[issue1663329] subprocess/popen close_fds perform poor if SC_OPEN_MAX is hi

[Mike Klaas]

Mike Klaas added the comment:

This problem has also afflicted us.

Attached is a patch which adds closerange(fd_low, fd_high) to the posix 
(and consequently os) module, and modifies subprocess to use it.  Patch 
is against trunk but should work for 2.5maint.

I don't really think that this is useful enough to add to the public 
api, but it results in a huge performance benefit for subprocess:

[python-trunk]$ ./python -m timeit -s 'import python_close' 
'python_close.go(100000)'
10 loops, best of 3: 358 msec per loop
[python-trunk]$ ./python -m timeit -s 'import os' 'os.closerange(4, 
100000)'
10 loops, best of 3: 20.7 msec per loop
[python-trunk]$ ./python -m timeit -s 'import python_close' 
'python_close.go(300000)'
10 loops, best of 3: 1.05 sec per loop
[python-trunk]$ ./python -m timeit -s 'import os' 'os.closerange(4, 
300000)'10 loops, best of 3: 63 msec per loop

[python-trunk]$ cat python_close.py
import os, sys
def go(N):
        for i in xrange(4, N):
                try:
                        os.close(i)
                except:
                        pass

----------
nosy: +klaas

_____________________________________
Tracker <report at bugs.python.org>
<http://bugs.python.org/issue1663329>
_____________________________________
-------------- next part --------------
A non-text attachment was scrubbed...
Name: posix_closerange.patch
Type: application/octet-stream
Size: 1915 bytes
Desc: not available
Url : http://mail.python.org/pipermail/python-bugs-list/attachments/20071016/e59cde84/attachment.obj