Closed Remi-Gau closed 3 years ago
Is it possible to reproduce locally (i.e. running mh_style on a specific revision on that repo will trigger changes in lib
)? If so can you send me a revision I could try this on?
Right I think I know what the problem is; in that if you provide a file to mh_style in an excluded dir, it is being checked. I.e.
mh_style .
(will exclude all things in lib/)
mh_style lib/foo.m
(will analyze foo.m)
The reason is that exclude_dir is meant for excluding external repositories; which themselves may or may not have a miss_hit config.
You could do:
mh_style --fix
(i.e. do not provide any paths)lib/
with enable: false
instead of excluding itSo I guess that https://github.com/cpp-lln-lab/CPP_SPM/blob/474d4e6f7aaf8b1f50f3ba4254e6114d831ed89d/.pre-commit-config.yaml#L11 will cause the CI tool to give all the files, instead of running on root
In which case only the enable: false
will work for you i think
The enable:false
worked: pre-commit-pr is now quiet and stays out of my lib
folder.
https://github.com/cpp-lln-lab/CPP_SPM/pull/383
Thanks for the tip.
Will close this.
Glad I could resolve the issue. As always it's really nice to see people using this in anger ;)
Hey,
Not sure this is purely miss_hit related but reporting it here in case someone has an idea.
I am using miss_hit on one of my repo with precommit.
I wanted to try the pre-commit CI as it can run
mh_style --fix
automatically and update an open PR with those fixes (not necessary if you are using precommit locally, but nice to have in case colleagues or contributors don't run miss_hit).More info here: https://pre-commit.ci/
So it is easy to set up and will run on your PR no problem (and update it): Example: https://github.com/cpp-lln-lab/CPP_SPM/pull/378
The problem seems to be that in CI miss_hit is not ignoring the folders that it should ignore: See this report where the
lib
folder is being "searched" by miss_hit eventhough the config tells it not to.https://results.pre-commit.ci/run/github/208231636/1623701596.tnPwryOERc2qQZ_-Qsg64w
This is not a problem that urgently needs a solution but in case someone else comes across the issue or has an idea on how to solve this.
Pinigng @wbekerom as he was involved in the pre-commit PR #185