle: usize, + max_idle_per_...
diff --git a/src/client/pool.rs b/src/client/pool.rs --- a/src/client/pool.rs +++ b/src/client/pool.rs @@ -773,7 +790,16 @@ mod tests { } fn pool_no_timer<T>() -> Pool<T> { - let pool = Pool::new(true, Some(Duration::from_millis(100)), &Exec::Default); + pool_max_idle_no_timer(::std::usize::MA...
Enable max idle Client Pool option There is currently a `max_idle` option for the `Client` connection pool, that claims it is the maximum number of idle connections that will be kept per host, but it's a lie. That config is currently ignored. It should be plugged into the pool so that each time a connection is ready...
2018-08-21T23:06:58Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,596
hyperium__hyper-1596
[ "1595" ]
396e6022e02bdbcabced5235528cb61bee29e3b7
diff --git a/examples/client_json.rs b/examples/client_json.rs --- a/examples/client_json.rs +++ b/examples/client_json.rs @@ -11,14 +11,32 @@ use hyper::rt::{self, Future, Stream}; fn main() { let url = "http://jsonplaceholder.typicode.com/users".parse().unwrap(); + let fut = fetch_json(url) + // use...
diff --git a/examples/web_api.rs b/examples/web_api.rs --- a/examples/web_api.rs +++ b/examples/web_api.rs @@ -27,32 +28,65 @@ fn response_examples(req: Request<Body>, client: &Client<HttpConnector>) }, (&Method::GET, "/test.html") => { // Run a web query against the web api below + + ...
JSON server example Would it be nice to have example for server responding with json? I think it's common use case (especially useful when API frequently change) It can be like this: (what I came up with) ```rust use futures::{future, Future}; use hyper::{self, Method, StatusCode, Request, Response, Body, Ser...
Perhaps this could be added to the web_api example that already exists? That probably could use some cleanup, and some comments explaining what each part does... But it otherwise seems like a good fit for this.
2018-07-06T15:47:17Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,594
hyperium__hyper-1594
[ "1402" ]
ced949cb6b798f25c2ffbdb3ebda6858c18393a7
diff --git a/examples/hello.rs b/examples/hello.rs --- a/examples/hello.rs +++ b/examples/hello.rs @@ -10,7 +10,6 @@ static PHRASE: &'static [u8] = b"Hello World!"; fn main() { pretty_env_logger::init(); - let addr = ([127, 0, 0, 1], 3000).into(); // new_service is run for each connection, creating ...
diff --git a/src/client/connect/mod.rs b/src/client/connect/mod.rs --- a/src/client/connect/mod.rs +++ b/src/client/connect/mod.rs @@ -268,6 +278,61 @@ impl Connected { self } */ + + // Don't public expose that `Connected` is `Clone`, unsure if we want to + // keep that contract... + pub(sup...
Get socket parameters in Client or Server When a `Client` makes a request, it may connect to one of many hosts associated with the URL. I would like to have a way to access the IP address associated with the remote host. This is accessible from tokio via `NetSocket::peer_addr()`. Logically, one should also be abl...
I agree there are cases where this is desirable, but there are also reasons it doesn't exist yet. These are properties specific to TCP, whereas hyper (and HTTP even) don't require the underlying transport to be TCP (Unix domain sockets don't have a concept of addresses, for example). It'd probably need some sort of ...
2018-07-05T18:54:23Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,572
hyperium__hyper-1572
[ "1564" ]
e4ebf4482372497f0cd5fa63d6492bd1c51d7b21
diff --git a/src/client/connect.rs b/src/client/connect.rs --- a/src/client/connect.rs +++ b/src/client/connect.rs @@ -6,9 +6,11 @@ //! establishes connections over TCP. //! - The [`Connect`](Connect) trait and related types to build custom connectors. use std::error::Error as StdError; +use std::mem; +use bytes...
diff --git a/src/client/connect.rs b/src/client/connect.rs --- a/src/client/connect.rs +++ b/src/client/connect.rs @@ -121,6 +261,121 @@ impl Connected { */ } +#[cfg(test)] +mod tests { + use super::Destination; + + #[test] + fn test_destination_set_scheme() { + let mut dst = Destination { + ...
Can't create or modify hyper::client::connect::Destination I'm trying to write a connector that takes a destination, modifies it and passes it to the next connector. It seems this is currently not possible because there's no way to get a mutable reference to the internal state and only hyper can create new `Destination...
I definitely get the use case! I'd like to find a way to support it, but I'd also like to list why the constructor was so far private: - The `Destination` type *may* gain more fields, such as in #1485 to support ALPN. - I worried that if a `Destination` gained more fields in new release, and a connector didn't know...
2018-06-18T21:21:52Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,563
hyperium__hyper-1563
[ "1395" ]
1c3fbfd6bf6b627f75ef694e69c8074745276e9b
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -29,7 +29,7 @@ iovec = "0.1" log = "0.4" net2 = { version = "0.2.32", optional = true } time = "0.1" -tokio = { version = "0.1.5", optional = true } +tokio = { version = "0.1.7", optional = true } tokio-executor = { version = "0.1.0", optiona...
diff --git a/src/proto/h1/conn.rs b/src/proto/h1/conn.rs --- a/src/proto/h1/conn.rs +++ b/src/proto/h1/conn.rs @@ -612,6 +598,10 @@ where I: AsyncRead + AsyncWrite, } } + pub(super) fn on_upgrade(&mut self) -> ::upgrade::OnUpgrade { + self.state.prepare_upgrade() + } + // Used in h1::d...
Client Protocol Upgrades (For handling server protocol upgrades, see #1323) HTTP/1 allows starting a different protocol with an HTTP/1 request, and the desired protocol in an `Upgrade` header. We should support this. Edit: removed bad API proposal.
This looks great! The channel based approach we discussed also seems to be equally ergonomic and a looks much simpler to implement without adding another type parameter to virtually all types in `proto`. Is there any reason you prefer this over that? Well, shoot. The reason I liked this was for the end user API, but...
2018-06-11T23:35:42Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,556
hyperium__hyper-1556
[ "1547" ]
f20afba57d6fabb04085968342e5fd62b45bc8df
diff --git a/src/headers.rs b/src/headers.rs --- a/src/headers.rs +++ b/src/headers.rs @@ -78,6 +78,13 @@ pub fn content_length_value(len: u64) -> HeaderValue { } } +pub fn set_content_length_if_missing(headers: &mut HeaderMap, len: u64) { + headers + .entry(CONTENT_LENGTH) + .unwrap() + ...
diff --git a/tests/integration.rs b/tests/integration.rs --- a/tests/integration.rs +++ b/tests/integration.rs @@ -166,6 +166,29 @@ t! { ; } +t! { + post_outgoing_length, + client: + request: + method: "POST", + uri: "/hello", + body: "hello, world!", + ...
HTTP2 outgoing messages should set content-length if payload knows it For instance, a server replying with `Response::new(Body::from("Hello, World!"))` over HTTP2 should set the `content-length: 13` header. ## Steps to fix - In `proto::h2::server`, before calling `SendResponse::send_response`, check `body.content...
2018-06-08T21:21:11Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,554
hyperium__hyper-1554
[ "1545" ]
396fe80e76840dea9373ca448b20cf7a9babd2f8
diff --git a/src/body/body.rs b/src/body/body.rs --- a/src/body/body.rs +++ b/src/body/body.rs @@ -35,6 +35,7 @@ pub struct Body { enum Kind { Once(Option<Chunk>), Chan { + content_length: Option<u64>, abort_rx: oneshot::Receiver<()>, rx: mpsc::Receiver<Result<Chunk, ::Error>>, ...
diff --git a/tests/client.rs b/tests/client.rs --- a/tests/client.rs +++ b/tests/client.rs @@ -1424,6 +1424,63 @@ mod conn { res.join(rx).map(|r| r.0).wait().unwrap(); } + #[test] + fn incoming_content_length() { + use hyper::body::Payload; + + let server = TcpListener::bind("127.0.0...
Incoming HTTP1 bodies should know content_length and is_end_stream For instance, if a server receives the follow request: ``` POST /foo HTTP/1.1 content-length: 15 ``` The `Body` in the `Request` should know these things, neither of which are currently true: - `Body::content_length()` should be `Some(15)`. ...
2018-06-08T03:56:07Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,503
hyperium__hyper-1503
[ "1486" ]
18f4dd240631df55064696de865b6d818cb76ba7
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -45,6 +45,7 @@ num_cpus = "1.0" pretty_env_logger = "0.2.0" spmc = "0.2" url = "1.0" +tokio-mockstream = "1.1.0" [features] default = [ diff --git a/src/error.rs b/src/error.rs --- a/src/error.rs +++ b/src/error.rs @@ -69,6 +69,7 @@ pub(cr...
diff --git /dev/null b/src/server/rewind.rs new file mode 100644 --- /dev/null +++ b/src/server/rewind.rs @@ -0,0 +1,208 @@ +use bytes::{Buf, BufMut, Bytes, IntoBuf}; +use futures::{Async, Poll}; +use std::io::{self, Read, Write}; +use std::cmp; +use tokio_io::{AsyncRead, AsyncWrite}; + +#[derive(Debug)] +pub struct Re...
Try parsing HTTP2 on server connections if HTTP1 parsing fails This would be a next step to allowing the server to have HTTP2 support on by default.
I'd like to take this one on. @estk cool! I haven't personally thought out an exact implementation yet... One way could be for `server::conn::Connection` to be able to inspect a returned error, and if it's a parse error, try to take the IO and read buffer back out and use them to try the `h2::server::handshake`... E...
2018-04-27T23:34:32Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,502
hyperium__hyper-1502
[ "1484" ]
5e3b43af09fda86df59486edbeb9e68c6801a503
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -24,7 +24,6 @@ include = [ bytes = "0.4.4" futures = "0.1.21" futures-cpupool = { version = "0.1.6", optional = true } -futures-timer = "0.1.0" http = "0.1.5" httparse = "1.0" h2 = "0.1.5" diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.to...
diff --git a/.travis.yml b/.travis.yml --- a/.travis.yml +++ b/.travis.yml @@ -18,7 +18,7 @@ cache: script: - ./.travis/readme.py - cargo build $FEATURES - - 'if [ "$BUILD_ONLY" != "1" ]; then RUST_LOG=hyper cargo test $FEATURES; fi' + - 'if [ "$BUILD_ONLY" != "1" ]; then RUST_LOG=hyper cargo test $FEATURES --...
Replace futures-timer usage with tokio-timer
I'll happily take this one! @srijs I try to port it last night, but tokio-timer require work with `Runtime`, so in many test cases, `res.join(...).wait()` not work well. Good luck! Oh, interesting! If tokio's `current_thread` executor could be updated to use a timer as well, we could probably change all `wait`s to use ...
2018-04-27T23:17:09Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,497
hyperium__hyper-1497
[ "1492" ]
988dc7c637bc54e97a325a95d1e10d39d5fc8b2d
diff --git a/src/client/conn.rs b/src/client/conn.rs --- a/src/client/conn.rs +++ b/src/client/conn.rs @@ -67,6 +67,7 @@ where pub struct Builder { exec: Exec, h1_writev: bool, + h1_title_case_headers: bool, http2: bool, } diff --git a/src/client/conn.rs b/src/client/conn.rs --- a/src/client/conn....
diff --git a/src/proto/h1/role.rs b/src/proto/h1/role.rs --- a/src/proto/h1/role.rs +++ b/src/proto/h1/role.rs @@ -857,6 +901,21 @@ mod tests { Client::decoder(&head, method).unwrap_err(); } + #[test] + fn test_client_request_encode_title_case() { + use http::header::HeaderValue; + u...
Headers are lower-cased when sent and no option to disable this feature. I am trying to use Hyper with a stubborn web service that does not like requests with headers that do not have the first letter of each word capitalized (`Content-Length` for example). I tried out some other HTTP client libraries for Rust (like `t...
I expected this problem to appear eventually. Clearly, the best thing to happen is for that server software to be fixed. But, I realized that there is likely other janky server software out there that also relies on title case, just rotting with no fix to ever come... Possibly a solution could be for hyper to add an...
2018-04-23T21:21:54Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,490
hyperium__hyper-1490
[ "1461" ]
71a15c25f583dcb8dacc21904cf44c658cf429b8
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -34,7 +34,6 @@ net2 = "0.2.32" time = "0.1" tokio = "0.1.5" tokio-executor = "0.1.0" -tokio-service = "0.1" tokio-io = "0.1" want = "0.0.3" diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benche...
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -71,20 +71,20 @@ fn post_one_at_a_time(b: &mut test::Bencher) { static PHRASE: &'static [u8] = include_bytes!("../CHANGELOG.md"); //b"Hello, World!"; fn spawn_hello(rt: &mut Runtime) -> SocketAddr {...
Update Service trait from tokio-service to tower
Hi, I want to have a try, any mentoring instructions for this change? This wouldn't be too much work in hyper directly, but it does require [tower](https://github.com/tower-rs/tower) to publish a version to crates.io. I'd probably expect it to do so only after futures 0.2, since it needs to update to the new futures.
2018-04-17T23:34:09Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,488
hyperium__hyper-1488
[ "1263" ]
35c38cba6e13e1d2122f46bd0e02eafcebf776f7
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -13,7 +13,8 @@ use tokio::runtime::Runtime; use tokio::net::TcpListener; use hyper::{Body, Method, Request, Response}; -use hyper::server::Http; +use hyper::client::HttpConnector; +use hyper::server...
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -21,8 +22,10 @@ fn get_one_at_a_time(b: &mut test::Bencher) { let mut rt = Runtime::new().unwrap(); let addr = spawn_hello(&mut rt); - let client = hyper::Client::configure() - .bu...
Http::bind does not scale to uses where a Handle is needed So, I have a [`ReverseProxy` type that requires a `Client`](https://docs.rs/hyper-reverse-proxy/0.2.1/hyper_reverse_proxy/struct.ReverseProxy.html#method.new) in order to be constructed. But in order to do that I need to [do the manual set up of the Tokio event...
Same here, exposing a handle is not enough: I'm running two HTTP servers on the same event loop (one http handling let's encrypt, and redirecting other requests to https), plus a pleingres database connection, and an SSH server. I would be able to spawn the pleingres and SSH server using just the current handle, but...
2018-04-16T19:04:36Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,480
hyperium__hyper-1480
[ "1338" ]
33874f9a7577eb20374c6d8943ce665feb9619e0
diff --git a/examples/client.rs b/examples/client.rs --- a/examples/client.rs +++ b/examples/client.rs @@ -40,8 +40,9 @@ fn main() { println!("Response: {}", res.status()); println!("Headers: {:#?}", res.headers()); - res.into_parts().1.into_stream().for_each(|chunk| { - ...
diff --git a/benches/server.rs b/benches/server.rs --- a/benches/server.rs +++ b/benches/server.rs @@ -78,7 +78,7 @@ fn throughput_fixedsize_large_payload(b: &mut test::Bencher) { fn throughput_fixedsize_many_chunks(b: &mut test::Bencher) { bench_server!(b, ("content-length", "1000000"), || { static S: &...
