search

adrienverge / yamllint

A linter for YAML files.
GNU General Public License v3.0
2.9k stars 278 forks source link

skip ignored files at discovery #646

Open james-powis opened 10 months ago

james-powis commented 10 months ago

Fixes: https://github.com/adrienverge/yamllint/issues/399

Not sure how coverage decreased, I did not touch that file ;)

ignored files should be skipped early on to prevent other issues such as tripping on intentionally ignored broken symlinks.

There are many reasons that a file is ignored, one of those reasons is that it is a symlink which for valid reasons is broken. Without this change, this results in an [Errno 2] No such file or directory: error which can cause CI/CD to fail.

There likely is efficiencies by skipping to read in the skipped files at startup, but that is not the point of this PR.

coveralls commented 10 months ago

Coverage Status

coverage: 99.797% (-0.03%) from 99.822% when pulling 5417f9619c03c060b8e689bcb951b9a023650d74 on james-powis:skip-ignored-files-on-load into 0d7ae7bf6b852fd739ba385b686ea83ed3164a79 on adrienverge:master.

adrienverge commented 10 months ago

Thanks for contributing!

Not sure how coverage decreased, I did not touch that file ;)

Coverage decreased indeed, because changes to yamllint source code should come with tests, to make sure there won't be regression. It was requested in the original issue #399 ;) A new test simulating the scenario of the initial message of #399 should be fine. See tests/test_config.py for how existing tests look like.

If you add a check not conf.is_file_ignored(filepath) early, should this check be removed later in yamllint execution? (I'm not 100% sure, but I remember that touching find_files_recursively() + conf.is_file_ignored() is tricky and could create problems, it needs special attention)

james-powis commented 10 months ago

@adrienverge I did not touch linter.py so its coverage decreasing is not related to my change to cli.py and likely a related to something else....

Furthermore there is a pretty significant gordian knot related to these checks inside and out within the code. Refactoring your code is not my interest. I was just trying to contribute a fix, which this does, and your users have been complaining about for over 2 years, which I also was impacted by. I have a monkey patch in place and do not need you to accept this. I just figured I would try and help out. This is an exceptionally trivial change, and one which is exceptionally difficult to write a test for. Its just not that valuable for me to deal with that silly difficulty. However I am sure you could piece together the test modifications in my other PR if you need inspiration.