Example of --regex flag in use

[luzpaz]

Can someone demonstrate how to use the codespell --regex= flag?

[peternewman]

See for example here:
https://github.com/OpenLightingProject/ola/blob/master/.travis-ci.sh#L147
codespell --check-filenames --quiet 2 --regex "[a-zA-Z0-9][\\-'a-zA-Z0-9]+[a-zA-Z0-9]"

This will check words within underscore separated variable names. It's not perfect, but it's the best I've got so far, I think due to some Python regex limitations.

[peternewman]

Thinking about it, this will only match three letter or more words. But I've a feeling I couldn't do +? for non-greedy in Python.

[dwo]

I was surprised that the default regex includes _ as part of words (via \w which is [a-zA-Z0-9_]):

codespell/codespell_lib/_codespell.py

Line 29 in 2d12d7a

word_regex_def = u"[\\w\\-'’`]+"

It's a surprising default since so much code uses snake_case to separate words in variable names.

It seems like using -r "[a-zA-Z0-9\-'’\`]+" (the default with an unrolled \w and dropping the underscore) would also achieve checking words separated by underscores?

I'll see if I can get a moment to write this up into the README.

[larsoner]

It seems reasonable to me to change the default actually, and document how to get the old behavior back for people who end up needing it

[Gabrielcarvfer]

See for example here: https://github.com/OpenLightingProject/ola/blob/master/.travis-ci.sh#L147 codespell --check-filenames --quiet 2 --regex "[a-zA-Z0-9][\\-'a-zA-Z0-9]+[a-zA-Z0-9]"

Thank you for the example. Ended up extending it a bit to catch subwords for [c|C]amelCase and snake_case. 😄
(?<![a-z])[a-z'`]+|[A-Z][a-z'`]*|[a-z]+'[a-z]*|[a-z]+(?=[_-])|[a-z]+(?=[A-Z])|\d+

https://regex101.com/r/0LFI8a/2