see here ZlibDecoder.cs and ZlibEncoder.cs:
https://github.com/matrix-xmpp/matrix-vnext/tree/master/src/Matrix/Network/Codecs
https://github.com/matrix-xmpp/matrix-vnext/tree/master/src/Matrix/Network/Codecs
There is also compression in the Http classes. All those classes did not exist when I wrote my compression stuff.
See: https://github.com/Azure/DotNetty/tree/dev/src/DotNetty.Codecs.Http
See: https://github.com/Azure/DotNetty/tree/dev/src/DotNetty.Codecs.Http
i have a problem, my application using .net core 3.1.201, DotNetty 0.6.0, and MongoDB. it's a MMOG Server, so it will send many Packet. on windows server, it runs very well, and performance is good; but on linux, half of the process time is running __pv_queued_spin_lock_slowpath. how can
how can i find the problem of my application?
thanks a lot
Are there any side effects that I should be aware of ?
I see you can get containers for some of the services. Is there a common way to deploy from docker compose/kubernetes? For things like cloud functions + api gateways.
When sending a IByteBuffer to a channel via WriteAsync/WriteAndFlushAsync - will it a) increment the reference counter when passed in, and b) decrement it once it has been sent?
Trying to work out when I need to decrement my reference (i.e. can I do it after calling .Async, or do I need to .ContinueWith to release it?)
I'm having issues where with libuv the channel is made immediately Inactive. When not using libuv, it seems to work (but perf is terrible)
the client should open and close connection for every request
and we should be able to throttle number of requests
@vsudhini see the samples here:
https://github.com/Azure/DotNetty/tree/dev/examples
The EchoClient is probably the easiest to get started with
_bootstrap = new Bootstrap();
_bootstrap
.Group(_eventGroup)
.Channel<TcpSocketChannel>()
.Option(ChannelOption.TcpNodelay, true)
.Option(ChannelOption.SoBacklog, 100)
.Option(ChannelOption.SoRcvbuf, int.MaxValue)
.Handler(new LoggingHandler("LSTN"))
.Handler(new ActionChannelInitializer<ISocketChannel>(channel =>
{
IChannelPipeline pipeline = channel.Pipeline;
//pipeline.AddLast(new DelimiterBasedFrameDecoder(int.MaxValue, BetaMessageEndDelimeter));
_bootstrap
.Group(_eventGroup)
.Channel<TcpSocketChannel>()
.Option(ChannelOption.TcpNodelay, true)
.Option(ChannelOption.SoBacklog, 100)
.Option(ChannelOption.SoRcvbuf, int.MaxValue)
.Handler(new LoggingHandler("LSTN"))
.Handler(new ActionChannelInitializer<ISocketChannel>(channel =>
{
IChannelPipeline pipeline = channel.Pipeline;
//pipeline.AddLast(new DelimiterBasedFrameDecoder(int.MaxValue, BetaMessageEndDelimeter));
pipeline.AddLast(new ClientHandler());
}));
ClientHandler is getting only partial message
how can i read the entire message coming back from server
@vsudhini Good question. You could create a state machine, but there are really many different possible ways to create a request and response system for any type of socket server, in any language that supports socket servers. Anyway, you could create a class that parses the responses and pass the socket object inside the constructor, and then create a class that takes the socket object to respond to as a parameter as well as the parsed message (or the parsed message object if you were to create a MessageParser class or something), and since you passed those two objects as parameters, you can use those object to send data back to the client using the socket object that was passed as a parameter in the responder class. Make sense? If not, DM me! :smile:
Hi... What would be the best way to block/wait reading more messages for a certain Channel/Context? I have a solution that is acting as a TCP bridge between a Gateway and an IoT Hub.. And while the application is connecting to the IoT hub, I would like to paus reading more messages from the Gateway. Is this solvable with DotNetty or do I have to create my own locking mechanism?
Have anyone experienced that your application after a while runs out of memory because the PooledByteBufferAllocator eats all the memory from the instance? And in that case does anyone have a solution for it?
The problem is that when a client disconnects it still doesn't release the memory, and our clients goes up and down quite often.
The problem is that when a client disconnects it still doesn't release the memory, and our clients goes up and down quite often.
