Commit Graph

1695 Commits

Author SHA1 Message Date
Trustin Lee
cbe26db0e4 Replaced email with homepage url 2010-06-14 11:07:44 +00:00
Trustin Lee
2571564eb2 * Updated the author tag
* Added svn:keywords property
2010-06-14 11:06:45 +00:00
Trustin Lee
6d0fb256c6 Fixed issue: NETTY-325 (QueryStringDecoder doesn't properly handle missing query string values and other corner cases)
* Rewrote QueryStringDecoder based on Benoit's work
* Added a test case for QueryStringDecoder
2010-06-14 10:57:48 +00:00
Trustin Lee
5097009f8c * Fixed an infiniloop in NioDatagramWorker which is triggered when a user writes an empty buffer.
* Reduced system calls on empty buffer
2010-06-07 02:11:04 +00:00
Trustin Lee
c032708ed7 Updated Javadoc 2010-06-07 01:50:02 +00:00
Trustin Lee
659eedb970 Better package description 2010-06-03 09:04:56 +00:00
Trustin Lee
4a390a1983 More Javadoc 2010-06-03 09:01:46 +00:00
Trustin Lee
3a92fb8a9e Fixed NETTY-320 ReplayingDecoderBuffer.toByteBuffer(int, int) throws java.lang.IndexOutOfBoundsException
* Added missing checkIndex() call
2010-06-01 09:19:19 +00:00
Trustin Lee
a55a1c6e72 Removed unnecessary lock 2010-06-01 08:38:51 +00:00
Trustin Lee
a2ef43cf3d Streamline LinkedTransferQueue.isEmpty() 2010-05-28 02:59:17 +00:00
Trustin Lee
11705c9915 Removed System.out 2010-05-27 12:40:50 +00:00
Trustin Lee
0d0b0c56c3 Fixed NETTY-319 HttpMessageDecoder incorrect parsing status line
* Specified an empty reason phrase if no reason phrase was specified
2010-05-27 12:38:54 +00:00
Trustin Lee
7d11a5369f Proper validation for HttpResponseStatus 2010-05-27 12:34:44 +00:00
Trustin Lee
d65cd371e7 Fixed a bug where HttpTunnelingClientSocketChannel's closeFuture is not notified if the connection is closed by the client 2010-05-27 12:02:49 +00:00
Trustin Lee
be544fa899 Added a TODO for 4.0 (or 3.3?) 2010-05-19 12:23:44 +00:00
Trustin Lee
c64e6866f6 Added a FIXME for later releases 2010-05-19 11:31:25 +00:00
Trustin Lee
d51ccd2211 The fix for RtspResponseDecoder should be applied to all RTSP decoders 2010-05-19 07:51:38 +00:00
Trustin Lee
0c9202ab75 Fixed a bug where RtspResponseDecoder does not decode a response with no 'Content-Length' header properly 2010-05-19 07:42:21 +00:00
Trustin Lee
3847fb00e5 Fixed issue: NETTY-310 Memory leak in netty-3.2.0.BETA1.jar
* Added a special internal use only constructor to AbstractChannel, which bypasses ID allocation
* EmbeddedChannel calls the special purpose constructor to avoid leak
2010-05-19 06:29:43 +00:00
Trustin Lee
93f9c4a2d6 * Added LengthFieldBasedFrameDecoder.extractFrame()
* ObjectDecoder overrides extractFrame() to avoid memory copy
2010-05-13 14:13:07 +00:00
Trustin Lee
d19aa4924e Simplified and hardened ObjectDecoder by extending LengthFieldBasedFrameDecoder 2010-05-13 13:40:36 +00:00
Trustin Lee
bdd213ec85 * Added missing Javadoc
* Proper getIntHeader() behavior
2010-05-06 07:50:19 +00:00
Trustin Lee
2c887daabe Added missing Javadoc 2010-05-06 07:37:27 +00:00
Trustin Lee
8bb9ad26ad Updated Javadoc 2010-05-06 07:33:26 +00:00
Trustin Lee
5edc4e16aa * StaticChannelPipeline's constructor should stop when a null is encountered
