Commit Graph

2901 Commits

Author SHA1 Message Date
Greg Wilkins cfe5abf8e6 Issue #660 NPE in Request
Made request more robust for calls after reset.
2016-06-24 14:33:18 +10:00
Greg Wilkins c3173932f0 Revert "Issue #660 - simplifying parameter extraction to prevent NPE"
This reverts commit ecea486d07.
2016-06-23 12:43:29 +10:00
Jesse McConnell ae931538ab [maven-release-plugin] prepare for next development iteration 2016-06-22 13:59:10 -05:00
Jesse McConnell df1ed4fb03 [maven-release-plugin] prepare release jetty-9.3.11.M0 2016-06-22 13:59:06 -05:00
Jesse McConnell 2b947e8f72 Merge branch 'jetty-9.3.x' into release-9.3.10 2016-06-22 13:22:27 -05:00
Joakim Erdfelt ecea486d07 Issue #660 - simplifying parameter extraction to prevent NPE
+ The checks in Request.restoreParameters() seem superfluous
  as MultiMap.addAllValues() is null and empty set safe
2016-06-22 09:53:24 -07:00
Joakim Erdfelt cce7837e64 Fixes #654 - ServletContext.getResourceAsStream("/") should return null
+ Jetty 9.2 (and earlier) used a Resource.getInputStream() implementation
   that would trigger an IOException on-construction (vs on-first-access)
   due its use of java.io.FileInputStream(File) which had a isDirectory
   check in its constructor.
 + Jetty 9.3 and onward uses java.nio.file.Files.newInputStream() which
   has the bad on-first-access behavior.
 + Changing the behavior of PathResource.getInputStream() to behave the
   same way as the prior FileResource.getInputStream(), as well as adding
   the Resource.isDirectory() check back into ContextHandler.getResourceAsStream(String)
   to prevent creation of the InputStream in the first place.
2016-06-22 09:09:05 -07:00
Jesse McConnell 030ec61837 [maven-release-plugin] prepare for next development iteration 2016-06-22 09:14:25 -05:00
Jesse McConnell 9a4cb41122 [maven-release-plugin] prepare release jetty-9.3.11.M0 2016-06-22 09:14:21 -05:00
Jesse McConnell f2264c96a6 Merge branch 'jetty-9.3.x' into release-9.3.10 2016-06-22 08:40:46 -05:00
Greg Wilkins d4c1bec340 Fixes #624 cleanup 2016-06-22 18:58:31 +10:00
Greg Wilkins 126e58c7cc Fixes #624 cleanup 2016-06-22 16:47:00 +10:00
Greg Wilkins 59540f552c Fixes #624 2016-06-22 16:31:25 +10:00
WalkerWatch baae86fcef Correcting path for #645 (#646)
* Correcting path for #645

Signed-off-by: WalkerWatch <ctwalker@gmail.com>

* Fixing path variables.

Signed-off-by: WalkerWatch <ctwalker@gmail.com>
2016-06-22 09:44:59 +10:00
Jesse McConnell e7034117ea [maven-release-plugin] prepare for next development iteration 2016-06-21 08:50:13 -05:00
Jesse McConnell 16c435c69d [maven-release-plugin] prepare release jetty-9.3.10.v20160621 2016-06-21 08:50:09 -05:00
Greg Wilkins a8e315a3f5 Fix #663 NPE during context stop
Avoid adding null beans
protect against null beans.
2016-06-15 18:04:33 +10:00
Greg Wilkins f724d36646 Cleanup #624
removed excess test code
2016-06-15 10:38:40 +10:00
Greg Wilkins bc87cf0ab4 Revert "Debug #624"
This reverts commit b40c2ae66f.
2016-06-15 10:32:10 +10:00
Simone Bordet e803bf6c81 Test code cleanup.
Converted static fields to non-static, and removed unused code.
2016-06-13 12:15:45 +02:00
Greg Wilkins b40c2ae66f Debug #624
This commit is instrumentation to be executed in the CI system to debug #624
This commit should be reverted once the test has failed again!
2016-06-11 10:33:48 +10:00
Greg Wilkins 2609c89d22 Issue #624 Fix onCompleted race 2016-06-08 16:22:53 +10:00
Greg Wilkins f4c13e5f54 Issue #623 Add gzip suffix to etags in 304 response 2016-06-08 14:33:49 +10:00
Greg Wilkins 40af755251 restructure failing test to better analyse failure 2016-06-07 10:44:33 +10:00
Greg Wilkins a2309057dc Test Harness for #596
Test harness to try to repeat problem with #596 of content-length added to a HEAD response.

