Introduced method Request.getConnection() to expose the Connection after at the request begin event.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Moved the call to destroy the CyclicTimeouts to a close() call that is always called.
Fixed NPE in ManagedSelector.getTotalKeys().
Fixed exception handling to avoid infinite recursion in SslConnection.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Bumped the rate control rate from 50 events/s to 128.
* Added rate control for all CONTINUATION frames.
* Added rate control for invalid PUSH_PROMISE frames.
* Added rate control for RST_STREAM frames.
* Added rate control for all SETTINGS frames.
* Fixed growth of header block accumulation buffer.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* improve testing of http2 client and refine default config
* improve testing in HTTP2Test for hpack
Signed-off-by: Lachlan Roberts <lachlan@webtide.com>
Now Content-Length and Content-Encoding are removed/modified by the decoder.
In this way, applications have a correct sets of headers to decide whether to decode the content themselves.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Now TunnelRequest.getURI() does not return null, so normalizeRequest() can properly apply the authentication headers.
Moved copy of a request to HttpRequest, so also the sub-type can be copied.
Fixed restore of destination in HttpProxy.HttpProxyClientConnectionFactory.newProxyConnection(): now doing it in the promise rather than in finally block.
Using the proxy destination (not the server's) to send subsequent CONNECT requests in case the first is not replied with 200.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Now allowing to specify a negative value for AbstractHTTP2ServerConnectionFactory.streamIdleTimeout, while 0 implies to use the default value (from the EndPoint).
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#8405 - onAllDataRead() is called twice under h2 if the stream times out
Per Servlet semantic, HTTP/2 stream timeout should be ignored.
The code was trying to fail the read via `_contentDemander.onTimeout()`, but
then it was still calling `onContentProducible()`, which was returning `true`
because the state of the read was IDLE (all the request content was read) and
the request was suspended.
Now the code checks if the read was really failed; if it is not, then
`onContentProducible()` is not called and so the idle timeout is ignored.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Method HTTP2StreamEndPoint.flush() has a "no pending operation" semantic, but the previous implementation was calling stream.data(), which may become a pending operation if the stream is congested.
Changed the implementation of flush() to return false in the IDLE and PENDING cases.
Now every flush() is converted to a write(), which has the same semantic as stream.data().
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
This is necessary because the remote encoder may encode fields assuming that the default size is 4096, and never send the table resize instruction.
In this case, the decoder table cannot start at zero, it must start at the default of 4096.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Implemented a few required error handlings.
* Changed `Parser.init()` to directly take the listener, rather than wrapping it.
The reason for this change was to be able to reconfigure the Parser upon receiving a SETTINGS frame.
* Initially setting the encoder and decoder max table capacity at the default of 4096, as per spec.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Regression introduced by #8678.
Now using awaitility to wait for the stream count to go to zero.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#8678 - Jetty client receives GO_AWAY and continue to send traffic on same connection
* Now upon receiving the GOAWAY, the connection is removed from the pool, so it cannot be used by new requests.
* HTTP2Session.removeStream() now happens _after_ notifying HEADERS and DATA events, although the Stream state change still happens before.
This is necessary to avoid that a "close" event is notified before a "headers" or "data" event.
With these changes, the race window of a client acquiring a connection while the server is closing it is reduced, but it is impossible to close it completely.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Don't care about stop failures in this test
* More consistent ConnectTimeoutTest
+ using IANA example.com (as it will route to a
real machine, and will never have this
specific port open)
Signed-off-by: Joakim Erdfelt <joakim.erdfelt@gmail.com>
Now a completed future is returned from shutdown() if the session is already closed.
Moved the notification of the CompletableFuture to terminate(), which is always invoked.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Joakim Erdfelt
Simone Bordet
Ludovic Orban
Lachlan
Jan Peter Stotz
Olivier Lamy
gregw
dependabot[bot]