Motivation:
People need to set all length fields manually when creating a memcache message and it's error prone. See #2736 for more dicussion.
Modifications:
This patch adds the logic to update the keyLength, extrasLength and totalBodyLength when key, extras or content is set.
Result:
The length fields of memcache messages will be updated automatically.
Motivation:
The key can be ByteBuf to avoid converting between ByteBuf and String. See #3689.
Modifications:
Replace the type of key with ByteBuf.
Result:
The type of key becomes ByteBuf.
Motivation:
Some duplicated methods in message types of codec-memcache can be cleaned using AbstractReferenceCounted.
Modifications:
Use AbstractReferenceCounted to avoid duplicated methods.
Result:
Duplicated methods are cleaned.
Motivation:
BinaryMemcacheObjectAggregator doesn't retain ByteBuf `extras`. So `io.netty.util.IllegalReferenceCountException: refCnt: 0, decrement: 1` will be thrown when aggregating a message containing `extras`. See the unit test for an example.
Modifications:
`ratain` extras to fix IllegalReferenceCountException.
Result:
`extras` is retained.
Motivation:
AbstractBinaryMemcacheDecoder.currentMessage is not retained after sending it out. Hence, if a message contains `extras`, `io.netty.util.IllegalReferenceCountException` will be thrown in `channelInactive`.
Modifications:
Retain AbstractBinaryMemcacheDecoder.currentMessage After putting it to `out` and release it when it's not used.
Result:
No IllegalReferenceCountException or leak.
Motivation
----------
Currently, only the fixed 24 bytes are allocated for the header and
then all the params as well as the optional extras and key are written
into the header section.
It is very likely that the buffer needs to expand at least two times
if either the extras and/or the key take up more space.
Modifications
-------------
Since at the point of allocation we know the key and extras length,
the buffer can be preallocated with the exact size, avoiding unnecessary
resizing and even allocating too much (since it uses power of two
internally).
Result
------
Less buffer resizing needed when encoding a memcache operation.
Motivation:
ChannelInboundHandler and ChannelOutboundHandler both can implement exceptionCaught(...) method and so we need to dispatch to both of them.
Modifications:
- Correctly first dispatch exceptionCaught to the ChannelInboundHandler but also make sure the next handler it will be dispatched to will be the ChannelOutboundHandler
- Add removeInboundHandler() and removeOutboundHandler() which allows to remove one of the combined handlers
- Let *Codec extends it and not ChannelHandlerAppender
- Remove ChannelHandlerAppender
Result:
Correctly handle events and also have same behavior as in 4.0
Motivation:
As part of merging / cherry-picking there were some Erroneous imports added to AbstractMemcacheObjectAggregator.
Modifications:
- Remove Imports from AbstractMemcacheObjectAggregator.
Result:
Code now builds.
Motivation:
The HttpObjectAggregator always responds with a 100-continue response. It should check the Content-Length header to see if the content length is OK, and if not responds with a 417.
Modifications:
- HttpObjectAggregator checks the Content-Length header in the case of a 100-continue.
Result:
HttpObjectAggregator responds with 417 if content is known to be too big.
Motivation:
Current AbstractMemcacheObjectEncoder does unnecessary message type checking if the message is MemcacheMessage type.
Modifications:
Returns after encoding MemcacheMessage message.
Result:
Small performance improvement for this encoder.
Related issue: #1133
Motivation:
There is no support for client socket connections via a proxy server in
Netty.
Modifications:
- Add a new module 'handler-proxy'
- Add ProxyHandler and its subclasses to support SOCKS 4a/5 and HTTP(S)
proxy connections
- Add a full parameterized test for most scenarios
- Clean up pom.xml
Result:
A user can make an outgoing connection via proxy servers with only
trivial effort.
If a connection is closed unexpectedly while
AbstractBinaryMemcacheDecoder decodes a message, the half-constructed
message's content might not be released.
Motivation:
Persuit for consistent method naming across all classes
Modifications:
Remove 'get' prefix for the getter methods in codec-memcache
Result:
More consistent method naming
