add Popen objects to _active only in __del__,

[atila-cheops]

BPO	1467770
Nosy	@loewis
Files	subprocess.py.patch

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2006-04-10.16:05:10.000>
created_at = <Date 2006-04-10.14:54:48.000>
labels = ['library']
title = 'add Popen objects to _active only in __del__, '
updated_at = <Date 2006-04-10.16:05:10.000>
user = 'https://bugs.python.org/atila-cheops'

bugs.python.org fields:

activity = <Date 2006-04-10.16:05:10.000>
actor = 'loewis'
assignee = 'astrand'
closed = True
closed_date = None
closer = None
components = ['Library (Lib)']
creation = <Date 2006-04-10.14:54:48.000>
creator = 'atila-cheops'
dependencies = []
files = ['7145']
hgrepos = []
issue_num = 1467770
keywords = ['patch']
message_count = 3.0
messages = ['49980', '49981', '49982']
nosy_count = 3.0
nosy_names = ['loewis', 'astrand', 'atila-cheops']
pr_nums = []
priority = 'normal'
resolution = 'accepted'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue1467770'
versions = ['Python 2.5']

[atila-cheops]

attached a patch for subprocess.py
this is for bug bpo-1460493 Why not drop the _active list?
popen2.py is patched (revision 43286), this is a (try)
to patch subprocess.py accordingly.
Bug bpo-1183780: Add Popen objects to _active only in __del__.
Cleanup terminated processes as well.

revision 43287 adds a small modification with an
assert, but that is not included.

I'm not sure about the wait(self) function under the
windows part, there is no check for an error, and the
_deadstate protection is not used

verified on windows XP SP2 32bits (python2.4 with
patched subprocess.py from 2.5a1) with the
'popen_bug.py -s '-script from bug bpo-1183780 but with
the following call: pipe = subprocess.Popen("calc")
then in windows xp do a "close group" action on all the
calculator windows (right click on the task bar).

[atila-cheops]

Logged In: YES
user_id=1276121

patch was not uploaded

[loewis]

Logged In: YES
user_id=21627

Thanks for the patch. I made a number of modifications:

• add poll(_deadstate) argument for Unix
• trace whether a child was actually created, else
  returncode would not have been set (not even to None) in __del__
  With these modifications, I committed it as r45234.