Serhiy Storchaka added the comment: For the worst case the drawback is significant:
$ ./python -m perf timeit -s "s = set('a%s' % i for i in range(100))" -- "s.add('test'); s.discard('test')" Unpatched: Median +- std dev: 861 ns +- 82 ns Patched: Median +- std dev: 2.81 us +- 0.18 us How large the benefit in the best case? I can't get any significant difference. $ ./python -m perf timeit -s "a = ['a%s' % i for i in range(1000)]" -- "set(a)" Unpatched: Median +- std dev: 130 us +- 6 us Patched: Median +- std dev: 127 us +- 8 us ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue29476> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com