Re: [PATCH v2 2/5] docs: automarkup.py: Fix regexes to solve sphinx 3 warnings

From: Jonathan Corbet
Date: Wed Oct 14 2020 - 16:16:20 EST


On Wed, 14 Oct 2020 20:09:10 +0000
Nícolas F. R. A. Prado <nfraprado@xxxxxxxxxxxxxx> wrote:

> One I had noted down was:
>
> WARNING: Unparseable C cross-reference: '调用debugfs_rename'
>
> which I believe occurred in the chinese translation.
>
> I think the problem is that in chinese there normally isn't space between the
> words, so even if I had made the regexes only match the beginning of the word
> (which I didn't, but I fixed this in this patch with the \b), it would still try
> to cross-reference to that symbol containing chinese characters, which is
> unparsable to sphinx.
>
> So since valid identifiers in C are only in ASCII anyway, I used the ASCII flag
> to make \w, and \d only match ASCII characters, otherwise they match any unicode
> character.

OK, this all makes sense, as does your fix. The one thing I would ask
would be to put that warning into the changelog for future reference.

Thanks,

jon