* Added a test case for StaticChannelPipeline
2010-05-06 07:00:52 +00:00
Trustin Lee
56cdf73cb9 Fixed issue: NETTY-314 ProtobufEncoder and ProtobufDecoder should accept MessageLite as well as Message. 2010-05-06 04:11:11 +00:00
Trustin Lee
d4049710bb * Fixed dangling HTTP header entries in the linked list
* Added a test case
2010-05-04 01:46:15 +00:00
Trustin Lee
b2f550cb3b Fixed a bug where RtspResponseDecoder generates a request object instead of a response object 2010-05-03 02:21:01 +00:00
Trustin Lee
5741721f71 Fixed a bug where HEAD response is not read completely if it's transfer encoding is 'chunked'. 2010-04-16 13:32:55 +00:00
Trustin Lee
2ac79dcb74 A little optimization 2010-04-16 06:53:12 +00:00
Trustin Lee
f1bba73b22 Typo in the comment 2010-04-16 06:44:48 +00:00
Trustin Lee
d38219a85c Fixed a bug where non-successful HEAD response is assumed to have a message body 2010-04-16 06:40:29 +00:00
Trustin Lee
3aa3b07a95 Hint on downcasting and multicasting 2010-04-16 05:07:37 +00:00
Trustin Lee
66882ae708 * Updated APIviz tags
* More help on file transfer
2010-04-16 05:01:55 +00:00
Trustin Lee
ba08d11356 * Code clean-up
* More strict sanity check
2010-04-16 04:22:00 +00:00
Trustin Lee
48696d909f Updated APIviz tags 2010-04-16 04:12:43 +00:00
Trustin Lee
4e69cfe281 Exception should be raised even if count is 0 2010-04-12 11:57:46 +00:00
Trustin Lee
318a4d901f Fixed a silly bug in file transfer - There's no way to specify the start offset in FileRegion.transferTo() 2010-04-12 11:56:52 +00:00
Trustin Lee
17a222e397 * Changed the sort order of LocalAddress.compareTo()
* Added a FIXME to LocalAddress.compareTo()
* Modified LocalAddress.compareTo() to throw an Error when impacted by the known bug in compareTo() - will fix this bug in the next major feature release since it will break the backward compatibility
2010-04-12 10:53:49 +00:00
Trustin Lee
2fc539a9bc Proper LocalAddress.compareTo() implementation 2010-04-12 10:36:04 +00:00
Trustin Lee
d0d2519871 Fixed issue: NETTY-308 ChunkedInput implementations send only the first chunk 2010-04-12 10:22:51 +00:00
Trustin Lee
cac77a47b2 Fixed a wrong usage of ExecutionHandler in the example 2010-04-06 09:40:35 +00:00
Trustin Lee
8474026c92 * Misuse warnings are logged at WARN level now (with higher threshold)
* Extracted duplicate code into SharedResourceMisuseDetector
2010-04-06 09:23:25 +00:00
Trustin Lee
11e5e2ba56 Fixed issue: NETTY-306 bug in passing large nonchunked data through ChunketWriteHandler 2010-04-06 08:43:49 +00:00
Trustin Lee
1ae76562b7 Simplified container integration by removing unused features 2010-03-31 03:20:47 +00:00
Trustin Lee
3f7fff3fa5 HttpChunk is mutable so no need to create a new chunk 2010-03-31 03:02:58 +00:00
Trustin Lee
8eb353eb5d HttpChunk is mutable so no need to create a new chunk 2010-03-31 03:02:20 +00:00
Trustin Lee
b210c36ef9 Do not remove 'Content-Length' header even for a message with empty content - HEAD requires that 2010-03-31 02:50:19 +00:00
Trustin Lee
745b8300f0 Do not decode when the content is empty 2010-03-31 02:48:57 +00:00
Trustin Lee
