[issue22653] Crash in insertdict

Antoine Pitrou added the comment: Ok, then I've committed the patch to 3.4 and 3.5. -- resolution: -> fixed stage: patch review -> resolved status: open -> closed ___ Python tracker ___

[issue22653] Crash in insertdict

Roundup Robot added the comment: New changeset 9ec84f9b61c6 by Antoine Pitrou in branch '3.4': Issue #22653: Fix an assertion failure in debug mode when doing a reentrant dict insertion in debug mode. https://hg.python.org/cpython/rev/9ec84f9b61c6 New changeset 4ff865976bb9 by Antoine Pitrou in

[issue22653] Crash in insertdict

Changes by Antoine Pitrou : -- versions: -Python 3.3 ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://m

[issue22653] Crash in insertdict

Georg Brandl added the comment: This is not a crash, but an abort, so it's not a security issue. -- ___ Python tracker ___ ___ Python-

[issue22653] Crash in insertdict

Antoine Pitrou added the comment: But 3.3 only receives security fixes, and this is not a security issue. It's up to Georg whether we really wants to backport the fix. -- nosy: +georg.brandl ___ Python tracker ___

[issue22653] Crash in insertdict

Arfrever Frehtes Taifersar Arahesis added the comment: Crash when running test_reentrant_insertion is reproducible also in 3.3 branch, but not in 3.2 branch. insertdict.patch applies and works in 3.3 branch. -- versions: +Python 3.3 ___ Python tracke

[issue22653] Crash in insertdict

Antoine Pitrou added the comment: Le 16/10/2014 15:43, Mark Shannon a écrit : > > I'll put a patch together this weekend, once I've had time to reproduce the > failure. Sorry, I was already writing mine when you said that :-) -- title: Crash in insertdict in debug mode -> Crash in ins

[issue22653] Crash in insertdict in debug mode

Antoine Pitrou added the comment: Here is a patch, with a test that fails deterministically without the fix. -- priority: high -> normal stage: -> patch review title: Crash in insertdict -> Crash in insertdict in debug mode Added file: http://bugs.python.org/file36949/insertdict.patch

[issue22653] Crash in insertdict

Mark Shannon added the comment: The assertion on line 855 is a duplicate of the assertion on line 821, which is just before the reentrant DECREF. The assertion on line 821 appears to be correct. I don't know why the assertion is at the end of the function rather than earlier, but hg blame say

[issue22653] Crash in insertdict

Antoine Pitrou added the comment: So after moving away that specific assert, there doesn't seem to be any problem (I ran the whole test suite in the same interpreter). If it's really a case of the assert being too strict, then the issue isn't very severe, as most people would you a non-debug b

[issue22653] Crash in insertdict

Antoine Pitrou added the comment: Looking down into the insertdict frame: (gdb) p key $14 = 'f' (gdb) p old_value $15 = (gdb) p *ep $16 = {me_hash = -3761688987579986997, me_key = 0x0, me_value = 0x0} So this seems to have hit the following line in insertdict(): if (old_value != NULL) {

[issue22653] Crash in insertdict

Changes by Antoine Pitrou : -- priority: normal -> high versions: +Python 3.4 ___ Python tracker ___ ___ Python-bugs-list mailing list

[issue22653] Crash in insertdict

Changes by Antoine Pitrou : Added file: http://bugs.python.org/file36948/dictcrash.py ___ Python tracker ___ ___ Python-bugs-list mailing list

[issue22653] Crash in insertdict

Antoine Pitrou added the comment: Ok, here is how to reproduce deterministically: 1) apply randomsecret.patch to hardcode the siphash initialization values 2) run "cat dictcrash.py | ./python -i": Python 3.5.0a0 (default:030fda7b1de8+, Oct 16 2014, 14:27:32) [GCC 4.8.1] on linux Type "help", "

[issue22653] Crash in insertdict

Changes by STINNER Victor : -- nosy: +haypo ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python

[issue22653] Crash in insertdict

Changes by Arfrever Frehtes Taifersar Arahesis : -- nosy: +Arfrever ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscri

[issue22653] Crash in insertdict

New submission from Antoine Pitrou: I got a weird crash in an interpreter session. Here is what I did: $ ./python Python 3.5.0a0 (default:fd658692db3a+, Oct 15 2014, 23:13:43) [GCC 4.8.1] on linux Type "help", "copyright", "credits" or "license" for more information. >>> f = open('toto', 'ab')