Message227251
Victor, the message in my patch is copied almost verbatim from the
current subprocess' documentation [1]
[1] https://hg.python.org/cpython/file/850a62354402/Doc/library/subprocess.rst#l57
People use `call(cmd, stdout=PIPE)` as a *broken* way to suppress
output i.e., when they actually want `call(cmd, stdout=DEVNULL)`
The issue with `call(cmd, stdout=PIPE)` that it *appears* to work
if cmd doesn't produce much output i.e., it might work in tests but
may hang in production.
It is unrelated to check_output(), getstatusouptut() or getoutput(). |
|
Date |
User |
Action |
Args |
2014-09-22 04:02:52 | akira | set | recipients:
+ akira, vstinner, juj |
2014-09-22 04:02:52 | akira | set | messageid: <1411358572.81.0.47017649792.issue22442@psf.upfronthosting.co.za> |
2014-09-22 04:02:52 | akira | link | issue22442 messages |
2014-09-22 04:02:52 | akira | create | |
|