From-trait based error conversion for request bodies Picking up that particular thread from #1328, I was wondering whether it might be worth it to implement a From/Into-based error coercion for request bodies. This would be different from #1129 in so far as I believe that it should be possible to ship it as a non-br...
That seems fine to me. I think I tried it out a few months ago, and ran into issues propagating the generics everywhere, but maybe I did it wrong!
2018-04-06T21:13:43Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,470
hyperium__hyper-1470
[ "1448" ]
5db85316a10d0b7bdd36524d85a746a23bd10190
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -22,20 +22,17 @@ include = [ [dependencies] bytes = "0.4.4" -futures = "0.1.17" -futures-cpupool = "0.1.6" -futures-timer = "0.1.0" +futures = "0.2.0-beta" +futures-timer = { git = "https://github.com/alexcrichton/futures-timer.git" } http =...
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -22,19 +23,20 @@ fn get_one_at_a_time(b: &mut test::Bencher) { let addr = spawn_hello(&mut rt); let client = hyper::Client::configure() - .build_with_executor(&rt.handle(), rt.executo...
Update to futures 0.2 It's not released yet, but this is just tracking that it needs to happen and is part of 0.12.
I'm super excited for this and would be interested in helping out! futures-0.2 is nearly code-complete, so we could probably get started soon. Do you have a branch you'd like PRs submitted to? There's a 0.12.x branch now on the repo where I've been trying to add new breaking changes to. Now that tokio 0.1 has landed in...
2018-03-20T09:55:12Z
0.12
a1609fbb332fccffbeb85d16f1cc0bf98c6ede21
hyperium/hyper
1,459
hyperium__hyper-1459
[ "1323" ]
eb15c660c106a2be08afd07e6cdfce162272d6cc
diff --git a/src/proto/h1/conn.rs b/src/proto/h1/conn.rs --- a/src/proto/h1/conn.rs +++ b/src/proto/h1/conn.rs @@ -500,6 +500,8 @@ where I: AsyncRead + AsyncWrite, Ok(encoder) => { if !encoder.is_eof() { Writing::Body(encoder) + } else if encoder.is_last...
diff --git a/tests/server.rs b/tests/server.rs --- a/tests/server.rs +++ b/tests/server.rs @@ -1,5 +1,6 @@ #![deny(warnings)] extern crate hyper; +#[macro_use] extern crate futures; extern crate spmc; extern crate pretty_env_logger; diff --git a/tests/server.rs b/tests/server.rs --- a/tests/server.rs +++ b/tests/s...
Server Protocol Upgrades Add support for servers to receive requests that wish to upgrade to a different protocol, such as websockets. A proposed API follows: ## Proposal - `Response` - `pub fn upgrade(proto: header::Upgrade) -> (Response, server::Upgrade)` This sets the `StatusCode::SwitchingProtoc...
This is a great idea, but what about handling CONNECT method at the server side (like in a HTTP proxy server)? I think the upgrade function should only provide the underlying I/O object, and it is the user's responsibility to return a proper HTTP response. @lqf96 I would definitely like to also offer a way of handling ...
2018-03-08T21:45:47Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,454
hyperium__hyper-1454
[ "1449" ]
0786ea1f871de9e007295238f20512825720853e
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -8,14 +8,13 @@ extern crate tokio_core; use std::net::SocketAddr; -use futures::{future, Future, Stream}; +use futures::{Future, Stream}; use tokio_core::reactor::{Core, Handle}; use tokio_core::...
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -42,13 +41,15 @@ fn get_one_at_a_time(b: &mut test::Bencher) { #[bench] fn post_one_at_a_time(b: &mut test::Bencher) { + extern crate pretty_env_logger; + let _ = pretty_env_logger::try_init()...
A lower level Connection API for the Client The `Client` is a higher level API that, among other things, provides a connection pool. Similar to how the server has higher and lower level APIs (`Server` vs `Serve`/`Connection`), the client should gain a lower level connection based API. This would allow finer grained con...
This would be great! The API I've been prototyping has a similar look to the [h2 client](https://docs.rs/h2/0.1.*/h2/client/index.html), where you provide an already connected `io`, and get back a sender (`SendRequest`?) and `Connection` pair. You probably want to just spawn the `Connection` into an executor, and then ...
2018-03-01T01:43:41Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,434
hyperium__hyper-1434
[ "1433" ]
8fb84d29a1c0ff848ff1c921f4c5bb7493939b97
diff --git /dev/null b/src/client/cancel.rs new file mode 100644 --- /dev/null +++ b/src/client/cancel.rs @@ -0,0 +1,148 @@ +use std::sync::Arc; +use std::sync::atomic::{AtomicBool, Ordering}; + +use futures::{Async, Future, Poll}; +use futures::task::{self, Task}; + +use common::Never; + +use self::lock::Lock; + +#[de...
diff --git a/src/proto/h1/dispatch.rs b/src/proto/h1/dispatch.rs --- a/src/proto/h1/dispatch.rs +++ b/src/proto/h1/dispatch.rs @@ -435,8 +431,6 @@ where #[cfg(test)] mod tests { - use futures::Sink; - use super::*; use mock::AsyncIo; use proto::ClientTransaction; diff --git a/src/proto/h1/dispatch...
Connection pool returns connections which are not ready I am running a high number of concurrent requests (~1000) to the AWS Kinesis service and I am getting an error log from the hyper connection pool: ``` root@doit-1800981089-sz6zt:~/kinesis-hyper-bug# cargo build --release && RUST_LOG=error RUST_BACKTRACE=full ....
2018-02-03T00:03:56Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,390
hyperium__hyper-1390
[ "1365" ]
cecef9d402b76af12e6415519deb2b604f77b195
diff --git a/src/proto/conn.rs b/src/proto/conn.rs --- a/src/proto/conn.rs +++ b/src/proto/conn.rs @@ -453,6 +453,14 @@ where I: AsyncRead + AsyncWrite, pub fn close_write(&mut self) { self.state.close_write(); } + + pub fn disable_keep_alive(&mut self) { + if self.state.is_idle() { + ...
diff --git a/src/proto/conn.rs b/src/proto/conn.rs --- a/src/proto/conn.rs +++ b/src/proto/conn.rs @@ -869,7 +881,7 @@ mod tests { other => panic!("unexpected frame: {:?}", other) } - // client + // client let io = AsyncIo::new_buf(vec![], 1); ...
Make graceful shutdown work at request rather than connection granularity The graceful shutdown logic currently waits until all `Service`s have dropped (i.e. all connections have closed). However, with keep-alive this is overconservative - what we'd really like is to wait until pending requests have finished. This c...
@seanmonstar pointed out that a simple way of doing this is to add a method to disable keep-alive on Connections. Then, `Service::drop`-based tracking will be accurate.
2017-11-29T07:06:16Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,387
hyperium__hyper-1387
[ "1383" ]
e4864a2bea59b40fb07e6d18329f75817803a3f3
diff --git a/src/client/pool.rs b/src/client/pool.rs --- a/src/client/pool.rs +++ b/src/client/pool.rs @@ -200,10 +200,13 @@ impl<T: Clone> KeepAlive for Pooled<T> { }; if pool.is_enabled() { pool.put(self.key.clone(), self.entry.clone()); + } else { + ...
diff --git a/tests/client.rs b/tests/client.rs --- a/tests/client.rs +++ b/tests/client.rs @@ -654,6 +654,46 @@ mod dispatch_impl { } + #[test] + fn no_keep_alive_closes_connection() { + // https://github.com/hyperium/hyper/issues/1383 + let _ = pretty_env_logger::init(); + + let ser...
All client connections are leaked when keep_alive(false) ```rust extern crate hyper; extern crate tokio_core; extern crate tokio_io; extern crate futures; extern crate env_logger; use tokio_core::reactor::{Core, Timeout}; use tokio_core::net::TcpStream; use tokio_io::{AsyncRead, AsyncWrite}; use hyper::{Clie...
2017-11-28T05:44:01Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,362
hyperium__hyper-1362
[ "1353" ]
8153cfaebf65779c279dc089767dece508d3359c
diff --git a/.travis.yml b/.travis.yml --- a/.travis.yml +++ b/.travis.yml @@ -8,6 +8,8 @@ matrix: env: FEATURES="--features nightly" - rust: beta - rust: stable + - rust: stable + env: HYPER_NO_PROTO=1 - rust: stable env: FEATURES="--features compat" ...
diff --git a/src/proto/conn.rs b/src/proto/conn.rs --- a/src/proto/conn.rs +++ b/src/proto/conn.rs @@ -727,7 +769,7 @@ mod tests { use futures::future; use tokio_proto::streaming::pipeline::Frame; - use proto::{self, MessageHead, ServerTransaction}; + use proto::{self, ClientTransaction, MessageHead, ...
Not reading the entire response body leaks the connection on the client side This program continually creates new connections without ever closing old ones: ```rust extern crate hyper; extern crate tokio_core; extern crate tokio_io; extern crate futures; extern crate env_logger; use tokio_core::reactor::{Cor...
2017-10-27T06:25:54Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,325
hyperium__hyper-1325
[ "1315" ]
971864c424495e9dd4cafe6c7fb4c7504a01e03d
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -28,6 +28,7 @@ language-tags = "0.2" log = "0.3" mime = "0.3.2" percent-encoding = "1.0" +relay = "0.1" time = "0.1" tokio-core = "0.1.6" tokio-proto = "0.1" diff --git a/src/client/pool.rs b/src/client/pool.rs --- a/src/client/pool.rs +++ ...
diff --git a/src/client/pool.rs b/src/client/pool.rs --- a/src/client/pool.rs +++ b/src/client/pool.rs @@ -364,4 +408,30 @@ mod tests { })).map(|(entry, _)| entry); assert_eq!(*checkout.wait().unwrap(), *pooled1); } + + #[test] + fn test_pool_checkout_drop_cleans_up_parked() { + futu...
Memory leak when using client.clone() for server (proxy) I'm writing a reverse proxy and use the HTTP client to connect to any upstream server. Code is in https://github.com/klausi/rustnish/blob/goal-06/src/lib.rs#L150 . The server is leaking memory so I must be doing something wrong. Steps to reproduce: 1. Make ...
The `Client` contains a pool of connections internally. Are the requests to different hosts? The pool is lazy about ejecting expired connections, so that could be part of it... The requests are always to the same host, http://localhost/ in this example. I will try to find out what happens on a cloned client and what ex...
2017-09-18T20:30:32Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,259
hyperium__hyper-1259
[ "1257" ]
5f47d72347231c60a953c228abec75b077e99f8d
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -7,10 +7,10 @@ use futures::task::Task; use tokio_io::{AsyncRead, AsyncWrite}; use tokio_proto::streaming::pipeline::{Frame, Transport}; -use header::{ContentLength, TransferEncoding}; use http::{self, Http1Transactio...
diff --git a/src/http/h1/parse.rs b/src/http/h1/parse.rs --- a/src/http/h1/parse.rs +++ b/src/http/h1/parse.rs @@ -421,63 +444,83 @@ mod tests { fn test_decoder_request() { use super::Decoder; + let method = &mut None; let mut head = MessageHead::<::http::RequestLine>::default(); ...
Response without a body receives `Transfer-Encoding: chunked` I want a server to make a `304 Not Modified` response like this: ```rust Response::new() .with_status(StatusCode::NotModified) .with_header(ETag(etag)) .with_header(LastModified(last_modified)) ``` My problem is *hyper* ...
It's a bug in hyper. I've been working on this week actually, to better handle when message shouldn't have bodies. Besides 304, also 204, and also when the request verb would not allow a response body, like `HEAD` or `CONNECT`.
2017-07-13T19:19:31Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,182
hyperium__hyper-1182
[ "650" ]
97abb81b3cc271a6dd258da86559bc25ab77c652
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -57,6 +57,7 @@ pub use self::transfer_encoding::TransferEncoding; pub use self::upgrade::{Upgrade, Protocol, ProtocolName}; pub use self::user_agent::UserAgent; pub use self::vary::Vary; ...
diff --git /dev/null b/src/header/common/link.rs new file mode 100644 --- /dev/null +++ b/src/header/common/link.rs @@ -0,0 +1,1110 @@ +use std::fmt; +use std::borrow::Cow; +use std::str::FromStr; +use std::ascii::AsciiExt; + +use mime::Mime; +use language_tags::LanguageTag; + +use header::parsing; +use header::{Header...
Link header Currently, as far as I know there's no facility to parse and generate Link headers, It would be very useful if we had one, many APIs (including github) uses it to navigate. Do we plan to add support for it? http://www.w3.org/wiki/LinkHeader http://www.rfc-editor.org/rfc/rfc5988.txt
Sounds fine to me. It seems like this header is more complicated than others, just by judging it's ABNF: ``` ABNF Link = "Link" ":" #link-value link-value = "<" URI-Reference ">" *( ";" link-param ) link-param = ( ( "rel" "=" relation-types ) | ( "anchor" "=" <"> URI-Reference <"> ) ...
2017-05-18T22:47:39Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
1,178
hyperium__hyper-1178
[ "1176" ]
33eb8f95a3537cfd02aa4956a69dedd0c840a604
diff --git a/src/uri.rs b/src/uri.rs --- a/src/uri.rs +++ b/src/uri.rs @@ -213,9 +213,8 @@ fn parse_scheme(s: &str) -> Option<usize> { fn parse_authority(s: &str) -> usize { let i = s.find("://").map(|p| p + 3).unwrap_or(0); - s[i..].find('/') - .or_else(|| s[i..].find('?')) - .or_else(|| s[i.....
diff --git a/src/uri.rs b/src/uri.rs --- a/src/uri.rs +++ b/src/uri.rs @@ -537,6 +536,28 @@ test_parse! { port = None, } +test_parse! { + test_uri_parse_absolute_form_with_empty_path_and_fragment_with_slash, + "http://127.0.0.1#foo/bar", + scheme = Some("http"), + authority = Some("127.0.0.1"), + ...
hyper::Uri doesn't parse Uris with empty path well Problem: Latest master hyper::Uri thinks `https://google.com#ab/cd` has a host `google.com#ab`, while it should have a host `google.com`. hyper::Uri is okay with `https://google.com/#ab/cd`.
