Repository navigation
getcwd hangs and leaks mem on Solaris <= 9 in very long file name case #51225
Description
| BPO | 6976 |
|---|---|
| Nosy | @skrah |
| Superseder | |
| Files |
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 2010-07-20.18:48:26.522>
created_at = <Date 2009-09-23.05:55:36.880>
labels = ['build', 'performance']
title = 'getcwd hangs and leaks mem on Solaris <= 9 in very long file name case'
updated_at = <Date 2010-07-20.18:48:26.520>
user = 'https://bugs.python.org/iandekit'bugs.python.org fields:
activity = <Date 2010-07-20.18:48:26.520>
actor = 'skrah'
assignee = 'none'
closed = True
closed_date = <Date 2010-07-20.18:48:26.522>
closer = 'skrah'
components = ['Build']
creation = <Date 2009-09-23.05:55:36.880>
creator = 'iandekit'
dependencies = []
files = ['14955']
hgrepos = []
issue_num = 6976
keywords = []
message_count = 3.0
messages = ['93028', '110925', '110945']
nosy_count = 3.0
nosy_names = ['skrah', 'iandekit', 'BreamoreBoy']
pr_nums = []
priority = 'normal'
resolution = 'duplicate'
stage = 'resolved'
status = 'closed'
superseder = '9185'
type = 'resource usage'
url = 'https://bugs.python.org/issue6976'
versions = ['Python 2.6']
Activity
iandekit commentedon Sep 23, 2009
test_posix hung on Solaris 9 ... traced to getcwd test hanging.
This in turn was traced to the very long filename case...
It seems posixmodule was modified (since py2.4.3 at least)
to retry getcwd with a bigger buffer if ERANGE occurs.
However on Solaris 9 its not documented that ERANGE also
occurs if getcwd(3) can't cope with the path length, even
if the buffer is big enough. This causes posix_getcwd() to
loop malloc'ing a bigger buffer, forever.
I enclose a patch that limits the damage, to 1Mbyte at least.
(not sure if more recent Solaris 9 patches than we have
provide another solution)
On Solaris 10, there is no problem as the getcwd() is implemented
as a system call.
BreamoreBoy commentedon Jul 20, 2010
Are there any Solaris people who could look at this? The patch file contains a very small change to posixmodule.c.
skrah commentedon Jul 20, 2010
This has just been fixed in bpo-9185.