Improving performance of ProtobufDecoder

[IanS]

I believe most of the difference is that CodedInputStream allocates a 4096
byte buffer every time, which can start becoming a lot of overhead for many
small messages.  I tried this based on a forum post I can't seem to locate
right now.  

Thanks for the suggestion about array, I changed the method to a more
'zero-copy' approach now:

	protected Object decode(ChannelHandlerContext ctx, Channel channel, Object
msg) throws Exception {
		if (!(msg instanceof ChannelBuffer)) {
			return msg;
		}

		ChannelBuffer buf = (ChannelBuffer) msg;
		if(buf.hasArray()) {			
			if(extensionRegistry == null) {
				return prototype.newBuilderForType().mergeFrom(buf.array(),
buf.arrayOffset(), buf.readableBytes()).build();
			}
			else {
				return prototype.newBuilderForType().mergeFrom(buf.array(),
buf.arrayOffset(), buf.readableBytes(), extensionRegistry).build();
			}
		}
		else
		{
			if (extensionRegistry == null) {
				return prototype.newBuilderForType().mergeFrom(
						new ChannelBufferInputStream((ChannelBuffer) msg)).build();
			} else {
				return prototype.newBuilderForType().mergeFrom(
						new ChannelBufferInputStream((ChannelBuffer) msg),
						extensionRegistry).build();
			}
		}
	}

-- 
View this message in context: http://netty-forums-and-mailing-lists.685743.n2.nabble.com/Improving-performance-of-ProtobufDecoder-tp5293790p5299238.html
Sent from the Netty User Group mailing list archive at Nabble.com.