2017-05-17T11:42:58Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,162
hyperium__hyper-1162
[ "1155" ]
df1095dfe79e199128bd605777dd0f3e58487e46
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -22,19 +22,19 @@ pub use tokio_service::Service; use header::{Headers, Host}; use http::{self, TokioBody}; +use http::response; +use http::request; use method::Method; use self::pool::{Pool, Pooled}; use uri::{s...
diff --git a/tests/server.rs b/tests/server.rs --- a/tests/server.rs +++ b/tests/server.rs @@ -503,7 +503,7 @@ fn test_server_disable_keep_alive() { .status(hyper::Ok) .header(hyper::header::ContentLength(quux.len() as u64)) .body(quux); - + let _ = req.write_all(b"\ GET /qu...
Unite Request and Response types Currently, `client::Request` is a different type from `server::Request`, and `client::Response` is different from `server::Response`. We want to unite them into the same type. That dramatically improves generic middleware that could apply on both a server and a client, and makes thin...
2017-05-01T18:43:59Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,152
hyperium__hyper-1152
[ "1145" ]
c3466bedf8bd634c9c1aba25a3790bf8ae5e89d7
diff --git a/src/header/common/cookie.rs b/src/header/common/cookie.rs --- a/src/header/common/cookie.rs +++ b/src/header/common/cookie.rs @@ -1,7 +1,10 @@ -use header::{Header, Raw}; -use std::fmt::{self, Display}; +use std::borrow::Cow; +use std::fmt; use std::str::from_utf8; +use header::{Header, Raw}; +use heade...
diff --git a/src/header/common/cookie.rs b/src/header/common/cookie.rs --- a/src/header/common/cookie.rs +++ b/src/header/common/cookie.rs @@ -59,16 +120,110 @@ impl Header for Cookie { } } +impl PartialEq for Cookie { + fn eq(&self, other: &Cookie) -> bool { + if self.0.len() == other.0.len() { + ...
Cookie Header should be Map-like We only ever receive headers like this: `Cookie: foo=bar; session=sean; hello=world`, and having that be a vector of `["foo=bar", "session=sean", "hello=world"]` doesn't really help anyone. Instead, anyone needing to accept cookies would only be looking for a certain name anyways, so...
For implementation, something like the `VecMap` inside `Headers` would probably be best, since it performs faster and with less memory than a `HashMap` when the number of items is small. I can't imagine the `Cookie` header, in the common case, having that many key-value pairs. So, it could look like: ```rust pub...
2017-04-26T20:25:54Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,147
hyperium__hyper-1147
[ "1065" ]
1cd8ea36f33486c352b53cbbf92bba889077e85a
diff --git a/src/header/common/origin.rs b/src/header/common/origin.rs --- a/src/header/common/origin.rs +++ b/src/header/common/origin.rs @@ -31,20 +31,39 @@ use header::parsing::from_one_raw_str; /// ); /// ``` -#[derive(Clone, Debug)] -pub struct Origin { - /// The scheme, such as http or https - scheme: C...
diff --git a/src/header/common/origin.rs b/src/header/common/origin.rs --- a/src/header/common/origin.rs +++ b/src/header/common/origin.rs @@ -104,26 +129,24 @@ impl FromStr for Origin { s => Cow::Owned(s.to_owned()) }; - Ok(Origin{ + Ok(Origin(OriginOrNull::Origin { ...
Support Opaque origin headers, serializing it to `null` Support Opaque origins: > An internal value, with no serialization it can be recreated from (it is serialized as "null" per ASCII serialization of an origin), for which the only meaningful operation is testing for equality. https://html.spec.whatwg.org/multi...
Sorry, in what context? The url? The `Origin` header? The origin header. I see. So it can either be `null`, or the full origin. It seems like any way to make this work with hyper's `Origin` header would require a breaking change. Since that's the case, my thinking is something like this, for inclusion in v0.11: ```r...
2017-04-20T20:27:31Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,136
hyperium__hyper-1136
[ "1124" ]
574ded90dd5bf52b2ea10420fe34379636b094bd
diff --git a/src/header/common/origin.rs b/src/header/common/origin.rs --- a/src/header/common/origin.rs +++ b/src/header/common/origin.rs @@ -84,6 +84,9 @@ impl Header for Origin { } } +static HTTP : &'static str = "http"; +static HTTPS : &'static str = "https"; + impl FromStr for Origin { type Err = ::E...
diff --git a/src/header/common/origin.rs b/src/header/common/origin.rs --- a/src/header/common/origin.rs +++ b/src/header/common/origin.rs @@ -121,14 +128,26 @@ impl PartialEq for Origin { mod tests { use super::Origin; use header::Header; + use std::borrow::Cow; + + macro_rules! assert_borrowed{ + ...
Reduce to_owned of scheme in Origin header Now that the `scheme` propertry is a `Cow<'static, str>`, and the scheme is pretty much always only ever "http" or "https", we should compare for those and then store a `Cow::Borrowed("http")` etc, instead of calling `scheme.to_owned()`.
2017-04-13T22:20:32Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,135
hyperium__hyper-1135
[ "1134" ]
f05a58a1b2288d30a601b0466bc08a38c5a76054
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -610,6 +610,9 @@ impl<B, K: KeepAlive> State<B, K> { } fn busy(&mut self) { + if let KA::Disabled = self.keep_alive.status() { + return; + } self.keep_alive.busy(); }
diff --git a/tests/server.rs b/tests/server.rs --- a/tests/server.rs +++ b/tests/server.rs @@ -146,10 +146,16 @@ fn connect(addr: &SocketAddr) -> TcpStream { } fn serve() -> Serve { - serve_with_timeout(None) + serve_with_options(Default::default()) } -fn serve_with_timeout(dur: Option<Duration>) -> Serve ...
Connection Keepalive Hi, I'm currently using the hyper master branch. I have written a Service and want that the connection is dropped after each request. I thought that using "Http::keep_alive(false)" should close the connection? But that does not work. Is this a bug or do I use the wrong functionality?
2017-04-13T00:05:44Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,116
hyperium__hyper-1116
[ "1111" ]
6e55fbe75d0ac98b162b4f44f746df76d71a778a
diff --git a/src/http/io.rs b/src/http/io.rs --- a/src/http/io.rs +++ b/src/http/io.rs @@ -55,31 +55,33 @@ impl<T: AsyncRead + AsyncWrite> Buffered<T> { } pub fn parse<S: Http1Transaction>(&mut self) -> ::Result<Option<MessageHead<S::Incoming>>> { - self.reserve_read_buf(); - match self.read_f...
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -745,7 +745,6 @@ mod tests { let mut conn = Conn::<_, http::Chunk, ServerTransaction>::new(io, Default::default()); conn.state.idle(); - assert!(conn.poll().unwrap().is_not_ready()); mat...
Example tokio client hangs when HTTP Response code is in its own packet Steps to reproduce: - Run nc -l -p 8081 - Run hyper/examples/client.rs http://localhost:8081/ - In the nc terminal, paste: ``` HTTP/1.1 200 OK Hello World ``` - Press ^C - This works fine. Now repeat, but: - Run nc -l -p 8081 - Run ...
2017-04-05T22:23:25Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,115
hyperium__hyper-1115
[ "1112" ]
66ad619d6ece49a2325bf22f85fdecc9e3eda3c4
diff --git a/src/uri.rs b/src/uri.rs --- a/src/uri.rs +++ b/src/uri.rs @@ -3,6 +3,7 @@ use std::fmt::{Display, self}; use std::str::{self, FromStr}; use http::ByteStr; +use bytes::{BufMut, BytesMut}; /// The Request-URI of a Request's StartLine. /// diff --git a/src/uri.rs b/src/uri.rs --- a/src/uri.rs +++ b/sr...
diff --git a/src/http/h1/parse.rs b/src/http/h1/parse.rs --- a/src/http/h1/parse.rs +++ b/src/http/h1/parse.rs @@ -344,7 +344,7 @@ mod tests { let (req, len) = parse::<http::ServerTransaction, _>(&mut raw).unwrap().unwrap(); assert_eq!(len, expected_len); assert_eq!(req.subject.0, ::Method::G...
Send invalid request line when there's no trailing slash in url If url is requested without trailing slash like "http://www.google.com", '404 not found' is returned because Request line doesn't have Request URI like "GET HTTP/1.1". It is ok if url is requested like "http://www.google.com/"
2017-04-05T21:26:06Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,110
hyperium__hyper-1110
[ "1108" ]
d63b7de44f813696f8ec595d2f8f901526c1720e
diff --git a/src/http/io.rs b/src/http/io.rs --- a/src/http/io.rs +++ b/src/http/io.rs @@ -134,15 +134,16 @@ impl<T: Write> Write for Buffered<T> { fn flush(&mut self) -> io::Result<()> { if self.write_buf.remaining() == 0 { - Ok(()) + self.io.flush() } else { ...
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -827,6 +827,7 @@ mod tests { assert!(conn.state.writing.is_queued()); assert!(conn.poll_complete().unwrap().is_ready()); assert!(!conn.state.writing.is_queued()); + assert!...
Client doesn't appear to flush when waiting to read a response I was implementing a custom I/O stream to send requests over and I saw that the request was never actually written when it was waiting for a response. Turns out I was accidentally enabling buffering on the abstraction I was using and so none of the data was...
Interesting! As `conn.poll_complete()` is called, we take that to mean to flush `Conn`s `write_buf` into the `io`, but then we don't actually call `io.flush()`. I assume that's the missing piece? I think so yeah, that'd make sense. This came up w/ a custom `AsyncRead + AsyncWrite` implementation so a call to `io.flush(...
2017-04-03T17:08:03Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,100
hyperium__hyper-1100
[ "1089" ]
e81184e53c5c59a6262201e212a04a772f089079
diff --git a/examples/client.rs b/examples/client.rs --- a/examples/client.rs +++ b/examples/client.rs @@ -24,8 +24,8 @@ fn main() { } }; - let url = hyper::Url::parse(&url).unwrap(); - if url.scheme() != "http" { + let url = url.parse::<hyper::Uri>().unwrap(); + if url.scheme() != Some("htt...
diff --git a/benches/end_to_end.rs b/benches/end_to_end.rs --- a/benches/end_to_end.rs +++ b/benches/end_to_end.rs @@ -28,7 +28,7 @@ fn get_one_at_a_time(b: &mut test::Bencher) { let client = hyper::Client::new(&handle); - let url: hyper::Url = format!("http://{}/get", addr).parse().unwrap(); + let url: ...
Remove Url depedency Now that `hyper::Uri` has sufficient parsing capability, and can represent any URI for the server or client, does it make sense to still also use the `Url` type? When looking at the rustdocs generated, it certainly looks confusing that there is both `Url` and `Uri` at the crate root. And issues lik...
At least for me in how I've worked with libs in the past it's pretty rare for me to have a parsed `Url` on hand. Typically I manage it via a string and then just pass it to a library (there's also typically very little manipulation of the URL as well). I could go either way, but in some sense reducing the suite of p...
2017-03-21T18:03:13Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,074
hyperium__hyper-1074
[ "1073" ]
dc97dd77f45486d9cb9a22a1859809c5af5579e2
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -4,6 +4,7 @@ use std::marker::PhantomData; use std::time::Instant; use futures::{Poll, Async, AsyncSink, Stream, Sink, StartSend}; +use futures::task::Task; use tokio::io::Io; use tokio_proto::streaming::pipeline::{F...
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -605,14 +639,14 @@ impl<'a, T: fmt::Debug + 'a, B: AsRef<[u8]> + 'a> fmt::Debug for DebugFrame<'a, #[cfg(test)] mod tests { - use futures::{Async, Stream, Sink}; + use futures::{Async, Future, Stream, Sink}; ...
Infinite loop if body sender is dropped Originally [reported here](https://github.com/tokio-rs/tokio-core/issues/177#issuecomment-281574226) it looks like this code: ```rust extern crate env_logger; extern crate futures; extern crate hyper; use hyper::{Body, Error}; use hyper::server::{Request, Response}; us...
I've been looking in to trying to park and unpark a little less, but one thing here strikes me as odd: if the Sender hasn't sent anything, why would tokio try to flush again, especially if the previous flush already said it was complete? IIRC "flush" in this case basically just happens pretty commonly, but @carllerche ...
2017-02-23T00:36:01Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,054
hyperium__hyper-1054
[ "1041" ]
04f169034a4bf34927a65408cf418dcfacfe15f0
diff --git a/src/client/request.rs b/src/client/request.rs --- a/src/client/request.rs +++ b/src/client/request.rs @@ -7,6 +7,7 @@ use http::{Body, RequestHead}; use method::Method; use uri::Uri; use version::HttpVersion; +use std::str::FromStr; /// A client request to a remote server. pub struct Request { diff ...
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -577,6 +577,8 @@ mod tests { use super::{Conn, Writing}; use ::uri::Uri; + use std::str::FromStr; + #[test] fn test_conn_init_read() { let good_message = b"GET / HTTP/1.1\r\n\r\n".to_vec()...
Make Uri use MemSlice internally
New version with internal enum. There would also need to be a change in `http/h1/parse`, instead of using `req.path.unwrap().parse()`. Instead, we want to create a `Uri` using the related `MemSlice`. It could look like this: ```rust let path = slice.slice(path_start..path_end); // path was found to be utf8 by http...
2017-02-09T19:49:24Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,038
hyperium__hyper-1038
[ "1026" ]
5c890321ee2da727a814c18d4ee2df5eddd6720e
diff --git a/src/header/common/content_length.rs b/src/header/common/content_length.rs --- a/src/header/common/content_length.rs +++ b/src/header/common/content_length.rs @@ -40,6 +40,7 @@ impl Header for ContentLength { static NAME: &'static str = "Content-Length"; NAME } + fn parse_header(...
diff --git a/src/header/common/retry_after.rs b/src/header/common/retry_after.rs --- a/src/header/common/retry_after.rs +++ b/src/header/common/retry_after.rs @@ -141,9 +141,7 @@ impl Header for RetryAfter { #[cfg(test)] mod tests { - extern crate httparse; - - use header::{Header, Headers}; + use header::...
Raw is now using Memslice Fixes #1024. I still need to find where I made a mistake to fix the tests. (And add some tests as well)
2017-01-27T19:09:33Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
1,014
hyperium__hyper-1014
[ "1000" ]
b64665635ff3a5dc7f47a2ed847d7a98e1f1eb67
diff --git a/examples/server.rs b/examples/server.rs --- a/examples/server.rs +++ b/examples/server.rs @@ -23,12 +23,12 @@ impl Service for Echo { fn call(&self, req: Request) -> Self::Future { ::futures::finished(match (req.method(), req.path()) { - (&Get, Some("/")) | (&Get, Some("/echo")) ...
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -574,6 +574,7 @@ mod tests { use mock::AsyncIo; use super::{Conn, Writing}; + use ::uri::Uri; #[test] fn test_conn_init_read() { diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/co...
