Jack DeVries <jdevries3...@gmail.com> added the comment:
Hi all! @eamanu, I went ahead and picked up where you left off. I stopped short of opening up a PR, because I don't want to step on any toes, but I definitely want to do what I can to give this bpo one final shove over the finish line. Despite not creating a PR, you can see the changes here: https://github.com/python/cpython/compare/main...jdevries3133:bpo-19217-slow-assertEq @eamanu, if you prefer, you can probably just merge this branch into `eamanu:fix_bpo-19217`, and the conversation can move forward on GH-11204. Also, below is a summary of what I've done: Revert Change to difflib ======================== Link to the original thread where this change was discussed a bit: https://github.com/python/cpython/pull/11204#discussion_r242369775 I assume this was a performance improvement. It's not clear to me why we are not interested in this part of the diff when the input is a list or tuple. I'm sure someone will help to explain, but this is why I reverted this change: 1. It is not an issue with unittest, and reaches beyond the scope of this bpo 2. The performance problem at the root of this bpo is solved without this change. 3. Reverting this change fixes most of the cascading failures throughout other parts of the test suite. It better constrains this fix to improving the implementation's performance issues while limiting changes to behavior. Obviously, I don't have knowledge of the reasoning behind this change sufficient to write a new bpo, but I think that this small change does deserve its own bpo with an independent explaination of what it does and why it should be done. It's definitely possible that I'm missing something here, so let me know if that's the case! Add Regression Test =================== The regression test basically attempts to reproduce the problem and asserts that the code will complete within one second. It's a pretty rough approach, but I think that it will catch a future regression. If anyone has any better ideas, I'm all ears. Misc ==== I also fixed the one failing test in unittest's own test case, and added a blurb. ---------- nosy: +jack__d _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue19217> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
