Repository navigation
file-contents-sorter Not working same way everytime #794
Description
Activity
yeah this set call breaks unique sorting -- it should probably just forbid the combination of unique and ignore case since the result is undefined https://github.com/pre-commit/pre-commit-hooks/blob/v4.3.0/pre_commit_hooks/file_contents_sorter.py#L36
@asottile please review if you have time. I'm hoping this PR is simple enough and goes with the spirit of forbidding the combinations of options as you mentioned. Also, I'd appreciate if you could label w/
hacktoberfest-acceptedso I can get a tree planted, thanks.I'm not going to review something which doesn't pass tests
@asottile Apologies for not seeing the failing tests before asking. I think its ready now.
yeah this set call breaks unique sorting -- it should probably just forbid the combination of unique and ignore case since the result is undefined
v4.3.0/pre_commit_hooks/file_contents_sorter.py#L36Could we respect the original word case if the uncased words are equal? For example, let
Prealways proceedprebecause it's upper case. This can be easily implemented withtuple-based sort keys.def key(s): return (s.lower(), s) ret = sorted(set(word_list), key=key)
Another option is to respect the original order (stable sort). This can use an "
OrderedSet" rather thanset.def unique(iterable): return list(OrderedDict.fromkeys(iterable)) # can be replaced with `dict` for Python 3.7+
Both approaches I list above are deterministic.
- no, it should forbid the combination as I said already…On Mon, Mar 27, 2023, 4:24 AM Xuehai Pan ***@***.***> wrote: yeah this set call breaks unique sorting -- it should probably just forbid the combination of unique and ignore case since the result is undefined v4.3.0/pre_commit_hooks/file_contents_sorter.py#L36 <https://github.com/pre-commit/pre-commit-hooks/blob/v4.3.0/pre_commit_hooks/file_contents_sorter.py?rgh-link-date=2022-07-29T11%3A55%3A28Z#L36> Could we respect the original word case if the uncased words are equal? For example, let Pre always proceed pre because it's upper case. This can be easily implemented with tuple-based sort keys. def key(s): return (s.lower(), s) ret = sorted(set(word_list), key=key) Another option is to respect the original order (stable sort). This can use an "OrderedSet" rather than set. def unique(iterable): return list(OrderedDict.fromkeys(iterable)) # can be replaced with `dict` for Python 3.7+ — Reply to this email directly, view it on GitHub <#794 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN2BH2U7K4LE7U7DJ5BN5TW6FFD3ANCNFSM55AOQ72Q> . You are receiving this because you were mentioned.Message ID: ***@***.***>
- added a commit that references this issue
on Aug 11, 2025 - added a commit that references this issue
on Jan 22, 2026
I get strange issue that file-contents-sorter hook does not work same way everytime. I made little test case which trickers the issue. If i run this 10 times it will sometimes pass and sometimes not. I did not have time to look what problem might be.
Add this to
file_contents_sorter_test.pyAnd run it with
for run in {1..10}; do pytest tests/file_contents_sorter_test.py; doneThis gives me that about 30% pass and 70% fails