Redesign the Uri type There are a couple of things that a redesign could help with: - Making it an opaque struct instead of an enum means fiddling with internals is no longer a breaking change - It'd be best if we could create less copies when parsing out a URI - To future-proof for when hyper supports HTTP2 We...
2017-01-18T00:48:46Z
0.11
eb15c660c106a2be08afd07e6cdfce162272d6cc
hyperium/hyper
994
hyperium__hyper-994
[ "993" ]
bef05bf2075cc0569213f8cad4cd1bf22cc4e736
diff --git a/examples/hello.rs b/examples/hello.rs --- a/examples/hello.rs +++ b/examples/hello.rs @@ -17,7 +17,7 @@ impl Service for Hello { type Response = Response; type Error = hyper::Error; type Future = ::futures::Finished<Response, hyper::Error>; - fn call(&mut self, _req: Request) -> Self::Fut...
diff --git a/src/client/connect.rs b/src/client/connect.rs --- a/src/client/connect.rs +++ b/src/client/connect.rs @@ -344,7 +344,7 @@ mod tests { fn test_non_http_url() { let mut core = Core::new().unwrap(); let url = Url::parse("file:///home/sean/foo.txt").unwrap(); - let mut connector =...
tokio branch does not compile after tokio_service mutability change i am gonna try to write a PR to fix this! here's the error for posterity. ``` error[E0053]: method `call` has an incompatible type for trait --> src/client/connect.rs:81:5 | 81 | fn call(&mut self, url: Url) -> Self::Future { | ...
2017-01-10T00:05:57Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
971
hyperium__hyper-971
[ "969" ]
17a7e2b0514d3cf8f0f072b7ebc972223a338dad
diff --git /dev/null b/src/body.rs new file mode 100644 --- /dev/null +++ b/src/body.rs @@ -0,0 +1,81 @@ +//! The Hyper Body, which is a wrapper around the tokio_proto Body. +//! This is used as the Body for a Server Request, Server Response, +//! and also for a Client Request and Client Response. +//! It is based on a...
diff --git a/src/client/request.rs b/src/client/request.rs --- a/src/client/request.rs +++ b/src/client/request.rs @@ -73,38 +73,6 @@ pub fn split(req: Request) -> (RequestHead, Option<Body>) { (req.head, req.body) } -pub trait IntoBody { - fn into(self) -> Body; -} - -impl IntoBody for Body { - fn into(s...
[Tokio Branch] hyper::server::response::{ Body and IntoBody } are private, propose to expose publically Currently, you can call `response.set_body()` with anything that implements `hyper::server::response::IntoBody`, and right now that is just `Body` itself, `Vec<u8>` and `&'static [u8]`. In my application, I have m...
In general things are private because I'm being much more conservative with what is made public, so internal optimizations can be made without breaking people. In this particular case, I didn't really like the idea of a `IntoBody` trait, but instead wanted `Into<Body>`. The `IntoBody` trait currently exists because ...
2016-12-14T02:03:34Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
920
hyperium__hyper-920
[ "882" ]
588ef9d25242f6154294c0c9cc5966d886c0a0f3
diff --git a/src/header/common/referrer_policy.rs b/src/header/common/referrer_policy.rs --- a/src/header/common/referrer_policy.rs +++ b/src/header/common/referrer_policy.rs @@ -54,19 +54,23 @@ impl Header for ReferrerPolicy { fn parse_header(raw: &Raw) -> ::Result<ReferrerPolicy> { use self::ReferrerP...
diff --git a/src/header/common/referrer_policy.rs b/src/header/common/referrer_policy.rs --- a/src/header/common/referrer_policy.rs +++ b/src/header/common/referrer_policy.rs @@ -90,3 +94,10 @@ fn test_parse_header() { let e: ::Result<ReferrerPolicy> = Header::parse_header(&"foobar".into()); assert!(e.is_err(...
Referrer-Policy header should support multiple values In order to support fallback policies when a subset of the full policies is supported (such as when a new policy is introduced in the specification), hyper's Referrer-Policy header should store the most recent value that is successfully parsed from the list of poten...
2016-10-06T20:21:50Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
910
hyperium__hyper-910
[ "883" ]
8b3c1206846cb96be780923952eafe0dde7850bf
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -58,6 +58,7 @@ pub use self::transfer_encoding::TransferEncoding; pub use self::upgrade::{Upgrade, Protocol, ProtocolName}; pub use self::user_agent::UserAgent; pub use self::vary::Vary; ...
diff --git /dev/null b/src/header/common/warning.rs new file mode 100644 --- /dev/null +++ b/src/header/common/warning.rs @@ -0,0 +1,174 @@ +use std::fmt; +use std::str::{FromStr}; +use header::{Header, HttpDate, Raw}; +use header::parsing::from_one_raw_str; + +/// `Warning` header, defined in [RFC7234](https://tools.i...
add Warning header I would like to use the `Warning` header to warn users of deprecated API routes as proposed in this SO answer: http://stackoverflow.com/a/29623798. Spec: https://tools.ietf.org/html/rfc7234#section-5.5
2016-09-05T17:55:06Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
899
hyperium__hyper-899
[ "896" ]
a228486a85d88eb10b03509228bd11937ad57261
diff --git a/doc/guide/server.md b/doc/guide/server.md --- a/doc/guide/server.md +++ b/doc/guide/server.md @@ -165,7 +165,7 @@ impl Handler<Http> for Text { fn on_request(&mut self, req: Request<Http>) -> Next { use hyper::RequestUri; let path = match *req.uri() { - RequestUri::Absolut...
diff --git a/.travis.yml b/.travis.yml --- a/.travis.yml +++ b/.travis.yml @@ -22,7 +22,7 @@ script: - ./.travis/readme.py - cargo build --verbose $FEATURES - cargo test --verbose $FEATURES - - 'for f in ./doc/**/*.md; do rustdoc -L ./target/debug -L ./target/debug/deps --test $f; done' + - 'for f in ./doc/...
Add `path()` method to `server::Request` ``` rust fn path(&self) -> Option<&str> ``` It'd need to check for a query string in the `AbsolutePath` variant.
2016-08-23T02:47:00Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
898
hyperium__hyper-898
[ "891" ]
74136de9609e9e272e3fc2de1b6b939e629df093
diff --git a/src/header/internals/cell.rs b/src/header/internals/cell.rs --- a/src/header/internals/cell.rs +++ b/src/header/internals/cell.rs @@ -86,6 +86,20 @@ impl<V: ?Sized + Any + 'static> PtrMapCell<V> { }.map(|val| &mut **val) } + #[inline] + pub fn into_value(self, key: TypeId) -> Option<B...
diff --git a/src/header/mod.rs b/src/header/mod.rs --- a/src/header/mod.rs +++ b/src/header/mod.rs @@ -751,6 +760,13 @@ mod tests { assert_eq!(headers.get_raw("Content-length"), None); } + #[test] + fn test_remove() { + let mut headers = Headers::new(); + headers.set(ContentLength(10...
Headers should offer method to remove and return header If I take a `Headers` by value (moving it into my function), I want to avoid copying them. For that reason it would be useful if either `Headers::remove` would return the popped item (as `HashMap::remove` does), or if a new method was added.
(I'm willing to write a PR for this) Hm, I agree with this entirely, 👍 . This would be a breaking change, since `remove` currently returns a `bool`, but it seems like the right one to make. As such, I'd aim it for master and not for the 0.9.x branch. Basically: ``` rust fn remove<H: Header>(&mut self) -> Option<H>; ...
2016-08-22T19:54:17Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
875
hyperium__hyper-875
[ "870" ]
e5841365dd82ff7f594f24309b48ae692cb09cc5
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -27,6 +27,7 @@ pub use self::content_disposition::{ContentDisposition, DispositionType, Disposi pub use self::content_length::ContentLength; pub use self::content_encoding::ContentEncoding...
diff --git /dev/null b/src/header/common/content_location.rs new file mode 100644 --- /dev/null +++ b/src/header/common/content_location.rs @@ -0,0 +1,44 @@ +header! { + /// `Content-Location` header, defined in + /// [RFC7231](https://tools.ietf.org/html/rfc7231#section-3.1.4.2) + /// + /// The header can ...
Add Content-Location header See https://tools.ietf.org/html/rfc7231#section-3.1.4.2
I'd like to take a shot at this. ABNF from RFC here is: ``` Content-Location = absolute-URI / partial-URI ``` ABNF is exactly same as with `Referer` which is implemented as ``` header! { // TODO Use URL (Referer, "Referer") => [String] // testcase } ``` `/TODO:? Use URL/` is mentioned for `Location` an...
2016-07-27T09:58:01Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
873
hyperium__hyper-873
[ "806" ]
50ccdaa7e7db574ec9890c220765ffd2da5e493b
diff --git a/doc/guide/server.md b/doc/guide/server.md --- a/doc/guide/server.md +++ b/doc/guide/server.md @@ -1,30 +1,384 @@ % Server Guide -# The `Handler` +# Hello, World -```ignore,no_run -extern crate hyper; -use hyper::server::{Handler, Request, Response, Decoder, Encoder, Next, HttpStream as Http}; +Let's s...
diff --git a/.travis.yml b/.travis.yml --- a/.travis.yml +++ b/.travis.yml @@ -22,7 +22,7 @@ script: - ./.travis/readme.py - cargo build --verbose $FEATURES - cargo test --verbose $FEATURES - - 'for f in ./doc/**/*.md; do rustdoc --test $f; done' + - 'for f in ./doc/**/*.md; do rustdoc -L ./target/debug -L ....
Add a Server Guide Add a guide on how to do more and more complicated things with the Server/Handler in `docs/guide/server.md`.
So what sort of spec would you hold for this guide? like what points should it cover? I could imagine the guide going in steps: - Introduce the `Handler` trait. Such as show how to make a simple "Hello World" handler, like from the `examples/hello.rs`. - Explain `WouldBlock`, and/or `try_read`/`try_write`, and repeati...
2016-07-26T00:44:05Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
867
hyperium__hyper-867
[ "859" ]
a22ae26cecf56b1a97b3d15bc57ac05a6231c8e6
diff --git a/src/net.rs b/src/net.rs --- a/src/net.rs +++ b/src/net.rs @@ -1,6 +1,7 @@ //! A collection of traits abstracting over Listeners and Streams. use std::io::{self, Read, Write}; use std::net::{SocketAddr}; +use std::option; use rotor::mio::tcp::{TcpStream, TcpListener}; use rotor::mio::{Selector, Token...
diff --git a/src/server/mod.rs b/src/server/mod.rs --- a/src/server/mod.rs +++ b/src/server/mod.rs @@ -375,8 +403,3 @@ where F: FnMut(http::Control) -> H, H: Handler<T>, T: Transport { self(ctrl) } } - -#[cfg(test)] -mod tests { - -} diff --git a/tests/server.rs b/tests/server.rs --- a/tests/server.rs ++...
Listen on multiple addresses in async hyper Given that hyper is now async it would be great to be able to easily have a single server listen on multiple addresses. It's already possible to get the local address of a connection via `Request::transport`. It's not possible to implement your own `Transport` to do this as...
Interesting. Your proposed use case is to listen on port 80 and 443 in the same event loop? At the moment I just want to listen for http on multiple IPV4 and 6 addresses. That just seemed like a natural extension. While I have personal feelings on having an app run on both 80 and 443 together (I'd be that the http li...
2016-07-18T07:04:45Z
0.10
97abb81b3cc271a6dd258da86559bc25ab77c652
hyperium/hyper
857
hyperium__hyper-857
[ "848" ]
220d09fc3a68cdd9a10589b2b454b54312518bb3
diff --git a/examples/client.rs b/examples/client.rs --- a/examples/client.rs +++ b/examples/client.rs @@ -48,7 +48,7 @@ impl hyper::client::Handler<HttpStream> for Dump { Err(e) => match e.kind() { io::ErrorKind::WouldBlock => Next::read(), _ => { - pri...
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -943,22 +956,13 @@ impl<'a, K: Key + 'a> Seed<'a, K> { pub trait MessageHandlerFactory<K: Key, T: Transport> { type Output: MessageHandler<T>; - fn create(&mut self, seed: Seed<K>) -> Self::Output; -} + fn cr...
Client panics with 'handler not in queue for key' Using latest master. I have a `Client` that sends HTTP POST requests to the same URL from multiple threads and sometime this happens: ``` 2016-06-29 19:39:49 TRACE:hyper::client: on_incoming MessageHead { version: Http11, subject: RawStatus(200, "OK"), headers: Headers...
I'm still getting this error on 220d09f: ``` thread 'hyper-client' panicked at 'handler not in queue for key', ../src/libcore/option.rs:699 stack backtrace: 1: 0x101062fab - std::sys::backtrace::tracing::imp::write::h3800f45f421043b8 2: 0x101065135 - std::panicking::default_hook::_$u7b$$u7b$closure...
2016-07-13T19:24:29Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
856
hyperium__hyper-856
[ "848" ]
5f273ef646a9e7b7154ecd52b1f65388b77ea422
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -5,6 +5,7 @@ use std::collections::HashMap; use std::fmt; +use std::io; use std::marker::PhantomData; use std::sync::mpsc; use std::thread; diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod...
diff --git a/src/http/mod.rs b/src/http/mod.rs --- a/src/http/mod.rs +++ b/src/http/mod.rs @@ -410,6 +411,19 @@ impl Next { } } +impl Next_ { + fn register(&self) -> Reg { + match *self { + Next_::Read => Reg::Read, + Next_::Write => Reg::Write, + Next_::ReadWrite => R...
Client panics with 'handler not in queue for key' Using latest master. I have a `Client` that sends HTTP POST requests to the same URL from multiple threads and sometime this happens: ``` 2016-06-29 19:39:49 TRACE:hyper::client: on_incoming MessageHead { version: Http11, subject: RawStatus(200, "OK"), headers: Headers...
2016-07-08T17:16:53Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
853
hyperium__hyper-853
[ "651" ]
b47affd94b1ca53d04642202728f194a1ceec04d
diff --git a/src/header/common/host.rs b/src/header/common/host.rs --- a/src/header/common/host.rs +++ b/src/header/common/host.rs @@ -1,6 +1,8 @@ use header::{Header, HeaderFormat}; use std::fmt; +use std::str::FromStr; use header::parsing::from_one_raw_str; +use url::idna::domain_to_unicode; /// The `Host` head...
diff --git a/src/header/common/host.rs b/src/header/common/host.rs --- a/src/header/common/host.rs +++ b/src/header/common/host.rs @@ -97,6 +62,35 @@ impl HeaderFormat for Host { } } +impl fmt::Display for Host { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + self.fmt_header(f) + } +} +...
