Re: [PATCH 03/12] docs: kdoc: backslashectomy in kdoc_parser

[Jonathan Corbet <corbet@lwn.net>]

Mauro Carvalho Chehab <mchehab+huawei@kernel.org> writes:

> In time: I mean *possible* POSIX violations. 
>
> I very much prefer more backslashs than needed or use re.escape()
> than to read thoughtfully POSIX and Python-specific specific rules.
>
> Python, in particular, is not very reliable between versions - each
> new version comes with a set of incompatible changes.

People like to complain about that ... but the LWN site code has been in
Python since 2002, and with the notable exception of the Python 3
transition, it has been really painless.

FWIW, the 3.13 re module will warn if it sees a construction like "[[]"
- evidently there is a possible future change that might make "[[" mean
something special and new.  So I avoided that combination.

> In summary, if you agree with always escape brackets, curly brackets and
> parenthesis inside brackets on kernel-doc, we should be free of not
> opened/not closed "symbols" with is an annoyance at least for me, and
> we should be freed of possible POSIX issues and undefined behavior(*).

It shall be a mighty struggle, but I think I can find a way to live with
that... :)

Thanks,

jon