andreacavalli/netty5
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
2,204 Commits 27 Branches 234 Tags 84 MiB
a91c56c84d
Commit Graph

2003 Commits

Author SHA1 Message Date
Norman Maurer
a91c56c84d Remove iostream support from 3.2 branch. This will only go into 4.x 2011-11-03 19:29:22 +01:00
Norman Maurer
24e0b01e8d Release the cumulation buffer after firing upstream so we don't end up 
with a "leak" because of a very big ChannelBuffer. This patch is
extracted of the pull request 39. Thanks to arya for the patch. See
NETTY-453
 2011-11-03 11:29:39 +01:00
Norman Maurer
d742a11c25 Merge pull request #22 from pfisterer/master. Which adds a 
IOStreamChannelFactory which can be used to connect to InputStream and
OutputStream
 2011-10-29 20:18:10 +02:00
norman
b1c27063a2 Make sure the fireExceptionCaught will not get fired while holding a 
lock. This will prevent a deadlock which you may see if you have an
ChannelHandler that will trigger Channel.close() on an Exception. See
NETTY-443
 2011-10-25 11:52:27 +02:00
norman
6728766158 ChannelLocal nows removes Channel instances from itself once the Channel 
was closed. This is configurable via a constructor parameter to allow
the user to choose what to do. The default is to not remove the Channel
from the ChannelLocal to not break usage which depends on the "old"
behavior. See NETTY-447
 2011-10-24 08:32:17 +02:00
Trustin Lee
e3386736c0 Fix NETTY-432 HttpContentEncoder should not encode if Content-Encoding is set already and it is not 'identity' 2011-10-22 22:08:49 -07:00
Trustin Lee
a67cdaeb48 Clean up FileRegion and its implementation changes 
* Fix broken javadoc tags
* Remove unnecessary public modifier
* Reorder method
* Make releaseAfterTransfer immutable
 2011-10-22 21:47:07 -07:00
Norman Maurer
2a4b91b59b Only release FileRegion after transfer was done if its configured todo 
so. The default is false to be sure we are compatible with earlier 3.2
releases. This should be changed to default true in the master branch.
See NETTY-440
 2011-10-22 14:56:53 +02:00
Trustin Lee
4b9b421af2 NETTY-439 Call channelBound/Connected from worker 
* Move channelBound() to RegisterTask
* Simplify the if block in RegisterTask
 2011-10-21 15:42:02 -07:00
Trustin Lee
beadadecf7 Fix NETTY-445 Memory leak in SslHandler 
* Ensure SSLEngine.closeInbound() and closeOutbound() is called on
handshake failure.
* Ensure connection is closed when a closure request is issued and
handshake failed.
 2011-10-21 13:54:07 -07:00
norman
48c6793d30 Revert invalid fix 2011-10-17 15:56:38 +02:00
norman
4be22a42ef ThreadNameDeterminer.CURRENT return the right name now. Was returning 
null before all the time
 2011-10-17 15:49:29 +02:00
norman
35401caac4 Remove unused imports and fix warnings 2011-10-13 12:08:06 +02:00
norman
fd1d11ec39 Add support for FileRegion in OIOWorker. 2011-10-12 17:09:02 +02:00
norman
2d25998eb4 Make sure FileRegion.releaseExternalResources() is called after the 
write was done. See NETTY-440
 2011-10-12 15:07:53 +02:00
norman
fb408778d1 Only handle the channelConnected in the worker thread. The channelOpen 
and channelBound was moved back to the boss thread. This was done after
reading NETTY-154
 2011-10-12 13:15:06 +02:00
norman
cb2e047f1d Remove code which I committed by mistake 2011-10-12 13:04:37 +02:00
norman
bbdc2032f0 Make sure the ChannelOpen, ChannelBound and ChannelConnected events get 
fired from within an IO-Worker Thread. This makes sure the Boss-Thread
will not get blocked by any user action
 2011-10-12 13:02:50 +02:00
norman
28120aa778 Add javadoc entry to explain that some events are executed in the boss 
thread.
 2011-10-12 10:35:12 +02:00
Trustin Lee
ba72bf7235 NETTY-438 Re-enable SSL re-negotiation by default 
* Re-enabled renegotiation
* Updated Javadoc
 2011-10-10 20:47:32 +09:00
Norman Maurer
b89487c445 Add patch of NETTY-434 to the mix to fix Deadlock in 
ChunkedWriteHandler. This patch was not written by me...
 2011-10-10 18:00:32 +09:00
Evan Meagher
256621246c Fix wording in "Summary" section. 2011-10-10 17:42:15 +09:00
Evan Meagher
f0b30aa8e1 Fix wording in "Advanced Components for More Rapid Development" section. 2011-10-10 17:42:07 +09:00
Evan Meagher
95b127748e Fix wording in "Interceptor Chain Pattern" section. 2011-10-10 17:42:00 +09:00
Evan Meagher
56d6217b53 Fix wording in "Universal Asynchronous I/O API" section. 2011-10-10 17:41:52 +09:00
Evan Meagher
d1a15c9aff Fix wording in "Rich Buffer Data Structure" section. 2011-10-10 17:41:45 +09:00
Trustin Lee
a78d37460c Fixed a compilation error in the example 2011-08-26 13:24:17 +09:00
Trustin Lee
b0d0502413 Added a TooLongFrameException recovery test for DelimiterBasedFrameDecoder 2011-08-19 11:11:16 +09:00
Trustin Lee
3effb0d1b0 NETTY-435 LengthFieldBasedFrameDecoder fails to recover from TooLongFrameException 
* Fixed a bug where TooLongFrameException is not raised immediately when the large frame was fully decoded at the first attempt
* Fixed a bug where LengthFieldBasedFrameDecoder does not reset its state completely after raising TooLongFrameException
 2011-08-19 11:05:24 +09:00
Trustin Lee
58cc6aec86 NETTY-431 HashedWheelTimer's TimerTask may execute after call to Timeout.cancel() 
* Replaced a volatile boolean flag and system date access with an atomic integer flag.
 2011-08-12 14:03:48 +09:00
Trustin Lee
1345a00a1a NETTY-430 ChunkedWriteHandler NPE 2011-08-12 13:39:18 +09:00
Trustin Lee
7eb39aaf93 NETTY-429 IllegalArgumentException when an HTTP server sends a '100 Continue' response to HttpContentDecoder. 
* Fixed a silly coding mistake where I sent the event in the opposite direction
 2011-08-03 21:26:31 +09:00
Trustin Lee
b969706832 NETTY-428 SslHandler does not trigger an exceptionCaught event for some handshake failure cases. 
* Made sure SslHandler.handshake() and SslHandler.close() trigger an exceptionCaught event when failed
 2011-08-02 15:39:25 +09:00
Trustin Lee
5f19f4c0b6 SSLEngine.beginHandshake() can throw a RuntimeException. 2011-08-02 15:09:35 +09:00
Trustin Lee
dc33e99872 NETTY-408 ChunkedStream generates 1-byte chunks for wrapped input streams that does have sane implementation of available() method 
* Updated JavaDoc so that users don't specify an InputStream with poor available() implementation.
 2011-08-02 09:48:34 +09:00
Trustin Lee
e0eb72618a NETTY-384 Another deadlock in ChunkedWriteHandler 
ChunkedWriteHandler.discard() do not issue write requests to trigger exceptionCaught events and to notify write futures anymore.  Instead, it triggers exceptionCaught events and notifies write futures by itself.  Therefore, no write lock is involved during discard(), avoiding the reported dead lock.