Origin Header It seems like the "Origin" header is missing [here](http://hyper.rs/hyper/hyper/header/index.html) http://tools.ietf.org/id/draft-abarth-origin-03.html
Is this in common use? I've never used it myself, and when checking the dev consoles, I notice neither Firefox nor Chrome set that header while browsing. The Origin header is set for CORS requests. On 15 October 2015 01:36:27 CEST, Sean McArthur notifications@github.com wrote: > Is this in common use? I've never use...
2016-07-06T20:03:14Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
845
hyperium__hyper-845
[ "843" ]
e682844431f9d3e83f9dfc8bd5f7ef873b8a9b16
diff --git a/examples/hello.rs b/examples/hello.rs --- a/examples/hello.rs +++ b/examples/hello.rs @@ -14,7 +14,7 @@ static PHRASE: &'static [u8] = b"Hello World!"; struct Hello; impl Handler<HttpStream> for Hello { - fn on_request(&mut self, _: Request) -> Next { + fn on_request(&mut self, _: Request<HttpStr...
diff --git a/tests/server.rs b/tests/server.rs --- a/tests/server.rs +++ b/tests/server.rs @@ -99,7 +99,7 @@ impl TestHandler { } impl Handler<HttpStream> for TestHandler { - fn on_request(&mut self, _req: Request) -> Next { + fn on_request(&mut self, _req: Request<HttpStream>) -> Next { //self.tx.se...
Remote Address Might Not Always Available So looking more carefully at how hyper does its thing it looks like the remote address (from the transport) might not always be available. In the case of a GET request there likely won't be a body and it looks like https://github.com/hyperium/hyper/blob/master/src/server/mod.r...
I've also simply felt that it'd be better to have access to the `Transport` in the `on_request` event. So to do this, I've been thinking of adding a `&'a T` field to the `Request` struct. ``` rust fn on_request<'a>(&mut self, req: Request<'a>) -> Next { let addr = req.transport().peer_addr().unwrap(); } ``` I won...
2016-06-23T22:29:49Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
841
hyperium__hyper-841
[ "833", "821" ]
6dab63fbac546fee78dadad20ec2617c7396f15a
diff --git a/src/header/internals/item.rs b/src/header/internals/item.rs --- a/src/header/internals/item.rs +++ b/src/header/internals/item.rs @@ -78,6 +78,9 @@ impl Item { Err(_) => () } } + if self.raw.is_some() && self.typed.get_mut(tid).is_some() { + self.raw...
diff --git a/src/header/mod.rs b/src/header/mod.rs --- a/src/header/mod.rs +++ b/src/header/mod.rs @@ -670,6 +670,7 @@ mod tests { fn test_get_mutable() { let mut headers = Headers::from_raw(&raw!(b"Content-Length: 10")).unwrap(); *headers.get_mut::<ContentLength>().unwrap() = ContentLength(20); ...
Kill raw part when getting mutable reference to typed header Fixes #821. get_mut doesn't seem to work I'm not sure what happened here: ``` rust extern crate hyper; use hyper::header::Headers; use hyper::header::{ContentLength, ContentType}; use hyper::http::RawStatus; use hyper::method::Method; use hyper::mime::{Attr...
I guess the mixture of raw and non-raw methods is breaking it, but I'm not sure. Does set_raw create two copies of the thing -- one typed, one untyped? I guess the issue is that typed_mut doesn't remove the header from the raw headers when inserting a typed one. Ah, this is a bug. The invariants I've been enforcing ...
2016-06-20T22:13:45Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
834
hyperium__hyper-834
[ "831" ]
43ac0dd09569f80625f7875ff27543bfaf339b99
diff --git a/src/http/conn.rs b/src/http/conn.rs --- a/src/http/conn.rs +++ b/src/http/conn.rs @@ -646,13 +646,6 @@ impl<H: MessageHandler<T>, T: Transport> State<H, T> { _ => Reading::Closed, }; let writing = match http1.writing { - Writing::Rea...
diff --git a/src/http/h1/encode.rs b/src/http/h1/encode.rs --- a/src/http/h1/encode.rs +++ b/src/http/h1/encode.rs @@ -335,7 +347,7 @@ mod tests { use mock::{Async, Buf}; #[test] - fn test_write_chunked_sync() { + fn test_chunked_encode_sync() { let mut dst = Buf::new(); let mut enco...
Client needs a way to "end" writing chunked requests
This cannot be `Next::end()`, as that signals the entire message is complete. There is need to signal "writing is done, but I still want to read from the server". Some options I've thought of: - Make the Handler write with zero bytes if chunked, ie `encoder.write(b"")` - Add a method to `Encoder<T>`, such that the han...
2016-06-17T11:08:53Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
827
hyperium__hyper-827
[ "804" ]
f20d5953c775868a139f3f55e7774cbbd6425dbf
diff --git a/src/http/buffer.rs b/src/http/buffer.rs --- a/src/http/buffer.rs +++ b/src/http/buffer.rs @@ -98,6 +98,12 @@ pub struct BufReader<'a, R: io::Read + 'a> { reader: &'a mut R } +impl<'a, R: io::Read + 'a> BufReader<'a, R> { + pub fn get_ref(&self) -> &R { + self.reader + } +} + impl<'a, ...
diff --git a/src/client/pool.rs /dev/null --- a/src/client/pool.rs +++ /dev/null @@ -1,291 +0,0 @@ -//! Client Connection Pooling -use std::borrow::ToOwned; -use std::collections::HashMap; -use std::io::{self, Read, Write}; -use std::net::{SocketAddr, Shutdown}; -use std::sync::{Arc, Mutex}; - -use std::time::Duration;...
Remote Address not set in Request The remote_addr member of hyper::server::Request is commented out and not yet set on each request. Presumably https://github.com/hyperium/hyper/blob/master/src/server/request.rs#L19 is commented out as a placeholder to be done for 0.10. Just making an issue so its not forgotten :+1: ...
2016-06-14T15:04:58Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
778
hyperium__hyper-778
[ "395" ]
1ec56fe6b68ecf74213cf7e285f80c3a4ee70ce9
diff --git a/.travis.yml b/.travis.yml --- a/.travis.yml +++ b/.travis.yml @@ -3,6 +3,8 @@ matrix: fast_finish: true include: - os: osx + rust: stable + env: FEATURES="--no-default-features --features security-framework" - rust: nightly env: FEATURES="--features ...
diff --git a/benches/client.rs /dev/null --- a/benches/client.rs +++ /dev/null @@ -1,111 +0,0 @@ -#![deny(warnings)] -#![feature(test)] -extern crate hyper; - -extern crate test; - -use std::fmt; -use std::io::{self, Read, Write, Cursor}; -use std::net::SocketAddr; -use std::time::Duration; - -use hyper::net; - -static...
Non-blocking/Evented I/O Hyper would be far more powerful of a client & server if it was based on traditional event-oriented I/O, single-threaded or multi-threaded. You should look into https://github.com/carllerche/mio or a wrapper around libuv or something of that sort. Another option is to split hyper up into mult...
We agree. We're actively looking into it. Mio looks promising. We also need a Windows library, and a wrapper combining the two. On Tue, Mar 24, 2015, 6:06 AM Jarred Nicholls notifications@github.com wrote: > Hyper would be far more powerful of a client & server if it was based on > traditional event-oriented I/O, sin...
2016-05-04T19:40:53Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
775
hyperium__hyper-775
[ "774" ]
3a3e08687b9711edcc2d08e1f307805322ff68bb
diff --git a/examples/client.rs b/examples/client.rs --- a/examples/client.rs +++ b/examples/client.rs @@ -20,7 +20,20 @@ fn main() { } }; - let client = Client::new(); + let client = match env::var("HTTP_PROXY") { + Ok(mut proxy) => { + // parse the proxy, message if it doesn't ...
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -479,8 +483,9 @@ mod tests { use std::io::Read; use header::Server; use http::h1::Http11Message; - use mock::{MockStream}; + use mock::{MockStream, MockSsl}; use super::{Client, RedirectPolicy...
Proxy handling does not properly set destination Host header Here is a script transcript of a working proxy session ``` Script started on Wed 27 Apr 2016 09:32:58 AM PDT [~/repos/rust/hyper-upstream] ~$ telnet my-proxy 8080 Trying x.x.x.x... Connected to my-proxy. Escape character is '^]'. GET http://xkcd.com http/1.1...
I see, I misunderstood the spec. I understood that the `Host` header is the host of the proxy server, with the absolute-uri in the request line aimed at the target host. Re-reading the spec, it seems I am indeed incorrect, as HTTP/1.0 proxies may just forward the `Host` header directly. It may be that the proxy you ar...
2016-04-27T18:21:16Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
771
hyperium__hyper-771
[ "531" ]
4828437551c7f5ed3f54acb1c1bf1fd50a6a3516
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -55,9 +55,9 @@ //! clone2.post("http://example.domain/post").body("foo=bar").send().unwrap(); //! }); //! ``` +use std::borrow::Cow; use std::default::Default; use std::io::{self, copy, Read}; -use std::iter::...
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -456,6 +478,8 @@ fn get_host_and_port(url: &Url) -> ::Result<(&str, u16)> { mod tests { use std::io::Read; use header::Server; + use http::h1::Http11Message; + use mock::{MockStream}; use super::{...
Add Proxy support to Client I imagine some sort of `Proxy` type being added a property of `Client`, with a corresponding `set_proxy(Option<Proxy>)` method. Use of a proxy can have different requirements: all requests, only https requests, only http... So it would probably mean `Proxy` would be an enum... and the proxy...
It would be awesome to have this feature! Yes please. Some of us need this to use Rust at work. Something like: ``` let mut client = Client::new(); // after checking config, env for HTTP_PROXY, whatever let proxy = Proxy::new(server, port, scheme); // AuthProxy::new for the more complicated cases client.add_proxy(p...
2016-04-25T22:51:19Z
0.9
220d09fc3a68cdd9a10589b2b454b54312518bb3
hyperium/hyper
750
hyperium__hyper-750
[ "747" ]
c85b056cabd1bbe62340af03696810c7e155fe17
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -44,6 +44,8 @@ pub use self::if_range::IfRange; pub use self::last_modified::LastModified; pub use self::location::Location; pub use self::pragma::Pragma; +pub use self::prefer::{Prefer, ...
diff --git /dev/null b/src/header/common/prefer.rs new file mode 100644 --- /dev/null +++ b/src/header/common/prefer.rs @@ -0,0 +1,203 @@ +use std::fmt; +use std::str::FromStr; +use header::{Header, HeaderFormat}; +use header::parsing::{from_comma_delimited, fmt_comma_delimited}; + +/// `Prefer` header, defined in [RFC...
Add prefer header I'd like to use the [`Prefer`](http://tools.ietf.org/html/rfc7240) header, could types for this header be added?
This is the first I've heard of this header. Is there a situation where its usage is common? If you look at the [PostgREST](http://postgrest.com/api/reading/) project you'll see it's being used to suppress counts and in a few other cases. The [spec](http://tools.ietf.org/html/rfc7240#section-2.1) also gives some inter...
2016-03-24T18:12:17Z
0.8
c85b056cabd1bbe62340af03696810c7e155fe17
hyperium/hyper
743
hyperium__hyper-743
[ "683" ]
028f5864324f95d4d9007c304358b6e7b91743fc
diff --git a/src/header/common/cache_control.rs b/src/header/common/cache_control.rs --- a/src/header/common/cache_control.rs +++ b/src/header/common/cache_control.rs @@ -1,7 +1,7 @@ use std::fmt; use std::str::FromStr; use header::{Header, HeaderFormat}; -use header::parsing::{from_one_comma_delimited, fmt_comma_de...
diff --git a/src/header/common/content_length.rs b/src/header/common/content_length.rs --- a/src/header/common/content_length.rs +++ b/src/header/common/content_length.rs @@ -79,7 +79,16 @@ __hyper__tm!(ContentLength, tests { test_header!(test1, vec![b"3495"], Some(HeaderField(3495))); test_header!(test_inv...
Response with multiple `Transfer-Encoding: chunked` header fields not parsed as chunk I've adapted this code from `client/response.rs`: ``` rust fn main() { let stream = MockStream::with_input(b"\ HTTP/1.1 200 OK\r\n\ Transfer-Encoding: chunked\r\n\ Transfer-Encoding: chunked\r\n\ \...
Indeed, it would seem the bug is in the parsing of the `TransferEncoding` header. It should handle more than 1 field occurrence. @seanmonstar I'm running into this issue myself, but I'm not sure how/where to best fix it. Any pointers where I should start looking? This occurs in the parsing. `TransferEncoding` uses [...
2016-03-11T20:08:29Z
0.8
c85b056cabd1bbe62340af03696810c7e155fe17
hyperium/hyper
718
hyperium__hyper-718
[ "715" ]
a4230eb510224b3a52e0f2d616fade8f0e8313b9
diff --git a/src/buffer.rs b/src/buffer.rs --- a/src/buffer.rs +++ b/src/buffer.rs @@ -79,7 +79,7 @@ unsafe fn grow_zerofill(buf: &mut Vec<u8>, additional: usize) { use std::ptr; let len = buf.len(); buf.set_len(len + additional); - ptr::write_bytes(buf.as_mut_ptr(), 0, buf.len()); + ptr::write_byt...
diff --git a/src/buffer.rs b/src/buffer.rs --- a/src/buffer.rs +++ b/src/buffer.rs @@ -151,4 +151,14 @@ mod tests { assert_eq!(rdr.pos, 0); assert_eq!(rdr.cap, 0); } + + #[test] + fn test_resize() { + let raw = b"hello world"; + let mut rdr = BufReader::with_capacity(&raw[..],...
Odd error when unwrapping a response I'm getting a weird error when trying to get data from this API. Any help would be appreciated. ``` rust let client = Client::new(); let mut res = client.get(&"https://data.medicare.gov/resource/pqp8-xrjv.json?$limit=1".to_string()) .send().unwrap(); ``` ``` thread '<main>' pani...
Hm, what version of hyper? I just tried that URL using https://github.com/hyperium/hyper/blob/master/examples/client.rs and it worked for me. Was on 0.6 when I posted the issue. Updated to 0.7 and still getting the same thing. Here's the log incase it helps ``` TRACE:hyper::header: Headers.set( "Connection", Connec...
2016-01-04T23:01:42Z
0.7
a4230eb510224b3a52e0f2d616fade8f0e8313b9
hyperium/hyper
699
hyperium__hyper-699
[ "698" ]
3f1b13c72d925b80601ab5468dbe36273622b451
diff --git a/src/server/request.rs b/src/server/request.rs --- a/src/server/request.rs +++ b/src/server/request.rs @@ -9,7 +9,7 @@ use std::time::Duration; use buffer::BufReader; use net::NetworkStream; use version::{HttpVersion}; -use method::Method::{self, Get, Head}; +use method::Method; use header::{Headers, Co...
diff --git a/src/server/request.rs b/src/server/request.rs --- a/src/server/request.rs +++ b/src/server/request.rs @@ -159,6 +157,24 @@ mod tests { assert_eq!(read_to_string(req).unwrap(), "".to_owned()); } + #[test] + fn test_get_with_body() { + let mut mock = MockStream::with_input(b"\ + ...
