aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)Author
2022-03-08Add a test for HTTP/2 empty DATA frame with fin setLoïc Hoguin
2022-03-08Handle any zero-sized Data in http:data with finMartin Björklund
Amended to make the test case hit the problem.
2022-03-08Don't add default SNI if already setJohan Bevemyr
2022-03-08Add tests for SNILoïc Hoguin
2022-03-08gun_pool Dialyzer fixesSergei Shuvatov
2022-03-08Fix flushing by stream referenceSergei Shuvatov
2022-03-08Make the stream_error_* functions return 'ok'Viktor Söderqvist
This will make it more obvious what callbacks return.
2022-03-08Return commands instead of state in remaining callbacksViktor Söderqvist
2021-05-24Gun 2.0.0-rc.22.0.0-rc.2Loïc Hoguin
2021-02-12Temporarily depend on Cowlib masterLoïc Hoguin
Because we depend on Cowboy master for tests we also need the more recent Cowlib commits otherwise some tests will fail.
2021-02-10Depend on Ranch 2.0.0 for compatibility with OTP-24Loïc Hoguin
2021-02-09Fix gun_pool:stop_pool/2,3 with default portLoïc Hoguin
2021-02-08No longer use ssl:ssl_accept/2Loïc Hoguin
2021-02-08Fix type that breaks gun_pool compilationLoïc Hoguin
2021-02-08Fix a typo in migrating_from_1.3.asciidocErlend Hamberg
2021-02-07Initial commit for Gun poolsLoïc Hoguin
The approach taken here is very similar to what browsers are doing. A separate pool is created for each host/port/scope. The authority (host header) is used to determine which pool will execute requests. A connection process is semi-randomly chosen, from the connections that have capacity. Maximum capacity is determined by the protocol (the HTTP/2 setting set by the server is used, for example). Multiple processes can process requests/responses on the same connection concurrently. There is no need to "give back" the response to the pool, the number of ongoing streams is maintained via an event handler. The implementation is currently not strict, there may be more attempts to create requests than there is capacity. I'm not sure if it should be made strict or if Gun should just wait before sending requests (it only matters in the HTTP/2 case at the moment). When there is no connection with capacity available in the pool (because they have too many streams, or are reconnecting, or any other reason), checking out fails. There is no timeout to wait for a connection to be available. On the other hand the checkout_retry option allows setting multiple timeouts to retry checking out a connection. Each retry attempt's wait time can have a different value. The initial implementation of this work was sponsored by Kobil and made at the suggestion of Ilya Khaprov.
2020-11-19Grammar fix on most recent changelog2.0.0-rc.1Loïc Hoguin
2020-11-19Update Erlang.mk and configure HexLoïc Hoguin
2020-11-19Update Cowlib to 2.10.1Loïc Hoguin
2020-11-12Reset host/port/scheme/transport/intermediaries on disconnectLoïc Hoguin
I've made some more test timeouts infinity in order to fix additional intermittent issues that popped up.
2020-11-12Return 'undefined' for raw|socks origin_scheme where applicableLoïc Hoguin
2020-11-12Fix a low hanging fruit todoLoïc Hoguin
2020-11-12Fail fast while closing if reconnect is offViktor Söderqvist
If a request/headers/connect/ws_upgrade is created when a connection is in state 'closing', such as after receiving an HTTP/2 GOAWAY frame or an HTTP/1.1 "Connection: close" header, an error message is sent back to the caller immediately, if reconnect is off (that is if the option retry is set to 0). This allows an application to retry the request on another connection without waiting for all streams on the current connection to complete.
2020-11-12Change version to Gun 2.0.0-rc.1Loïc Hoguin
We are about one week before the actual release.
2020-11-12Update sponsorsLoïc Hoguin
2020-11-12Update copyright yearsLoïc Hoguin
2020-11-12Replace gun:ws_send/2 with gun:ws_send/3Loïc Hoguin
Switching from /2 to /3 should be easy enough. Also update the documentation about HTTP/2 Websocket support.
2020-11-12Review and remove many todosLoïc Hoguin
HTTP/1.1 Upgrade to HTTP/2 will not be implemented. There are discussions for this functionality to be removed from the HTTP/2 spec. HTTP/1.1 Upgrade to TLS will most likely not be implemented.
2020-11-12Remove an unused key from the internal tunnel_info()Loïc Hoguin
2020-11-12Add more new keys to gun:info and gun:stream_info manualsLoïc Hoguin
2020-11-12Use gun:stream_ref() where applicableLoïc Hoguin
2020-11-12Update gun_tunnel_up manualLoïc Hoguin
It was missing the StreamRef from the message.
2020-11-12Enable optional0042 cookie testLoïc Hoguin
Cowlib master has the fix for this test. Temporarily depend on Cowlib master until a release is made.
2020-11-11Initial internals chapter about TLS over TLSLoïc Hoguin
2020-11-09Update source path for test tool h2specdBjorn Svensson
The main source file for the test tool h2specd was moved a time ago, but correct one is used now. Also make sure that gun sends the GOAWAY directly to avoid timing issues in test 6.9.1 in h2specd_SUITE.
2020-11-09Update the migration guide and the gun(3) manualLoïc Hoguin
2020-11-09Add the gun_event manualLoïc Hoguin
2020-11-06Make CONNECT responses produce a response_end eventLoïc Hoguin
While the stream has not ended, the response has.
2020-11-06Update Cowlib to 2.10.0Loïc Hoguin
2020-11-06Fix IsFin mismatch in HTTP/2 CONNECT responseLoïc Hoguin
The response ends when the tunnel is established, even if the stream itself does not. The data coming in on the stream after is not part of the response. This makes both HTTP/1.1 and HTTP/2 send 'fin' to successful CONNECT responses.
2020-11-02Initial implementation of Websocket over HTTP/2http2-websocketLoïc Hoguin
2020-10-21Fix cookie handling when tunnel and origin schemes mismatchLoïc Hoguin
The cookie_ignore_informational has been moved to http_opts and http2_opts. Also fix an issue when using 'protocols' in gun:open. When connecting via TLS the protocol's options were discarded.
2020-10-20Disconnect when setopts returns an errorViktor Söderqvist
There is a race condition: After receiving data, before setting the socket back to {active, once}, the server may have closed the socket or an error may have occurred. Since the socket was in passive mode, no message is received. Handling the return value of setopts solves the issue. Amended by Loïc Hoguin to handle the return value from the active/1 function call.
2020-10-19Fix cookies for tunnelsLoïc Hoguin
There are still small issues left to fix. In particular the set_cookie command should be replaced with doing the same in the protocol itself so that the scheme is correct. So CookieStore must be propagated to all callbacks.
2020-10-16Remove commented codeLoïc Hoguin
2020-10-16Remove the transport_changed eventLoïc Hoguin
It doesn't provide any new information compared to the other events.
2020-10-16Add or fix events inside or related to CONNECT tunnelsLoïc Hoguin
2020-10-07Propagate timeouts to the right layer in HTTP/2 tunnelsLoïc Hoguin
This temporarily depends on Cowlib master.
2020-10-07Update Erlang.mkLoïc Hoguin
2020-10-07Default keepalive to infinity for all protocolsLoïc Hoguin
For HTTP/2 some servers are not fond of receiving pings, especially if there's no stream at the time. (particularly gRPC servers). This is an odd decision for sure. Rather than trying to accomodate for their logic the ping is no longer sent by default. For Websocket sending a ping is a new feature in Gun 2.0 so I'm just being conservative there.
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-04 03:39:38 +0000