vibul
6009a413b9
Issue #283 - Support max frame length for web socket to limit chance of DOS attack
2012-04-26 10:09:12 +10:00
norman
a8b9e27c92
NioDatagramWorker.ChannelRegistionTask should handle ClosedChannelException gracefully. See #281 and #277
2012-04-25 09:24:51 +02:00
Norman Maurer
476cf97b97
Make the cumulation usage more memory efficient. See #280
2012-04-24 20:41:10 +02:00
Norman Maurer
6e2e9fb3c5
Use gathering writes if java version is >= 7 . See #269 and #271
2012-04-24 20:32:06 +02:00
Norman Maurer
77d2f9c4ef
Upgrade and Connection header must be matched in a case-insensitive manner in WebSocket 08 and 13. See #278
2012-04-22 12:56:37 +02:00
Norman Maurer
980d96cf58
Fix possible NPE. See #274
2012-04-21 16:12:08 +02:00
Norman Maurer
5f6b419bb9
Allow to specify the local address when connect. See #276
2012-04-21 16:00:23 +02:00
Norman Maurer
ccf01d133a
Make sure we always cleanup once ReplayingDecoder handles a message. See #259
2012-04-20 20:17:30 +02:00
Norman Maurer
ec409751e1
Correctly handle the stripping of the zoneId / scopeId in all cases. See #267
2012-04-19 16:42:55 +02:00
Norman Maurer
2ecef07c4a
Allow to disable the use of sun.misc.Unsafe via a System property. See #272
2012-04-19 12:07:17 +02:00
Norman Maurer
e719f23f7d
Only log if logging level is enabled
2012-04-19 12:03:09 +02:00
Norman Maurer
54559a9595
Make it configurable if the HttpClientCodec should throw an exception on close when the response and request count does not match. Default is false. See #266
2012-04-17 20:54:58 +02:00
Norman Maurer
d808cd0475
Fix compile error
2012-04-17 20:54:38 +02:00
Norman Maurer
b3b5fb1de6
Fallback to LegacyLinkedTransferQueue if using LinkedTransferQueue fails. See #268
2012-04-17 14:17:56 +02:00
Norman Maurer
02dc9ea8c1
Add workaround for connection problems with IPv6 link-local addresses and jdk < 7. See #267
2012-04-17 14:16:37 +02:00
Norman Maurer
2db3e59a6c
Add workaround for connection problems with IPv6 link-local addresses and jdk < 7. See #267
2012-04-17 09:56:15 +02:00
Norman Maurer
aa6c16b136
Make sure PrematureChannelClosureException is not thrown incorrectly sometimes. See #266
2012-04-17 09:25:53 +02:00
Norman Maurer
cd92478bfc
Use Sec-WebSocket-Origin in WebSocketClientHandshaker08 as replacement for Origin. See #264
2012-04-16 15:30:43 +02:00
norman
00e59da7ba
Add port to Origin HTTP Header if the port is non default (80/443). See
...
#262
2012-04-16 13:26:35 +02:00
norman
0a84ce6f00
Merge branch 'master' of ssh://github.com/netty/netty
2012-04-16 13:12:36 +02:00
norman
987e152f08
Add port to Origin HTTP Header if the port is non default (80/443). See
...
#262
2012-04-16 13:12:12 +02:00
Norman Maurer
978168e8c6
Fix typo
2012-04-16 11:45:59 +03:00
norman
c44f365ee7
Add some javadocs notes that explain the behavior of CookieEncoder.encode(). See #94
2012-04-16 10:43:21 +02:00
norman
05615c4779
Throw IllegalStateException if DynamicChannelBuffer exceed the maximum
...
ChannelBuffer size of 2gb. See #258
2012-04-16 09:55:03 +02:00
Norman Maurer
f25453cd36
Merge pull request #261 from CruzBishop/logger-usage
...
