2

Possible Duplicate:
Bypassing buffering of subprocess output with popen in C or Python

I'm building a wrapper around a server cmd line script that should run indefinitely. What I need to do is to get the current stout without waiting for the subprocess to finish.

I mean, if I run the following, everything works fine:

ls = Popen(["ls"], stdout=PIPE)
output = ls.stdout.read()

But if I do the same with an indefinitelly running program:

server = Popen(["python","-m","SimpleHTTPServer"], stdout=PIPE)
output = server.stdout.read()

It will not come back...

Update: Even

 output = server.stdout.read(1)

hangs...

Do you know if there's a way to capture partial output from a Popen (or similar threading implementation) in an OS independent way?

Community
  • 1
  • 1
Santi
  • 4,428
  • 4
  • 24
  • 28
  • See the solutions [here](http://stackoverflow.com/questions/1410849/manipulating-pipe-buffer-size-in-c-or-python) – Ned Deily Oct 05 '09 at 21:43
  • Is this cross platform? (it doesn't look like) – Santi Oct 05 '09 at 21:56
  • There is a description of the basic problem and a couple of solutions there. If you can modify the running process to do flushes, that is best. If you can't, Alex outlined the pexpect/wexpect trick which should cover both Unix-y and Windows platform. If those don't work and you can't change the way the child process writes/flushes to stdout and stderr, you're likely out of luck. – Ned Deily Oct 05 '09 at 22:09
  • @Santi: confirmed -- pexpect on all Unix variants inc. Mac, wexpect on Windows (if you're running on anything else there may or may not be a solution for other peculiar platforms). – Alex Martelli Oct 05 '09 at 22:25
  • Alright, I'll check those out to see if I'm lucky :) Tks both. – Santi Oct 05 '09 at 22:34

1 Answers1

-1

i would guess that read() returns the entire contents? If you read a fixed size chunk in a loop you may get better results.

sean riley
  • 2,633
  • 1
  • 22
  • 22