Robert Collins added the comment: Oh, I got a profile from the test case for my own interest.
6615 seconds .. some highlights that jumped out at me 20001 0.127 0.000 6610.025 0.330 difflib.py:868(compare) which means we're basically using ndiff, which is cubic rather than quadratic - the same performance issue the html module had. I think we'll probably make a better tradeoff by using unified_diff instead which is quadratic. ---------- Added file: http://bugs.python.org/file37001/issue19217-profile.txt _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue19217> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com