ok, ticket was created without any description... nice
So here is a pd-patch that demonstrates the issue.
It seems to appear only after the client get connected.
The patch is similar to the one reported in issue #1135, so it crashes on 64bit Ubuntu.
Here the client1 connect to server and then send float 1, the server should send it back to the client thanks to the [send <socket> <list>( message. But sometimes, client1 receive float 2, targeted to client2. Then the client disconnect after receiving data and connect again.
[send <socket> <list>(
Tested on Ubuntu 12.04 32bit.
You seem to have CSS turned off.
Please don't fill out this field.
confirmed on ubuntu 13.10 / 64 bit
It may be that when closing/opening the sockets, client 1 on socket 11 becomes client 1 on socket 12, since the threads and Pd can be a bit out of sync. The problem goes away if you add a small delay to one of the client sends.
There is something that I don't understand.
Here is the behavior I'am expecting :
1- [tcpserver] listen for connection on port 48000
2- [client1] connect to the server, so a socket is opened and [tcpserver] seems to report it right
3- after 2 seconds, [client1] sends some data to the server.
4- the server reports the socket on which data comes and then reports the data on the right outlet
5- the server send those data back to the client's socket
6- the client gets the data and disconnect
Am I right or am I expecting surrealistic behavior ?
According to some [print] I've put in my patch, it appears that [tcpserver] reports the right socket number, but send the wrong data to that socket.
Or maybe the client's socket changed between client send and receive routine.
How this is possible ?
Putting a big delay between connection and sending data doesn't help. Or at least, 2 seconds is not enough.
I also tried with a random delay between 2 and 2.1 sec on each client, I still got wrong data.
So I do think there is still an issue with this but I don't know if it is on server or client side.
Sign up for the SourceForge newsletter: