Issue 816476: Fatal Python error: GC object already tracked (original) (raw)

This issue has been migrated to GitHub: https://github.com/python/cpython/issues/39352

classification

process

Created on 2003-10-02 10:42 by thenninger, last changed 2022-04-10 16:11 by admin. This issue is now closed.

Messages (6)
msg18507 - (view)	Author: Thorsten Henninger (thenninger)	Date: 2003-10-02 10:42
Fatal Python error: GC object already tracked OS: Windows 2000 Server (Service Pack 4) Python: 2.3.1 (#47) pycurl: 7.10.6 I got this Fatal Error while fetching Webpages with pycurl. I do not have any other addiotional information. This error is reproducible and did not occur with Python2.2. Please contact me, if you need further information
msg18508 - (view)	Author: Michael Hudson (mwh)	Date: 2003-10-02 10:51
Logged In: YES user_id=6656 Why do you think this is a bug in Python and not pycurl?
msg18509 - (view)	Author: Tim Peters (tim.peters) *	Date: 2003-10-02 16:25
Logged In: YES user_id=31435 You should report this to the pycurl project. The specific error message you report is new in Python 2.3, doing a check for errnoneous usage of the Python C API. That check didn't exist in Python 2.2. While it's possible there's a previously unreported error in Python 2.3 here, it's much more likely that the problem is in pycurl's C code, and that the error there simply went undetected before Python 2.3, appearing to work by luck.
msg18510 - (view)	Author: Thorsten Henninger (thenninger)	Date: 2003-10-02 18:07
Logged In: YES user_id=878819 thank you, that makes sense! i will notice pycurl and have a look in th pycurl source!
msg18511 - (view)	Author: Tim Peters (tim.peters) *	Date: 2003-10-04 17:18
Logged In: YES user_id=31435 Reopening until we know more.
msg18512 - (view)	Author: Tim Peters (tim.peters) *	Date: 2004-01-01 05:01
Logged In: YES user_id=31435 Closed as (presumed) 3rd-party, won't fix. There's been no followup for 3 months; no other report of this symptom has been filed; and someone on c.l.py today reported what appears to be exactly the same problem *again* coming from use of pycurl. So all the evidence points to pycurl, and none so far to Python.