optimize check_case_conflict.parents by guykisel · Pull Request #626 · pre-commit/pre-commit-hooks

guykisel

fixes #625

performance for this approach on my team's repo:

λ py test.py
         10001077 function calls in 4.023 seconds

   Ordered by: standard name

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
        1    0.039    0.039    4.023    4.023 <string>:1(<module>)
        4    0.000    0.000    0.000    0.000 _weakrefset.py:38(_remove)
        4    0.000    0.000    0.000    0.000 _weakrefset.py:81(add)
        3    0.000    0.000    0.289    0.096 check_case_conflict.py:13(lower_set)
        3    0.192    0.064    0.289    0.096 check_case_conflict.py:14(<setcomp>)
  3189661    1.037    0.000    2.010    0.000 check_case_conflict.py:17(parents)
        2    0.000    0.000    2.817    1.409 check_case_conflict.py:25(directories_for)
        2    0.807    0.404    2.817    1.409 check_case_conflict.py:26(<setcomp>)
        1    0.146    0.146    3.984    3.984 check_case_conflict.py:29(find_conflicting_filenames)
        2    0.000    0.000    0.000    0.000 subprocess.py:1023(_internal_poll)
        2    0.000    0.000    0.021    0.011 subprocess.py:1040(wait)
        2    0.000    0.000    0.608    0.304 subprocess.py:1067(_communicate)
       22    0.000    0.000    0.000    0.000 subprocess.py:179(Close)

So it's 4 seconds with this approach vs 20 seconds with the previous approach.