Now Connection: close is ignored for 2xx responses to a CONNECT method.
In this way the tunnel is kept open, and bad proxies that were sending
Connection: close are now supported as apparently they are still out there.
Fixes also #6483.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#8770 - Review whether to send request body in redirects.
Now the original request body is re-sent only if the redirect status code is 307 or 308.
In the other cases, it is a redirect to a GET method, so the Location is followed without resending the body, and the content headers are removed.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
On Windows, Mac, OpenSolaris (verified) and probably Docker (did not verify, but likely),
DNS resolution of "localhost" yields [127.0.0.1, ::1], while on Linux only [127.0.0.1].
HttpClient will try the first address, and in case of failure will try the second address.
For those tests that want to explicitly fail the connection attempt, we don't want them to
try the second address, which likely yields test failures because of unexpected results
(for example, a different exception type).
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Removed "since" attribute (it's commonly related to Java versions) to avoid confusion.
* Avoid using getProxies() to modify the Proxy object in ProxyServletTest.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#7993 - HttpClient idleTimeout configuration being ignored/overridden
The problem was that the timeout scheduling was not happening,
because for TunnelRequest the timeouts were set in normalizeRequest(),
which runs after the scheduling.
Now a call to request.sent() is made also after normalizeRequest()
so that the timeouts is scheduled (if it was not scheduled before).
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Updating to version 10.0.12
* Updating to version 10.0.13-SNAPSHOT
* VERSION.TXT =+ content of jetty-9.4.49.v20220914 release
* remove strange line and reorder
* Fixes#8532 - Review System.nanoTime() usages.
Introduced o.e.j.util.NanoTime class to deal with nanoTimes.
Now NanoTime.now() should be used instead of System.nanoTime(),
and various <unit>[elapsed|since|until]() methods to calculate nanoTimes.
Furthermore, comparing 2 nanoTimes should be done via isBefore(),
rather than using the < operator, which is wrong as specified in
the System.nanoTime() javadocs.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
To avoid excess allocation, in particular for higher values of
`MaxRequestsQueuedPerDestination`, the per-destination request queue is
now initialized with a capacity of 32 entries and configured to grow 32
entries at a time until the maximum is reached.
Resolves#8319.
All `ByteBufferPool` can now be accessed as `RetainableByteBufferPools`.
Users now need to configure only a single buffer pool and there is just the additional retained parameter that needs consideration.
Default buffer pool has been changed to logarithmic, but we may wish to review that before next release.
Default factor size has been increased to 4096.
Fixes#8014 - Review HttpRequest URI construction.
Now always adding a "/" before the path, if not already present.
Disabled flakey HTTP/3 test.
Parse CONNECT URIs as Authority
Co-authored-by: Greg Wilkins <gregw@webtide.com>
Improved documentation for the clear-text dynamic transport case.
Fixed HttpClientTransportDynamic javadocs.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
* Fixes#7348 - Slow CONNECT request causes NPE (#7349)
Added NPE guard in `HttpReceiverOverHTTP.onUpgradeFrom()`.
Expanded logic in `HttpReceiverOverHTTP.parse()` to return true in case of CONNECT + 200.
Fixed `ProxyConnection.toConnectionString()` to avoid NPEs.
Fixed `HttpClientTest.testCONNECTWithHTTP10()` logic
after changes to fix this issue.
Now a tunneled connection is not put back into the connection pool,
and if applications explicitly want to use it, they must re-enable
fill interest, similarly to what should be done after upgrade+101.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
(cherry picked from commit 5eb7b70df7)
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
- Fixed max streams semantic for HTTP/3.
It's not *concurrent* streams but *cumulative* streams that are limited.
Fixed MultiplexConnectionPool to take that into account with the introduction
of ConnectionPool.MaxUsable.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
- More work on making HTTP semantic layer work on top of HTTP/3.
- Various fixes and improvement to HTTP client transport tests.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
- Made sure Connection.Listeners are properly notified.
- Fixed removal of QuicStreamEndPoints from QuicSession in case of successful request/response exchanges.
- Avoid spurious wakeups by shutting down input after receiving a frame with last=true.
- Updated HttpClient transport tests to work with UNIX_DOMAIN.
- Started updating HttpClient transport tests to work with HTTP/3.
Signed-off-by: Simone Bordet <simone.bordet@gmail.com>
Simone Bordet
Joakim Erdfelt
Olivier Lamy
gregw
dependabot[bot]
Gili Tzabari
Jan Vojt
Olivier Lamy
Ludovic Orban
Marc Philipp
Kaiyue Zhou