Request body should be allowed with GET/HEAD requests Currently, the body is ignored when the request method is either GET or HEAD. I had a look through the HTTP 1.1 spec and didn't find any mention of restricting request bodies to certian methods (http://www.w3.org/Protocols/rfc2616/rfc2616-sec4.html#sec4.3). I hav...
2616 has been [superseded](https://www.mnot.net/blog/2014/06/07/rfc2616_is_dead). See http://tools.ietf.org/html/rfc7231#section-4.3.2 for the relevant MUST NOT.
2015-11-29T17:10:41Z
0.7
a4230eb510224b3a52e0f2d616fade8f0e8313b9
hyperium/hyper
693
hyperium__hyper-693
[ "655" ]
072d4fe36af21e1551316242ba5ebb5a011d38ca
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -7,6 +7,7 @@ //! is used, such as `ContentType(pub Mime)`. pub use self::accept::Accept; +pub use self::access_control_allow_credentials::AccessControlAllowCredentials; pub use self::ac...
diff --git /dev/null b/src/header/common/access_control_allow_credentials.rs new file mode 100644 --- /dev/null +++ b/src/header/common/access_control_allow_credentials.rs @@ -0,0 +1,89 @@ +use std::fmt::{self, Display}; +use std::str; +use unicase::UniCase; +use header::{Header, HeaderFormat}; + +/// `Access-Control-A...
Access-Control-Allow-Credentials header the Access-Control-Allow-Credentials header is missing http://www.w3.org/TR/2008/WD-access-control-20080912/#access-control-allow-credentials ABNF: ``` Access-Control-Allow-Credentials: "Access-Control-Allow-Credentials" ":" "true" ```
2015-11-22T08:55:39Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
687
hyperium__hyper-687
[ "686" ]
d44ee5980f744b34bc33ad34b1ae2cafcf7ad6e7
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -25,11 +25,11 @@ unicase = "1.0" url = "0.2" [dependencies.cookie] -version = "0.1" +version = "0.2" default-features = false [dependencies.openssl] -version = "0.6.4" +version = "0.7" optional = true [dependencies.solicit] diff --git...
diff --git a/src/header/common/set_cookie.rs b/src/header/common/set_cookie.rs --- a/src/header/common/set_cookie.rs +++ b/src/header/common/set_cookie.rs @@ -142,7 +140,7 @@ impl SetCookie { #[test] fn test_parse() { let h = Header::parse_header(&[b"foo=bar; HttpOnly".to_vec()][..]); - let mut c1 = Cookie::n...
Can we bump the version of the cookie crate from 0.1 to 0.2? I'm trying to build https://github.com/cyderize/rust-websocket. Now that fails (see https://github.com/cyderize/rust-websocket/issues/54) because websocket wants a newer version of openssl than hyper/cookie does, which causes a conflict. Websocket could of c...
2015-11-20T19:15:49Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
680
hyperium__hyper-680
[ "561" ]
b4a9227204e1c08b047643b0c2f655fbf49e30b0
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -23,6 +23,7 @@ pub use self::allow::Allow; pub use self::authorization::{Authorization, Scheme, Basic, Bearer}; pub use self::cache_control::{CacheControl, CacheDirective}; pub use self::...
diff --git /dev/null b/src/header/common/content_disposition.rs new file mode 100644 --- /dev/null +++ b/src/header/common/content_disposition.rs @@ -0,0 +1,328 @@ +// # References +// +// "The Content-Disposition Header Field" https://www.ietf.org/rfc/rfc2183.txt +// "The Content-Disposition Header Field in the Hypert...
Content-Disposition header Not strictly part of HTTP/1.1 but comes up in common usage. See: - RFC 6266 - Browser conformance tests at: http://greenbytes.de/tech/tc2231/ - IANA assignment at: http://www.iana.org/assignments/cont-disp/cont-disp.xhtml (I'm just documenting this. I'm creating a heavily stripped down vari...
If anyone takes this on, you might want to start with the one at https://github.com/mikedilger/formdata/blob/master/src/headers.rs and then extend it to comply with the standards listed above.
2015-11-11T20:06:56Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
677
hyperium__hyper-677
[ "673" ]
becced4ef293d131ae2ca245621e6ac65bb32c29
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -10,6 +10,7 @@ pub use self::accept::Accept; pub use self::access_control_allow_headers::AccessControlAllowHeaders; pub use self::access_control_allow_methods::AccessControlAllowMethods; ...
diff --git /dev/null b/src/header/common/access_control_expose_headers.rs new file mode 100644 --- /dev/null +++ b/src/header/common/access_control_expose_headers.rs @@ -0,0 +1,60 @@ +use unicase::UniCase; + +header! { + /// `Access-Control-Expose-Headers` header, part of + /// [CORS](http://www.w3.org/TR/cors/#a...
Header request: Access-Control-Expose-Headers http://www.w3.org/TR/cors/#access-control-expose-headers-response-header
2015-11-02T21:04:17Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
647
hyperium__hyper-647
[ "646" ]
32e09a04292b0247456a8fb9003a75a6abaa998e
diff --git a/src/http/h1.rs b/src/http/h1.rs --- a/src/http/h1.rs +++ b/src/http/h1.rs @@ -225,6 +225,7 @@ impl HttpMessage for Http11Message { SizedReader(stream, len) } else if headers.has::<ContentLength>() { trace!("illegal Content-Length: {:?}", headers.get_raw("Conte...
diff --git a/src/http/h1.rs b/src/http/h1.rs --- a/src/http/h1.rs +++ b/src/http/h1.rs @@ -893,10 +894,12 @@ mod tests { use std::error::Error; use std::io::{self, Read, Write}; + use buffer::BufReader; use mock::MockStream; + use http::HttpMessage; - use super::{read_chunk_size, parse_req...
Panic when getting empty response from server As noted in this other (closed) issue: https://github.com/hyperium/hyper/issues/365 When receiving empty response data, hyper consistently panics with: ``` thread '<unnamed>' panicked at 'thread '<unnamed>Http11Message lost its underlying stream somehow.', /home/jneves/.c...
Yep, try updating hyper, this was fixed in a patch. On Fri, Sep 4, 2015, 12:26 AM João Neves notifications@github.com wrote: > As noted in this other (closed) issue: #365 > https://github.com/hyperium/hyper/issues/365 > > When receiving empty response data, hyper consistently panics with: > > thread '<unnamed>' pan...
2015-09-04T23:13:55Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
645
hyperium__hyper-645
[ "640" ]
8dbc38c75acfc93557f7acdba58edeb8d27dfa2d
diff --git a/src/client/pool.rs b/src/client/pool.rs --- a/src/client/pool.rs +++ b/src/client/pool.rs @@ -34,7 +34,7 @@ impl Default for Config { #[derive(Debug)] struct PoolImpl<S> { - conns: HashMap<Key, Vec<S>>, + conns: HashMap<Key, Vec<PooledStreamInner<S>>>, config: Config, } diff --git a/src/c...
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -446,8 +446,10 @@ fn get_host_and_port(url: &Url) -> ::Result<(String, u16)> { #[cfg(test)] mod tests { + use std::io::Read; use header::Server; use super::{Client, RedirectPolicy}; + use super::poo...
Client Connection Pool wrongly handles HEAD response with body
2015-09-01T23:49:03Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
636
hyperium__hyper-636
[ "635" ]
44a4010537dd9abbdf803061226505b56a053bc8
diff --git a/src/header/common/connection.rs b/src/header/common/connection.rs --- a/src/header/common/connection.rs +++ b/src/header/common/connection.rs @@ -4,6 +4,9 @@ use unicase::UniCase; pub use self::ConnectionOption::{KeepAlive, Close, ConnectionHeader}; +const KEEP_ALIVE: UniCase<&'static str> = UniCase("...
diff --git a/src/header/common/connection.rs b/src/header/common/connection.rs --- a/src/header/common/connection.rs +++ b/src/header/common/connection.rs @@ -131,6 +136,7 @@ mod tests { fn test_parse() { assert_eq!(Connection::close(),parse_option(b"close".to_vec())); assert_eq!(Connection::keep...
Connection Header case insensitivity http://tools.ietf.org/html/rfc7230#section-6.1: ``` ... "Connection options are case-insensitive." ... ``` I noticed that Apache Bench sends "Keep-Alive" and hyper misinterprets it ``` DEBUG:hyper::server::request: Headers { Connection: Keep-Alive, Host: 192.168.66.62:8881, Accep...
2015-08-27T22:39:32Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
632
hyperium__hyper-632
[ "629" ]
da2d29309a986ba25c5af6e95c71d033c48e4b2c
diff --git a/src/uri.rs b/src/uri.rs --- a/src/uri.rs +++ b/src/uri.rs @@ -1,4 +1,5 @@ //! HTTP RequestUris +use std::fmt::{Display, self}; use std::str::FromStr; use url::Url; use url::ParseError as UrlError;
diff --git a/src/uri.rs b/src/uri.rs --- a/src/uri.rs +++ b/src/uri.rs @@ -72,6 +73,17 @@ impl FromStr for RequestUri { } } +impl Display for RequestUri { + fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { + match *self { + RequestUri::AbsolutePath(ref path) => f.write_str(path), + ...
implement fmt::Display for RequestUri
2015-08-18T20:29:29Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
625
hyperium__hyper-625
[ "436" ]
af062ac954d5b90275138880ce2f5013d6664b5a
diff --git a/src/client/response.rs b/src/client/response.rs --- a/src/client/response.rs +++ b/src/client/response.rs @@ -47,9 +47,9 @@ impl Response { version: version, headers: headers, url: url, - message: message, status_raw: raw_status, - i...
diff --git a/src/header/common/content_length.rs b/src/header/common/content_length.rs --- a/src/header/common/content_length.rs +++ b/src/header/common/content_length.rs @@ -1,37 +1,86 @@ -header! { - #[doc="`Content-Length` header, defined in"] - #[doc="[RFC7230](http://tools.ietf.org/html/rfc7230#section-3.3.2...
When a server returns NoContent, reading the response should not hang When a client makes a request that results in a `NoContent` status code, the read methods on the repsonse should return an immediate EOF status, rather than blocking.
2015-08-05T23:47:36Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
621
hyperium__hyper-621
[ "315" ]
421422b620206c43bc85e05b973042da1b20f130
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -47,5 +47,5 @@ env_logger = "*" default = ["ssl"] ssl = ["openssl", "cookie/secure"] serde-serialization = ["serde"] -nightly = [] - +timeouts = [] +nightly = ["timeouts"] diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod...
diff --git a/src/lib.rs b/src/lib.rs --- a/src/lib.rs +++ b/src/lib.rs @@ -2,6 +2,7 @@ #![cfg_attr(test, deny(missing_docs))] #![cfg_attr(test, deny(warnings))] #![cfg_attr(all(test, feature = "nightly"), feature(test))] +#![cfg_attr(feature = "timeouts", feature(duration, socket_timeout))] //! # Hyper //! diff ...
Support for pervasive timeouts We need to provide timeouts for all of our blocking APIs and ways to set timeouts for all internal blocking actions.
Indeed, and we had a PR offering timeouts, but the IO reform currently means there is no API to do so for TcpStreams. There is now going to be a stable way to timeout on condition variables, so we could implement _something_. Link? On Wed, Apr 1, 2015, 2:13 PM Jonathan Reem notifications@github.com wrote: > There i...
2015-07-27T16:58:37Z
0.6
072d4fe36af21e1551316242ba5ebb5a011d38ca
hyperium/hyper
567
hyperium__hyper-567
[ "468" ]
c37d85728f67997cef4887736c5228322c8967ec
diff --git a/src/header/common/accept.rs b/src/header/common/accept.rs --- a/src/header/common/accept.rs +++ b/src/header/common/accept.rs @@ -27,6 +27,51 @@ header! { #[doc="* `audio/*; q=0.2, audio/basic` (`*` value won't parse correctly)"] #[doc="* `text/plain; q=0.5, text/html, text/x-dvi; q=0.8, text/x-c...
diff --git a/src/header/common/accept_charset.rs b/src/header/common/accept_charset.rs --- a/src/header/common/accept_charset.rs +++ b/src/header/common/accept_charset.rs @@ -18,6 +18,35 @@ header! { #[doc=""] #[doc="# Example values"] #[doc="* `iso-8859-5, unicode-1-1;q=0.8`"] + #[doc=""] + #[doc=...
Add example usage for each Header The header types docs could show an example usage on each of their respective pages.
This is solved by #503 Ah, when I meant example usage, I meant showing in a rust code fench how to create a header to to send in a Request/Response.
2015-06-15T03:44:31Z
0.5
c37d85728f67997cef4887736c5228322c8967ec
hyperium/hyper
518
hyperium__hyper-518
[ "495" ]
38f40c7f6a17bd33951bbab1adddf542cbb96de6
diff --git a/benches/client.rs b/benches/client.rs --- a/benches/client.rs +++ b/benches/client.rs @@ -8,7 +8,7 @@ use std::fmt; use std::io::{self, Read, Write, Cursor}; use std::net::SocketAddr; -use hyper::net; +use hyper::net::{self, ContextVerifier}; static README: &'static [u8] = include_bytes!("../README....
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -401,6 +406,8 @@ mod tests { use header::Server; use super::{Client, RedirectPolicy}; use url::Url; + use mock::ChannelMockConnector; + use std::sync::mpsc::{self, TryRecvError}; mock_connec...
Client: Setting a custom SSL verifier discards previous Connector When the `Client::set_ssl_verifier` method is called, it will discard any previously existing `Connector` instance that the `Client` was instantiated to use. This wasn't a problem before, when the connectors were largely stateless, but with the introduct...
This is true. Before, it was possible when `Client` was generic over the NetworkConnector. Perhaps such a method `set_ssl_verifier` should be added to the `NetworkConnector` trait...
2015-05-09T18:02:10Z
0.4
38f40c7f6a17bd33951bbab1adddf542cbb96de6
hyperium/hyper
499
hyperium__hyper-499
[ "480" ]
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -42,7 +42,7 @@ pub use self::referer::Referer; pub use self::server::Server; pub use self::set_cookie::SetCookie; pub use self::transfer_encoding::TransferEncoding; -pub use self::upgrade...
diff --git a/src/header/common/upgrade.rs b/src/header/common/upgrade.rs --- a/src/header/common/upgrade.rs +++ b/src/header/common/upgrade.rs @@ -31,47 +29,107 @@ header! { (Upgrade, "Upgrade") => (Protocol)+ test_upgrade { + // Testcase from the RFC test_header!( test1, ...