Replace System.out logging with InternalLogger
2012-04-15 23:06:14 -07:00
Norman Maurer
e02225a80f
Add also tests for HttpClientCodec that tests handling chunked responses. See #256 amd #259
2012-04-15 20:58:40 +02:00
Norman Maurer
bc141f695b
Add test for HttpClientCodec that tests handling of missing responses. See #256 amd #259
2012-04-15 15:47:05 +02:00
Cruz Bishop
11e974ace3
Fix up the last logging mishits
2012-04-15 19:22:02 +10:00
Cruz Bishop
a682b018b2
More logging
2012-04-15 19:18:35 +10:00
Norman Maurer
fa3c9c8eee
Merge pull request #260 from CruzBishop/logger-usage
...
Replace System.*.println calls with logging
2012-04-15 01:22:26 -07:00
Cruz Bishop
a20ab9184e
Use loggers in some more classes
2012-04-15 17:49:04 +10:00
Cruz Bishop
efabc3c285
Use a logger in ZStream
2012-04-15 17:34:51 +10:00
Cruz Bishop
db97e4eb35
Use a logger in SocketAddresses
2012-04-15 17:29:15 +10:00
norman
d363f73fd8
Only increment the counter if the encode did not fail. See #256
2012-04-12 10:55:18 +02:00
Trustin Lee
c7bd0b41e6
Remove unnecessary parameter
2012-04-12 17:54:56 +09:00
Trustin Lee
22a815eaf8
Revamp channel handler API
...
- Merged LifeCycleAwareChannelHandler into ChannelHandler
- Replaced ChannelUpstreamHandler and ChannelDownstreamHandler with
ChannelReader and ChannelWriter
- These two new interfaces are much more type-safe than its ancestor.
- Simplified channel state model as described in #68
- Handler creates send/receive buffer.
- Previously, Netty created them, but it led to more memory copies and
inflexibility. I'm going to allow a handler to create a bounded
queue for example.
- It currently uses Queue<T> but I'll define a new interface and make
ChannelBuffer implement it (e.g. Queue<Byte>)
- Introduced AttributeMap which replaces attachments in Channel and
ChannelHandlerContext and ChannelLocal
2012-04-12 17:39:01 +09:00
norman
5ed04c3ada
Correctly handle chunked requests/responses. See #256
2012-04-12 10:34:16 +02:00
norman
b9c60bd518
Throw a PrematureChannelClosureException if the channel was closed before all responses were received for the sent requests. See #256
2012-04-12 10:22:10 +02:00
norman
941e71de36
cleanup comments
2012-04-12 09:29:59 +02:00
norman
05204025cc
Fix regression in Zlib which was introduced while refactoring the code. See #255
2012-04-11 10:55:04 +02:00
norman
97561315fd
Add final keyword
2012-04-11 10:54:53 +02:00
norman
16c625cfd0
Allow to share a WorkerPool for boss and worker threads but also allow to have them separate. See #240
2012-04-11 10:54:41 +02:00
norman
f88cd3120d
Fix a bug which lead to only use two threads for all tasks all the time, even if the WorkerPool contained more. See #240
2012-04-11 10:54:27 +02:00
norman
4c0a5886ca
Make sure multicast tests pass on all os'es
2012-04-11 10:54:14 +02:00
norman
3f8c13f138
Fix regression in Zlib which was introduced while refactoring the code. See #255
2012-04-11 10:26:29 +02:00
norman
962a67dc03
Add final keyword
2012-04-11 09:16:27 +02:00
norman
5b53b66fbf
Allow to share a WorkerPool for boss and worker threads but also allow to have them separate. See #240
2012-04-11 09:15:02 +02:00
norman
470c1a898a
Fix a bug which lead to only use two threads for all tasks all the time, even if the WorkerPool contained more. See #240
2012-04-11 08:45:51 +02:00
norman
a37d7bb5f3
Make sure multicast tests pass on all os'es
2012-04-11 08:09:03 +02:00
Norman Maurer
4436f60e2f
Workaround to have multicast tests work on osx, linux and windows
2012-04-10 21:00:44 +02:00