However, this is a temporary solution, and eventually Netty must introduce more robust event thread model.
 2011-08-02 09:33:53 +09:00
Trustin Lee
654da6d02a NETTY-422 Add getPipeline() to CodecEmbedder so that a user can modify the pipeline dynamically. 2011-08-02 08:49:13 +09:00
Trustin Lee
8f30a11c67 NETTY-423 Add ChannelPipeline.getNames() 2011-08-02 08:44:15 +09:00
Trustin Lee
4321135452 NETTY-420 HttpClientCodec doesn't handle HEAD requests with Transfer-Encoding: chunked responses 
* Reverted back the workaround for servers that sends unnecessary zero-byte chunk in a HEAD response because it actually breaks the codec when the server that conforms to the RFC.
 2011-08-02 08:35:12 +09:00
Trustin Lee
5cdcc67086 NETTY-419 ObjectEncoder/ObjectDecoder fails with NPE when deserializing the java.lang.Class instance of an interface type 
* Fixed a bug where deserializing an interface class fails due to wrong lookup.
* Implemented class lookup caching
 2011-08-02 08:25:26 +09:00
Trustin Lee
99daeebe4a NETTY-417 client channel still open after close and wait 
* Fixed a bug in SslHandler where a write attempt made after SSLEngine is closed result in unnotified future.  Fixed by respecting the case where SslEngineResult.Status == CLOSED
 2011-08-02 07:59:14 +09:00
Trustin Lee
07e9378423 NETTY-417 client channel still open after close and wait 
* Fixed a race condition where NioSocketChannel's state variable is updated *after* its close future is notified
* Removed unnecessary use of ChannelFutureListeners in NioSocketChannel and AbstractChannel

Conflicts:

	src/main/java/org/jboss/netty/channel/AbstractChannel.java
	src/main/java/org/jboss/netty/channel/socket/nio/NioSocketChannel.java
 2011-08-02 07:58:16 +09:00
Trustin Lee
caf8b39172 Fixed a bug where the future associated with the close request is not notified when the channel was closed already when SslHandler attempts to send a closure_notify. 
This shouldn't be a problem to most applications because they usually do not specify their own future object for the close request.
 2011-08-02 07:49:45 +09:00
Felix Trepanier
7e50bd5160 do not encode if empty content 2011-08-02 06:48:01 +09:00
Trustin Lee
483f093036 NETTY-415 ChannelFuture.setFailure() not called when exception thrown handling Channel.close() 
* Made sure the ChannelFuture associated with a downstream event is marked as failure when an exception is raised before it reaches at ChannelSink.
 2011-08-02 06:37:00 +09:00
Trustin Lee
a804c3495e NETTY-418 isCompleteFailure() and isCompleteSuccess() are both true at the same time 
* Fixed incorrect isPartialSuccess(), isPartialFailure(), isCompleteFailure() implementation
 2011-08-01 06:12:01 +09:00
Trustin Lee
2d35bf2155 NETTY-425 ChannelBuffers.compare does not handle unsigned bytes correctly 
* Fixed a bug where signed comparison is made where unsigned comparison is expected
 2011-08-01 04:28:29 +09:00
Trustin Lee
fba8c7b7a4 NETTY-426 Prevent a user from reusing an upstream MessageEvent to write 
something

Modified the pipeline implementations so that it rejects the attempt to
send an UpstreamMessageEvent to downstream
 2011-08-01 04:17:39 +09:00
Trustin Lee
bf41f4b099 NETTY-410 NioWorker write deadlock 
Applied the patch provided by Greg Dhuse.
 2011-08-01 04:00:26 +09:00
Trustin Lee
d1b40a3c8b NETTY-407 HttpTunnelingClientSocketChannel failed close() is not sent to 
calling futures.

* HttpTunnelingClientSocketChannel.writeLastChunk should return a failed
future instead of raising an exception.
 2011-08-01 03:54:47 +09:00