Reminder: Upgrade header Protocols should support version As defined in http://tools.ietf.org/html/rfc7230#section-6.7 a protocol consists of a name and an optional version token.
I don't understand what the action is for this issue. Does the `Upgrade` header need to be changed? Yes we need to change `Upgrade` header field. (What means "need" we can also keep it but do not implement the RFC) The RFC says: ``` Upgrade = 1#protocol protocol = protocol-name ["/" protoc...
2015-05-02T17:30:05Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
498
hyperium__hyper-498
[ "497" ]
6d7ae81e36096947b28ce5a5414df3c75d1fbae9
diff --git a/src/http.rs b/src/http.rs --- a/src/http.rs +++ b/src/http.rs @@ -409,8 +409,8 @@ impl<'a> TryParse for httparse::Response<'a, 'a> { httparse::Status::Complete(len) => { let code = res.code.unwrap(); let reason = match StatusCode::from_u16(code).canonical_reas...
diff --git a/src/http.rs b/src/http.rs --- a/src/http.rs +++ b/src/http.rs @@ -460,7 +460,7 @@ mod tests { use buffer::BufReader; use mock::MockStream; - use super::{read_chunk_size, parse_request}; + use super::{read_chunk_size, parse_request, parse_response}; #[test] fn test_write_chunke...
Real status text gets lost This is blocking the Servo rustup. Code went from ``` rust let reason = match str::from_utf8(cursor.into_inner()) { Ok(s) => s.trim(), Err(_) => return Err(HttpStatusError) }; let reason = match from_u16::<StatusCode>(code) { Some(status) => match status....
For what is the reason phrase needed in servo? https://xhr.spec.whatwg.org/#the-statustext-attribute
2015-05-01T22:09:54Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
492
hyperium__hyper-492
[ "490" ]
1426a4ce343bb4533234ac51e7f10d110ec701a6
diff --git /dev/null b/examples/headers.rs new file mode 100644 --- /dev/null +++ b/examples/headers.rs @@ -0,0 +1,13 @@ +#![deny(warnings)] + +#[macro_use] +// TODO: only import header!, blocked by https://github.com/rust-lang/rust/issues/25003 +extern crate hyper; + +// A header in the form of `X-Foo: some random str...
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -95,6 +95,21 @@ macro_rules! deref( } ); +macro_rules! tm { + ($id:ident, $tm:ident{$($tf:item)*}) => { + #[allow(unused_imports)] + mod $tm{ + use std::s...
Breaking change in the header! macro not documented I updated to 0.3.15 and my code is now failing due to changes in `header!` macro: ``` src/lib.rs:250:43: 250:44 error: unexpected end of macro invocation src/lib.rs:250 header!{ (TenantId, "TenantId") => [String] } ``` This is most likely due to commits https://gith...
cc @pyfisch as he is the one which added the commits mentioned above I'd say this was an oversight. We shouldn't have made a breaking change. A reason making it easier to miss is that all of our internal uses _do_ include a test function. I think part of a fix so as to make sure we don't break the exported `header!` ...
2015-04-30T22:05:09Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
491
hyperium__hyper-491
[ "388" ]
92ee51acdbf9ac1e2f2d3770f5566196c13b20bd
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -34,6 +34,7 @@ pub use self::if_match::IfMatch; pub use self::if_modified_since::IfModifiedSince; pub use self::if_none_match::IfNoneMatch; pub use self::if_unmodified_since::IfUnmodified...
diff --git /dev/null b/src/header/common/if_range.rs new file mode 100644 --- /dev/null +++ b/src/header/common/if_range.rs @@ -0,0 +1,73 @@ +use header::{self, EntityTag, HttpDate}; + +/// `If-Range` header, defined in [RFC7233](http://tools.ietf.org/html/rfc7233#section-3.2) +/// +/// If a client has a partial copy o...
Implement If-Range header parsing Spec: `http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.27`
RFC2616 is obsolte. New and better spec: https://tools.ietf.org/html/rfc7233#section-3.2 I will do this.
2015-04-30T17:51:04Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
486
hyperium__hyper-486
[ "41" ]
362044c32c2e6a5dcb701c48339d0dddde27effc
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -46,15 +46,17 @@ use status::StatusClass::Redirection; use {Url, HttpResult}; use HttpError::HttpUriError; +pub use self::pool::Pool; pub use self::request::Request; pub use self::response::Response; +pub mod p...
diff --git /dev/null b/src/client/pool.rs new file mode 100644 --- /dev/null +++ b/src/client/pool.rs @@ -0,0 +1,227 @@ +//! Client Connection Pooling +use std::borrow::ToOwned; +use std::collections::HashMap; +use std::io::{self, Read, Write}; +use std::net::{SocketAddr, Shutdown}; +use std::sync::{Arc, Mutex}; + +use...
Implement keep-alive
This is really important for SSL requests. server side is done
2015-04-28T01:26:13Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
485
hyperium__hyper-485
[ "475" ]
f7f0361626a808293e52cbe3fc07bc89490a7c03
diff --git a/src/header/common/accept_language.rs b/src/header/common/accept_language.rs --- a/src/header/common/accept_language.rs +++ b/src/header/common/accept_language.rs @@ -1,39 +1,4 @@ -use header::QualityItem; -use std::str::FromStr; -use std::fmt; - -/// A language tag. -/// See http://www.w3.org/Protocols/rfc...
diff --git a/src/header/common/accept_language.rs b/src/header/common/accept_language.rs --- a/src/header/common/accept_language.rs +++ b/src/header/common/accept_language.rs @@ -57,7 +22,7 @@ header! { #[cfg(test)] mod tests { - use header::{Header, qitem, Quality, QualityItem}; + use header::{Header, Langua...
Implement Content-Language header https://tools.ietf.org/html/rfc7231#section-3.1.3.2 This will be useful for Servo.
2015-04-27T19:12:52Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
455
hyperium__hyper-455
[ "437" ]
dac2f4db8ac4cd3e529a8a8333a575dbb5b37839
diff --git a/src/http.rs b/src/http.rs --- a/src/http.rs +++ b/src/http.rs @@ -12,7 +12,7 @@ use method::Method; use status::StatusCode; use uri::RequestUri; use version::HttpVersion::{self, Http10, Http11}; -use HttpError:: HttpTooLargeError; +use HttpError::{HttpIoError, HttpTooLargeError}; use {HttpError, HttpRe...
diff --git a/src/http.rs b/src/http.rs --- a/src/http.rs +++ b/src/http.rs @@ -440,8 +447,10 @@ pub struct RawStatus(pub u16, pub Cow<'static, str>); mod tests { use std::io::{self, Write}; - use super::{read_chunk_size}; + use buffer::BufReader; + use mock::MockStream; + use super::{read_chunk_si...
HttpTooLargeError after the response Ever since the new buffering code, I get `HttpTooLargeError`s after almost every response. It is after the handler returns and the response has been sent, and so the site continues to function. But clearly something is amiss. The requests are rather normal small requests, nothin...
I maked a pull request: https://github.com/hyperium/hyper/pull/454
2015-04-15T18:20:52Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
407
hyperium__hyper-407
[ "406" ]
ce5231508e615276333dc2dff7c8c27a6716a93d
diff --git a/src/buffer.rs b/src/buffer.rs --- a/src/buffer.rs +++ b/src/buffer.rs @@ -27,7 +27,12 @@ impl<R: Read> BufReader<R> { pub fn get_mut(&mut self) -> &mut R { &mut self.inner } pub fn get_buf(&self) -> &[u8] { - self.buf.get_ref() + let pos = self.buf.position() as usize; + if...
diff --git a/src/buffer.rs b/src/buffer.rs --- a/src/buffer.rs +++ b/src/buffer.rs @@ -93,3 +98,18 @@ fn reserve(v: &mut Vec<u8>) { v.reserve(cmp::min(cap * 4, MAX_BUFFER_SIZE) - cap); } } + +#[cfg(test)] +mod tests { + + use std::io::BufRead; + use super::BufReader; + + #[test] + fn test_con...
hyper 0.3.5 is looping handle() is being called repeatedly on just 1 request until some resource becomes exhausted and response.start() fails. 0.3.4 worked fine. I'll post more details once I look into the code.
This commit is the culprit: cb59f609c61a097d5d9fa728b9df33d79922573b fix(http): read more before triggering TooLargeError I wont try to fix, I'll leave it for you Sean as you're more familiar with what you just did. To test, the client that makes the server loop is a GET request via this firefox extension: http://ww...
2015-03-30T04:21:21Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
401
hyperium__hyper-401
[ "389" ]
b04f6d8e7ad9550588bceadcc0e21f1434ab244a
diff --git /dev/null b/src/buffer.rs new file mode 100644 --- /dev/null +++ b/src/buffer.rs @@ -0,0 +1,95 @@ +use std::cmp; +use std::iter; +use std::io::{self, Read, BufRead, Cursor}; + +pub struct BufReader<R> { + buf: Cursor<Vec<u8>>, + inner: R +} + +const INIT_BUFFER_SIZE: usize = 4096; +const MAX_BUFFER_SIZ...
diff --git a/src/client/response.rs b/src/client/response.rs --- a/src/client/response.rs +++ b/src/client/response.rs @@ -103,9 +104,10 @@ impl Read for Response { mod tests { use std::borrow::Cow::Borrowed; use std::boxed::BoxAny; - use std::io::{self, Read, BufReader}; + use std::io::{self, Read}; ...
Hyper client fails if the header isn't fully received in the first read Some HTTP servers do not send the entire HTTP header in a single chunk. Hyper assumes that the entire header is received when calling fill_buf, which isn't always the case. Running the client example in the documentation against a [Flask test serv...
True, I was hoping that the BufReader would fill it's entire buffer before yielding, which [by default is 64kb](https://github.com/rust-lang/rust/blob/master/src/libstd/io/mod.rs#L50). Is the test server head bigger than this? Or BufReader is yielding earlier? The head is a few bytes long. By looking at Wireshark I co...
2015-03-27T18:02:42Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
383
hyperium__hyper-383
[ "381" ]
7469e62d1e54d9b6f51e9435306e0455de05e1d0
diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -28,7 +28,7 @@ Hello World Server: extern crate hyper; use std::io::Write; -use std::net::IpAddr; +use std::net::Ipv4Addr; use hyper::Server; use hyper::server::Request; diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -4...
diff --git a/benches/client.rs b/benches/client.rs --- a/benches/client.rs +++ b/benches/client.rs @@ -1,5 +1,5 @@ #![deny(warnings)] -#![feature(collections, io, net, test)] +#![feature(collections, test)] extern crate hyper; extern crate test; diff --git a/benches/server.rs b/benches/server.rs --- a/benches/serv...
hyper doesn't built with new rustc. `IpAddr` was splitted to `Ipv4Addr` and `Ipv6Addr` ``` src\server/mod.rs:4:16: 4:22 error: unresolved import `std::net::IpAddr`. There is no `IpAddr` in `std::net` src\server/mod.rs:4 use std::net::{IpAddr, SocketAddr}; ``` rustc 1.0.0-dev (c10918905 2015-03-18) (built 2015-03-18) ...
2015-03-20T09:32:33Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
382
hyperium__hyper-382
[ "381", "381" ]
7469e62d1e54d9b6f51e9435306e0455de05e1d0
diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -28,7 +28,7 @@ Hello World Server: extern crate hyper; use std::io::Write; -use std::net::IpAddr; +use std::net::Ipv4Addr; use hyper::Server; use hyper::server::Request; diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -4...
diff --git a/benches/server.rs b/benches/server.rs --- a/benches/server.rs +++ b/benches/server.rs @@ -5,7 +5,7 @@ extern crate test; use test::Bencher; use std::io::{Read, Write}; -use std::net::IpAddr; +use std::net::Ipv4Addr; use hyper::method::Method::Get; use hyper::server::{Request, Response}; diff --git ...
hyper doesn't built with new rustc. `IpAddr` was splitted to `Ipv4Addr` and `Ipv6Addr` ``` src\server/mod.rs:4:16: 4:22 error: unresolved import `std::net::IpAddr`. There is no `IpAddr` in `std::net` src\server/mod.rs:4 use std::net::{IpAddr, SocketAddr}; ``` rustc 1.0.0-dev (c10918905 2015-03-18) (built 2015-03-18) ...
2015-03-19T08:16:38Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
377
hyperium__hyper-377
[ "369" ]
fe8c6d9914b7c213edd97fefeddc0ef831e02d4b
diff --git /dev/null b/src/header/common/expect.rs new file mode 100644 --- /dev/null +++ b/src/header/common/expect.rs @@ -0,0 +1,37 @@ +use std::fmt; + +use header::{Header, HeaderFormat}; + +/// The `Expect` header. +/// +/// > The "Expect" header field in a request indicates a certain set of +/// > behaviors (expec...
diff --git a/src/server/mod.rs b/src/server/mod.rs --- a/src/server/mod.rs +++ b/src/server/mod.rs @@ -184,11 +193,78 @@ pub trait Handler: Sync + Send { /// Receives a `Request`/`Response` pair, and should perform some action on them. /// /// This could reading from the request, and writing to the respo...
hyper does not comply with `Expect: 100-continue` statement I'm using hyper to receive githook commit messages and when the payload is longer than 1024 bytes, `request.read_to_string()` is much much slower than it probably should be (something like a factor 100 on localhost)... I have created a very simple test case i...
Hmm. Hyper is using a BufReader, which by [default uses 64kbs](https://github.com/rust-lang/rust/blob/master/src/libstd/io/mod.rs#L55). Is it significantly bigger than that? It is much smaller: 1024 bytes (as indicated by the title of the Issue :wink: ). And the issue starts getting visible at precisely this value. ...
2015-03-16T23:00:08Z
0.3
b916a7b18cdbbecc872dba8f4ed12d5697cc531c
hyperium/hyper
354
hyperium__hyper-354
[ "347" ]
7235d3f74a6f058a9bedc840519be80384e292da
diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -27,12 +27,13 @@ Hello World Server: ```rust extern crate hyper; -use hyper::status::StatusCode; -use hyper::server::Server; -use hyper::server::request::Request; -use hyper::server::response::Response; +use std::io::Write; +use std::net::IpAddr;...
diff --git a/benches/client.rs b/benches/client.rs --- a/benches/client.rs +++ b/benches/client.rs @@ -1,33 +1,53 @@ -#![feature(core, old_io, test)] +#![feature(collections, io, net, test)] extern crate hyper; extern crate test; use std::fmt; -use std::old_io::net::ip::Ipv4Addr; -use hyper::server::{Request, Re...
