Author haypo
Recipients Demur Rumed, Mark.Shannon, berker.peksag, haypo, python-dev, serhiy.storchaka, yselivanov
Date 2016-09-11.09:01:16
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1473584477.03.0.60426116057.issue27213@psf.upfronthosting.co.za>
In-reply-to
Content
Author: Berker Peksag (berker.peksag) * (Python committer) 	Date: 2016-09-11 08:55

Berker: "I'm getting segfaults with a77756e480c2 (thanks to 'hg bisect') on Ubuntu 12.04 (...)
test_recursionlimit_recovery (...) test_main_recursion_error (...)"

We should check if the change counts correctly the number of frames.

In callfunc-10.patch, Serhiy modified a counter related to recursion limit/number of frames to fix test_traceback (issue #28050). He wrote on the review:

"Seems the number 50 is arbitrary. Now we should increase it. Why? This issue needs further investigation. Maybe this is a sign of regression."

To me, it's suspicious that we have to modify an unit test. Changing the opcode or internals must not have an impact on unit tests on recursion error.

Note: It might also be related to my FASTCALL changes, but you wrote that "hg bisect" pointed you to this change.


Berker: "I'm also getting the following error with Serhiy's latest patch: python: Objects/abstract.c:2385: _PyStack_AsDict: Assertion `PyDict_GetItem(kwdict, key) != ((void *)0)' failed."

This one might be related to the compact dict change.


Note: Just to make be sure that it is not a local issue, please try a "make distclean" and make sure that all .pyc files are removed, and then recompile Python.
History
Date User Action Args
2016-09-11 09:01:17hayposetrecipients: + haypo, Mark.Shannon, python-dev, berker.peksag, serhiy.storchaka, yselivanov, Demur Rumed
2016-09-11 09:01:17hayposetmessageid: <1473584477.03.0.60426116057.issue27213@psf.upfronthosting.co.za>
2016-09-11 09:01:17haypolinkissue27213 messages
2016-09-11 09:01:16haypocreate