[Tutor] basic threading question

[Cameron Simpson]

On 07Jan2016 14:26, richard kappler <richkappler at gmail.com> wrote:
>See previous posts on 'looping generator' for details about the code and
>project.
>
>The brief version, I am reading and parsing a data stream through a socket,
>several actually. Each new connection spawns a thread that reads and
>parses. Should the client close, I want the thread to terminate. Everything
>I've read says you don't kill threads, but if the client closes, the socket
>closes, and the thread is just sitting there, hanging.
>
>If the socket.recv returns 0 bytes, the docs tell me that means the client
>closed and therefore the server socket closes as well. If I do something in
>the thread target function like:
>
>        data_chunks = connection.recv(8192)
>        if len(data_chunks) == 0:
>             break
>
>len(data_chunks)==0 tells me the socket is closed, but does the break kill
>the thread? How do I prove that if it does?

The break doesn't kill the thread, but if the recv loop is all your thread's 
main function does, then sure: when the function exits, the thread exits. For 
example:

  T = Thread(target=gather_func)
  T.start()
  ...

  def gather_func():
    while True:
      data_chunks = connection.recv(8192)
      if len(data_chunks) == 0:
        break
      ... do stuff with data_chunks ...

Then you break exits the loop. Then the function returns since there's no more 
code in the function after the loop. And then the Thread exits because the 
function it was running has finished. There is no need to kill the Thread here.

You can wait for the Thread with T.wait() and probe it with T.is_alive(); see 
the threading module documentation.

BTW, I still recommend you drop use of .recv() and use .read1() to assemble 
packets. See your main discussion.

Cheers,
Cameron Simpson <cs at zip.com.au>