c1a70b3579 HTTPS support 2010-03-31 02:26:51 +00:00
Trustin Lee
7c139cf1b7 * Added HttpClientCodec and HttpServerCodec
* Additional Javadoc about handling HEAD and CONNECT requests and their responses
2010-03-31 02:19:45 +00:00
Trustin Lee
c1b1db276d More useful Javadoc 2010-03-30 08:02:32 +00:00
Trustin Lee
4a4f129381 Better documentation 2010-03-24 05:07:27 +00:00
Trustin Lee
807291a2db Removed outdated XNIO transport - it needs to be rewritten in the next major feature release. 2010-03-24 04:00:49 +00:00
Trustin Lee
5d36109464 Fixed Javadoc errors 2010-03-16 09:39:56 +00:00
Trustin Lee
fa6f7a4e6d Fixed a bug where the pooled send buffer is not released on a failed write 2010-03-08 04:17:28 +00:00
Trustin Lee
605d2969d2 Added @Deprecated annotation to the implementations of deprecated methods in ChannelBuffer (basically to silence nagging compiler) 2010-03-04 06:34:00 +00:00
Trustin Lee
1695e32b0f Removed unnecessary methods 2010-03-03 23:11:39 +00:00
Trustin Lee
8c81a1c8d8 Oops, wrong validation 2010-03-03 09:17:30 +00:00
Trustin Lee
6aa6694822 Related issue: NETTY-297 Non-ssl clients connecting to ssl server should be able to fail fast
* More strict SSL packet validation in SslHandler
2010-03-03 09:15:06 +00:00
Trustin Lee
3bbad77075 Resolved issue: NETTY-297 Non-ssl clients connecting to ssl server should be able to fail fast
* More strict SSL packet validation in SslHandler
2010-03-03 09:01:38 +00:00
Trustin Lee
5558f02003 Resolved issue: NETTY-298 Make ChannelBuffer setter/writer methods to accept int for narrower types 2010-03-03 05:35:01 +00:00
Trustin Lee
a3bae359bd Javadoc clean up 2010-02-24 00:24:41 +00:00
Trustin Lee
52ef2a5330 Upgrade your JDK to use FileRegion 2010-02-24 00:23:33 +00:00
Trustin Lee
049e97f03b Modified HTTP file server to utilize FileRegion support properly 2010-02-23 07:29:10 +00:00
Trustin Lee
aa4b8a2a6c * Changed the type of WriteCompletionEvent.amount from int to long
* Added back FileRegion since mapped buffer does not outperform transferTo()
* Refactored SocketSendBufferPool to support FileRegion
2010-02-23 07:18:58 +00:00
Trustin Lee
1989bd037b Collection -> List 2010-02-23 05:45:53 +00:00
Trustin Lee
69450ee6ec Doubled the default number of I/O threads which seems to be more sensible better performing default 2010-02-23 05:42:39 +00:00
Trustin Lee
b010cd15a5 Imlemented progress notification for NIO writes 2010-02-23 01:12:36 +00:00
Trustin Lee
43a603cfde * Removed unnecessary conditional
* Optimized the use of memory bandwidth in NioWorker.read
2010-02-23 00:56:04 +00:00
Trustin Lee
03268034c7 Removed unnecessary conditional 2010-02-23 00:43:15 +00:00
Trustin Lee
cc3b07c655 Removed FileRegion since a FileChannel can be mapped into a ChannelBuffer. What a lovely abstraction. 2010-02-22 06:18:39 +00:00
Trustin Lee
9ecd573937 Fixed Eclipse compiler warnings 2010-02-22 06:09:54 +00:00
Trustin Lee
59052be709 Related issue: NETTY-293 sendfile() support for NIO TCP transport
Support for progress notification of time-consuming I/O operations
* Added FileRegion
* Added ChannelFutureProgressListener
* Added ChannelFuture.setProgress()
* Performance seems to stay same as before fortunately
2010-02-19 09:58:38 +00:00
Trustin Lee
b97648f95e Comment fix 2010-02-19 09:18:10 +00:00
Trustin Lee
28bac1978f This should be volatile anyhow 2010-02-19 09:08:01 +00:00
Trustin Lee
0066f9c3ff Removed unnecessary volatiles from the examples 2010-02-19 09:02:57 +00:00
Trustin Lee
a462d0e85e Removed volatile from some internal state member variables in decoders because they are always either:
1) modified from the same threads or
2) protected by happens-before condition that is met by Executor's internal synchronization constructs
2010-02-19 09:00:00 +00:00
Trustin Lee
aa8e48df9f Removed 'volatile' from the member variables that are protected by ThreadPoolExecutor's BlockingQueue operations 2010-02-19 08:43:13 +00:00
Trustin Lee
e8c320c6c3 * Removed 'volatile' from the member variables that are protected by synchronized (this) block
* Updated comments regarding thread safety
2010-02-19 08:23:48 +00:00
Trustin Lee
de8671df93 Documentation fix 2010-02-19 05:13:48 +00:00
Trustin Lee
baf0978709 Fixed documentation 2010-02-19 04:49:52 +00:00
Trustin Lee
9980a25f7b Reverted back for backward compatibility 2010-02-19 04:37:36 +00:00
Trustin Lee
81d019e76a Reverted back for backward compatibility 2010-02-19 04:36:30 +00:00
Trustin Lee
d1c6ed6531 Added ChunkedFile interface 2010-02-19 04:18:35 +00:00
Trustin Lee
c7179aa28a Renamed ChunkedFile to ChunkedOioFile. ChunkedFile will be added again as an interface 2010-02-19 04:11:59 +00:00
Trustin Lee
d9de1675d0 Applied send buffer pool to nio datagram transport 2010-02-19 03:28:11 +00:00
Trustin Lee
32ef9e3e52 * Implemented send buffer pooling
* Renamed ReadBufferPool to SocketReceiveBufferPool
2010-02-19 03:10:30 +00:00
Trustin Lee
1a04586b54 Tiny optimization 2010-02-19 02:40:55 +00:00
Trustin Lee
2f601dcbaf WriteBuffer -> WriteRequestQueue 2010-02-19 01:41:32 +00:00
Trustin Lee
4be652ad04 Removed an unused method in ReadBufferPool 2010-02-19 01:13:38 +00:00
Trustin Lee
558c02b0df Tiny optimization 2010-02-19 00:57:23 +00:00
Trustin Lee
f4d7f573a1 Upgraded to Google Protobuf 2.3.0 2010-02-19 00:14:25 +00:00
Trustin Lee
cee5f945dc Tiny code clean up 2010-02-19 00:03:33 +00:00
Trustin Lee
545acfdf42 Resurrected the good old direct buffer pool as a read buffer pool instead of using it both for reads and writes. (write buffer pool needs different treatment.) 2010-02-18 15:51:06 +00:00
Trustin Lee
a9c90d0e4a Reverted back the changes about direct buffer pool - it doesn't seem to improve the performance as much as additional memory consumption and the latest one often led to OOME during testing 2010-02-18 13:00:31 +00:00
Trustin Lee
263e1ed601 It seems like using buffer pool during read doesn't seem to help performance at all. 2010-02-18 11:34:34 +00:00
Trustin Lee
e7c329cc4e Documentation 2010-02-18 11:33:54 +00:00
Trustin Lee
3412003bc2 Fixed an infinite loop 2010-02-18 04:29:58 +00:00
Trustin Lee
4ef4a92281 * Added ChunkedInput.isEndOfInput() and changed the contract of ChunkedInput.nextChunk() and hasNextChunk() to support slow streams 2010-02-18 04:24:41 +00:00