In the process added a much better getResponse mechanism to the local connector that avoids
using the idle time.
2016-06-03 14:15:25 +10:00
Greg Wilkins 89032b7eac improved test isolation 2016-06-02 14:24:18 +10:00
Greg Wilkins c2831bf09c Issue #608 reset encoding set from content type
Use an enum to track where a content encoding came from and selectively clear/reset
2016-06-01 15:29:05 +10:00
Simone Bordet 031bc0fed9 Issue #85 - Expose TLS protocol used for connection in SecureRequestCustomizer.
Exposed the SSLSession as attribute only if the attribute key is valid.
2016-05-25 00:50:01 +02:00
Simone Bordet c6aafecceb Code cleanups. 2016-05-25 00:47:58 +02:00
Simone Bordet ddddc3a675 Fixes #574 - Introduce a TLS handshake completed listener.
Introduced class SslHandshakeListener that can be registered as a
bean in both the ServerConnector and in clients such as HttpClient
and HTTP2Client.

When creating SslConnection instances, the factory will query the
connector (client or server) for SslHandshakeListener beans and, if
present, will be added to the SslConnection.
2016-05-24 16:09:19 +02:00
Simone Bordet d73c60db14 Issue #464 - Improve reporting of SSLHandshakeException. 2016-05-24 11:15:18 +02:00
Simone Bordet f95daca8b4 Code cleanups. 2016-05-24 11:15:18 +02:00
Greg Wilkins d73bf08515 cleaned up constructors 2016-05-20 15:53:46 +10:00
Simone Bordet faec7259b2 Fixed dispatch to application in case of empty chunked body. 2016-05-20 01:12:24 +02:00
Simone Bordet aed6bb6d6a Code cleanups. 2016-05-20 01:12:24 +02:00
Jesse McConnell 4bf91b5ab3 Merge remote-tracking branch 'origin/release-9.3.9' into jetty-9.3.x 2016-05-19 09:55:58 -05:00
Simone Bordet ff1bfa599f Issue #354 (Spin loop in case of exception thrown during accept()).
Improved the solution by sleeping only if we are still accepting.
This avoids the sleep when the connector is stopping, speeding up tests.
2016-05-19 11:48:51 +02:00
Greg Wilkins 79e5c31029 improved unit test, fixed deferred on chunking 2016-05-19 17:46:50 +10:00
Greg Wilkins cdd73eb4da improved javadoc 2016-05-19 11:27:09 +10:00
Jesse McConnell dc4574f380 [maven-release-plugin] prepare for next development iteration 2016-05-17 17:30:17 -05:00
Jesse McConnell 0bbebec1e0 [maven-release-plugin] prepare release jetty-9.3.9.v20160517 2016-05-17 17:30:13 -05:00
Simone Bordet 55e5f74889 Fixes #558 - HTTP/2 server hangs when thread pool is low on threads.
Fixed test that was broken after changes for #557.
2016-05-16 15:40:35 +02:00
Joakim Erdfelt 97af3d663f Fixes #556 - Improve Resource.getAlias() checks on Windows
+ Reworks PathResource and FileResource alias checking to
  use originally passed paths, before Windows JVM has a
  chance to normalize and "correct" the bad paths.
2016-05-13 11:57:23 -07:00
Greg Wilkins 298a340851 Issue #558
The testEPCExitsLowThreadsMode has been broken, so ignored for now
2016-05-12 13:26:14 +10:00
Simone Bordet 9b6d42317b Improved logging. 2016-05-11 13:05:16 +02:00
Greg Wilkins bdafea3e96 Issue #561 Fixed test timer 2016-05-11 09:24:02 +10:00
Simone Bordet caa45283c7 Fixes #547 (ExecuteProduceConsume (EWYK) does not exit low threads mode)
Exposed getters/setters for ExecutionStrategy.Factory in
ServerConnector and SelectorManager, to allow explicit configuration
and testing.

Added test to verify that EPC exits low threads mode.
2016-05-06 17:10:46 +02:00
Simone Bordet 8bb6d4d7dc Code cleanups. 2016-05-06 17:10:46 +02:00
Greg Wilkins 2a5b120fc3 Issue #525 fix blockForContent spin
abort request in bad request handling
2016-05-06 20:31:55 +10:00
Greg Wilkins 6591137ba0 Issue #547 EWYK leaves low resource mode
Directly implement PEC in EPC low resource mode
2016-05-05 11:22:37 +10:00