Implement similar fix to issue #225 (fixed in #388) for V1Decoder

[rheatley-pervasid]

We have seen a customer suffer a crash with the following stack trace

Description: The process was terminated due to an unhandled exception.
Exception Info: System.OverflowException
   at NetMQ.GCBufferPool.Take(Int32)
   at NetMQ.Msg.InitPool(Int32)
   at NetMQ.Core.Transports.V1Decoder.EightByteSizeReady()
   at NetMQ.Core.Transports.V1Decoder.Next()
   at NetMQ.Core.Transports.DecoderBase.Decode(NetMQ.Core.Transports.ByteArraySegment, Int32, Int32 ByRef)
   at NetMQ.Core.Transports.StreamEngine.ProcessInput()
   at NetMQ.Core.Transports.StreamEngine.Handle(Action, System.Net.Sockets.SocketError, Int32)
   at NetMQ.Core.Transports.StreamEngine.FeedAction(Action, System.Net.Sockets.SocketError, Int32)
   at NetMQ.Core.Transports.StreamEngine.InCompleted(System.Net.Sockets.SocketError, Int32)
   at NetMQ.Core.IOObject.InCompleted(System.Net.Sockets.SocketError, Int32)
   at NetMQ.Core.Utils.Proactor.Loop()
   at System.Threading.ThreadHelper.ThreadStart_Context(System.Object)
   at System.Threading.ExecutionContext.RunInternal(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object, Boolean)
   at System.Threading.ExecutionContext.Run(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object, Boolean)
   at System.Threading.ExecutionContext.Run(System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object)
   at System.Threading.ThreadHelper.ThreadStart()

This was on version 4.0.1.6. The crash is in our application which has a RouterSocket bound to "tcp://127.0.0.1:9202"

I tracked down issues #225 and #388 and believe this is the same problem. I have so far failed to reproduce the issue using NetMQ or python zmq, it isn't clear to me how to send a large enough message.

I can provoke the crash with some contrived python code

import socket
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect(("127.0.0.1", 9202))
s.send(b'\xFF' * 9 + b'\x00' + b'\x42' * 1000000)

Taking the fix from #388 and applying to V1Decoder.cs prevents the above python code crashing NetMQ. There is a comment in that PR - "Other decoders/encoders are worth checking out too I guess." - I assume this never happened?

Let me know your thoughts

[rheatley-pervasid]

@drewnoakes anything I can do to help progress this?

[rheatley-pervasid]

Oh, looks like it got fixed differently in #1030 Assume you are happy to lose support for messages >2GB and I'll close it?

[drewnoakes]

Apologies for dropping the ball on this.

I'm on my phone and don't have the spec handy. If the type is unsigned then the fix here would be more comprehensive.