Switch to new std::io Installed Rust nightly build, after not using Rust for 3 weeks. Hyper won't compile any more. The I/O libraries changed. Again. Not your fault. Sigh. ``` Compiling hyper v0.2.1 /home/john/.cargo/registry/src/github.com-1ecc6299db9ec823/hyper-0.2.1/src/net.rs:215:25: 215:29 error: mismatched types...
2015-03-01T03:23:46Z
0.2
7235d3f74a6f058a9bedc840519be80384e292da
hyperium/hyper
331
hyperium__hyper-331
[ "326" ]
41fd8de243dc1183b994d46a9624b1ee6c2dcdd2
diff --git a/src/header/mod.rs b/src/header/mod.rs --- a/src/header/mod.rs +++ b/src/header/mod.rs @@ -206,6 +206,11 @@ impl Headers { self.data.insert(UniCase(name.into_cow()), Item::new_raw(value)); } + /// Remove a header set by set_raw + pub fn remove_raw(&mut self, name: &str) { + self...
diff --git a/src/header/mod.rs b/src/header/mod.rs --- a/src/header/mod.rs +++ b/src/header/mod.rs @@ -662,6 +667,14 @@ mod tests { assert_eq!(headers.get(), Some(&ContentLength(20))); } + #[test] + fn test_remove_raw() { + let mut headers = Headers::new(); + headers.set_raw("content...
Removing a raw header Currently, I can insert or overwrite a header by doing, for example: ``` request_headers.set_raw("Accept-Encoding", vec![b"gzip, deflate".to_vec()]); ``` But apparently, I cannot remove it, only set it to a different value. Wouldn't it be useful to add a remove_raw() method to Headers? Raw heade...
Good call, this would be needed to support XHR's `removeHeader` method.
2015-02-22T12:24:43Z
0.2
7235d3f74a6f058a9bedc840519be80384e292da
hyperium/hyper
329
hyperium__hyper-329
[ "328" ]
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
diff --git a/examples/client.rs b/examples/client.rs --- a/examples/client.rs +++ b/examples/client.rs @@ -1,4 +1,4 @@ -#![feature(env, io)] +#![feature(env, old_io)] extern crate hyper; use std::env; diff --git a/examples/hello.rs b/examples/hello.rs --- a/examples/hello.rs +++ b/examples/hello.rs @@ -1,4 +1,4 @@ ...
diff --git a/benches/client.rs b/benches/client.rs --- a/benches/client.rs +++ b/benches/client.rs @@ -1,4 +1,4 @@ -#![feature(core, io, test)] +#![feature(core, old_io, test)] extern crate hyper; extern crate test; diff --git a/benches/client_mock_tcp.rs b/benches/client_mock_tcp.rs --- a/benches/client_mock_tcp.r...
Build failure with rust nightly Hi, I'm not sure what this error means, or why there are so many of these "note: write `[...]` instead" lines without any souce file and line number information... ``` $ cargo build Compiling hyper v0.1.13 note: write `[..]` instead note: write `[..]` instead note: write `[..]` inste...
2015-02-21T23:07:22Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
319
hyperium__hyper-319
[ "314" ]
f8776f4c241e6c064b7cdcaa62e4eab180784c69
diff --git a/src/header/shared/quality_item.rs b/src/header/shared/quality_item.rs --- a/src/header/shared/quality_item.rs +++ b/src/header/shared/quality_item.rs @@ -26,6 +27,12 @@ impl<T> QualityItem<T> { } } +impl<T: PartialEq> cmp::PartialOrd for QualityItem<T> { + fn partial_cmp(&self, other: &QualityIt...
diff --git a/src/header/shared/quality_item.rs b/src/header/shared/quality_item.rs --- a/src/header/shared/quality_item.rs +++ b/src/header/shared/quality_item.rs @@ -5,6 +5,7 @@ use std::fmt; use std::str; +use std::cmp; #[cfg(test)] use super::encoding::*; /// Represents an item with a quality value as define...
Implement `PartialOrd` for `QualityItem` Sort `QualityItems` according to their `q` value. This is useful for comparison and to sort `QualityItems`. It makes it easier to find the preferred item, that matches the provided items.
2015-02-15T17:04:29Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
313
hyperium__hyper-313
[ "301" ]
f554c09e12da805a219675fe7eceeffca4e41fec
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -20,6 +20,7 @@ pub use self::date::Date; pub use self::etag::Etag; pub use self::expires::Expires; pub use self::host::Host; +pub use self::if_match::IfMatch; pub use self::if_modified_s...
diff --git /dev/null b/src/header/common/if_match.rs new file mode 100644 --- /dev/null +++ b/src/header/common/if_match.rs @@ -0,0 +1,68 @@ +use header::{EntityTag, Header, HeaderFormat}; +use header::parsing::{from_comma_delimited, fmt_comma_delimited, from_one_raw_str}; +use std::fmt; + +/// The `If-Match` header +/...
feat(headers): add IfMatch header Add support for the If-Match http header.
2015-02-14T20:00:41Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
310
hyperium__hyper-310
[ "309" ]
f554c09e12da805a219675fe7eceeffca4e41fec
diff --git a/src/http.rs b/src/http.rs --- a/src/http.rs +++ b/src/http.rs @@ -3,7 +3,10 @@ use std::borrow::Cow::{Borrowed, Owned}; use std::borrow::IntoCow; use std::cmp::min; use std::old_io::{self, Reader, IoResult, BufWriter}; +use std::old_io::util as io_util; +use std::mem; use std::num::from_u16; +use std::...
diff --git a/src/lib.rs b/src/lib.rs --- a/src/lib.rs +++ b/src/lib.rs @@ -1,5 +1,5 @@ #![feature(core, collections, hash, io, os, path, std_misc, - slicing_syntax, box_syntax)] + slicing_syntax, box_syntax, unsafe_destructor)] #![deny(missing_docs)] #![cfg_attr(test, deny(warnings))] #![cfg_att...
Read body for all requests with a Content-Length header The framing of a HTTP request doesn't depend on the method [1], so any request with a non-zero Content-Length has a corresponding body. This does not change in the case of idempotent methods like GET; what is different is that for such methods it is incorrect for ...
Good catch. I've also noticed that for requests that should have a body, such as POST, and the user doesn't read it out, it bleeds into the next request also. We could write a Drop impl for Request that tries to read out the remainder of the body.
2015-02-14T02:39:22Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
302
hyperium__hyper-302
[ "238" ]
f836ee89c13e6afa44429d3da2517530175bad25
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -21,6 +21,7 @@ pub use self::etag::Etag; pub use self::expires::Expires; pub use self::host::Host; pub use self::if_modified_since::IfModifiedSince; +pub use self::if_none_match::IfNoneMa...
diff --git /dev/null b/src/header/common/if_none_match.rs new file mode 100644 --- /dev/null +++ b/src/header/common/if_none_match.rs @@ -0,0 +1,84 @@ +use header::{Header, HeaderFormat, EntityTag}; +use header::parsing::{from_comma_delimited, fmt_comma_delimited, from_one_raw_str}; +use std::fmt::{self}; + +/// The `I...
add If-None-Match header This is needed for https://github.com/servo/servo/pull/4117/. Spec: https://tools.ietf.org/html/rfc7232#section-3.2
2015-02-07T20:57:28Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
292
hyperium__hyper-292
[ "291" ]
eee462568657ac54b1cf44d473f5227979241255
diff --git a/src/header/common/authorization.rs b/src/header/common/authorization.rs --- a/src/header/common/authorization.rs +++ b/src/header/common/authorization.rs @@ -32,9 +32,9 @@ impl<S: Scheme> Header for Authorization<S> { match (from_utf8(unsafe { &raw[].get_unchecked(0)[] }), Scheme::scheme(None:...
diff --git a/benches/client_mock_tcp.rs b/benches/client_mock_tcp.rs --- a/benches/client_mock_tcp.rs +++ b/benches/client_mock_tcp.rs @@ -1,4 +1,4 @@ -#![feature(core, collections, io, test)] +#![feature(collections, io, test)] extern crate hyper; extern crate test; diff --git a/src/header/mod.rs b/src/header/mod....
fix(rustup): update FromStr
There were the following issues with your Pull Request - Commit: 12746b6076b62dd8167af2670ac345dc2497c36f - Commits must be in the following format: **%{type}(%{scope}): %{description}** Guidelines are available at https://github.com/hyperium/hyper/blob/master/CONTRIBUTING.md --- This message was auto-generated by...
2015-02-04T02:58:09Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
289
hyperium__hyper-289
[ "252" ]
e9af13c3a6b1de4cc1f0774c085bd20ded4d41d6
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -15,7 +15,6 @@ keywords = ["http", "hyper", "hyperium"] cookie = "*" log = ">= 0.2.0" mime = "*" -mucell = "*" openssl = "*" rustc-serialize = "*" time = "*" diff --git /dev/null b/src/header/cell.rs new file mode 100644 --- /dev/null +++ b...
diff --git a/src/header/mod.rs b/src/header/mod.rs --- a/src/header/mod.rs +++ b/src/header/mod.rs @@ -695,37 +710,58 @@ mod tests { } #[bench] - fn bench_header_get(b: &mut Bencher) { + fn bench_headers_new(b: &mut Bencher) { + b.iter(|| { + let mut h = Headers::new(); + ...
Mucell often breaks Hyper Mucell breaks Hyper really often. Like for example just now: https://travis-ci.org/hyperium/hyper/builds/47357537 I do not really know what mucell makes it superior to the standard library types, but for me it looks like it only provides a minor speed increasement. So do we really need it or c...
@pyfisch it's fixed now, but I do see your concern. I guess a benchmark (of hyper) using both mucell and refcell would be useful to know if it's worth using something outside of std.
2015-02-02T20:16:36Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
283
hyperium__hyper-283
[ "281" ]
93821fc731fe00295dad13b488b8e4b4097d15a8
diff --git a/src/header/shared/quality_item.rs b/src/header/shared/quality_item.rs --- a/src/header/shared/quality_item.rs +++ b/src/header/shared/quality_item.rs @@ -28,7 +28,12 @@ impl<T> QualityItem<T> { impl<T: fmt::Display> fmt::Display for QualityItem<T> { fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Res...
diff --git a/src/header/shared/quality_item.rs b/src/header/shared/quality_item.rs --- a/src/header/shared/quality_item.rs +++ b/src/header/shared/quality_item.rs @@ -79,7 +84,7 @@ pub fn qitem<T>(item: T) -> QualityItem<T> { #[test] fn test_quality_item_show1() { let x = qitem(Chunked); - assert_eq!(format!(...
Don't print "; q=1" when QualityItem is set to 1 See https://github.com/servo/servo/issues/4734
2015-01-28T05:53:47Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
267
hyperium__hyper-267
[ "237" ]
fb92a260c0506f9d1764504527648a7bd7be3560
diff --git a/src/header/common/mod.rs b/src/header/common/mod.rs --- a/src/header/common/mod.rs +++ b/src/header/common/mod.rs @@ -23,6 +23,7 @@ pub use self::host::Host; pub use self::if_modified_since::IfModifiedSince; pub use self::last_modified::LastModified; pub use self::location::Location; +pub use self::prag...
diff --git /dev/null b/src/header/common/pragma.rs new file mode 100644 --- /dev/null +++ b/src/header/common/pragma.rs @@ -0,0 +1,61 @@ +use std::fmt; +use std::ascii::AsciiExt; + +use header::{Header, HeaderFormat, parsing}; + +/// The `Pragma` header defined by HTTP/1.0. +/// +/// > The "Pragma" header field allows ...
add Pragma header This is needed for https://github.com/servo/servo/pull/4117/. Spec: https://tools.ietf.org/html/rfc7234#section-5.4
2015-01-22T21:12:20Z
0.1
e8833c0c894a2b87b5d9b48b07858ec7fc890b7a
hyperium/hyper
213
hyperium__hyper-213
[ "211", "211" ]
27b262c22678ed785ed269da4f4a4ddc154526d1
diff --git a/src/http.rs b/src/http.rs --- a/src/http.rs +++ b/src/http.rs @@ -544,6 +544,9 @@ pub fn read_header<R: Reader>(stream: &mut R) -> HttpResult<Option<RawHeaderLine } }; } + // Remove optional trailing whitespace + let real_len = value.len() - value.iter().rev().take_while(|&...
diff --git a/src/header/mod.rs b/src/header/mod.rs --- a/src/header/mod.rs +++ b/src/header/mod.rs @@ -594,6 +594,13 @@ mod tests { assert!(headers.get::<CrazyLength>().is_none()); } + #[test] + fn test_trailing_whitespace() { + let headers = Headers::from_raw(&mut mem("Content-Length: 10 ...
Whitespace trailing header values causes incorrect parsing of header field [RFC7230](https://tools.ietf.org/html/rfc7230#section-3.2) allows optional trailing whitespace for header fields. So `Content-Length: 10 \r\nContent-Type: text/plain\r\n\r\n` is a valid header. It causes the parser to fail, it thinks the value ...
True! It seems like in [read_header](https://github.com/hyperium/hyper/blob/master/src/http.rs#L509) should trim the OWS. True! It seems like in [read_header](https://github.com/hyperium/hyper/blob/master/src/http.rs#L509) should trim the OWS.
2014-12-29T11:18:29Z
0.0
27b262c22678ed785ed269da4f4a4ddc154526d1
hyperium/hyper
206
hyperium__hyper-206
[ "205" ]
33f61213ce9d58722647a03f2f04514c1c14ade9
diff --git a/Cargo.toml b/Cargo.toml --- a/Cargo.toml +++ b/Cargo.toml @@ -19,6 +19,8 @@ typeable = "*" cookie = "*" time = "*" mucell = "*" +log = "*" +rustc-serialize = "*" [dev-dependencies] curl = "*" diff --git a/src/header/common/accept.rs b/src/header/common/accept.rs --- a/src/header/common/accept.rs +++...
diff --git a/src/client/mod.rs b/src/client/mod.rs --- a/src/client/mod.rs +++ b/src/client/mod.rs @@ -381,7 +381,7 @@ mod tests { client.set_redirect_policy(RedirectPolicy::FollowAll); let res = client.get("http://127.0.0.1").send().unwrap(); - assert_eq!(res.headers.get(), Some(&Server("moc...
Multiple matching crates for log I'm not sure if this is a cargo bug, but a rust project that has a log dependency causes hyper to not be compiled. Cargo confuses liblog in rustlib with the liblog that cargo provided. ``` ~/.cargo/registry/src/github.com-1ecc6299db9ec823/hyper-0.0.16/src/lib.rs:132:23: 132:40 error: m...
Same issue here. I've simply made a hello world application which depends on hyper. My Cargo's dependencies looks like this: ``` hyper = "0.0.16" ``` Yep, things moved around. I'm fixing for latest nightly currently.
2014-12-23T21:10:33Z
0.0
27b262c22678ed785ed269da4f4a4ddc154526d1