Re: [PATCH 2/8] Documentation/sphinx: fix Python string escapes

[Benjamin Gray]

On 14/8/23 11:35 pm, Jonathan Corbet wrote:
> Benjamin Gray <bgray@xxxxxxxxxxxxx> writes:
>
> > Python 3.6 introduced a DeprecationWarning for invalid escape sequences.
> > This is upgraded to a SyntaxWarning in Python 3.12, and will eventually
> > be a syntax error.
> >
> > Fix these now to get ahead of it before it's an error.
> >
> > Signed-off-by: Benjamin Gray <bgray@xxxxxxxxxxxxx>
> > ---
> >   Documentation/sphinx/cdomain.py             | 2 +-
> >   Documentation/sphinx/kernel_abi.py          | 2 +-
> >   Documentation/sphinx/kernel_feat.py         | 2 +-
> >   Documentation/sphinx/kerneldoc.py           | 2 +-
> >   Documentation/sphinx/maintainers_include.py | 8 ++++----
> >   5 files changed, 8 insertions(+), 8 deletions(-)
>
> So I am the maintainer for this stuff...is there a reason you didn't
> copy me on this work?

Sorry, I thought the list linux-doc@xxxxxxxxxxxxxxx itself was enough. I 
haven't done a cross tree series before, I was a bit adverse to CC'ing 
everyone that appears as a maintainer for every patch.

> > diff --git a/Documentation/sphinx/cdomain.py b/Documentation/sphinx/cdomain.py
> > index ca8ac9e59ded..dbdc74bd0772 100644
> > --- a/Documentation/sphinx/cdomain.py
> > +++ b/Documentation/sphinx/cdomain.py
> > @@ -93,7 +93,7 @@ def markup_ctype_refs(match):
> >   #
> >   RE_expr = re.compile(r':c:(expr|texpr):`([^\`]+)`')
> >   def markup_c_expr(match):
> > -    return '\ ``' + match.group(2) + '``\ '
> > +    return '\\ ``' + match.group(2) + '``\\ '
>
> I have to wonder about this one; I doubt the intent was to insert a
> literal backslash.  I have to fire up my ancient build environment to
> even try this, but even if it's right...

Yeah, there is even a file that just has a syntax error. I don't have a 
way to verify the original script was correct, but I have verified this 
series doesn't change the parsed AST.

In this case though, it's generating reST, so it might just be 
conservatively guarding against generating bad markup[1]

[1]: 
https://www.sphinx-doc.org/en/master/usage/restructuredtext/basics.html#inline-markup 


> >   #
> >   # Parse Sphinx 3.x C markups, replacing them by backward-compatible ones
> > diff --git a/Documentation/sphinx/kernel_abi.py b/Documentation/sphinx/kernel_abi.py
> > index b5feb5b1d905..b9f026f016fd 100644
> > --- a/Documentation/sphinx/kernel_abi.py
> > +++ b/Documentation/sphinx/kernel_abi.py
> > @@ -138,7 +138,7 @@ class KernelCmd(Directive):
> >                   code_block += "\n    " + l
> >               lines = code_block + "\n\n"
> >   
> > -        line_regex = re.compile("^\.\. LINENO (\S+)\#([0-9]+)$")
> > +        line_regex = re.compile("^\\.\\. LINENO (\\S+)\\#([0-9]+)$")
>
> All of these really just want to be raw strings - a much more minimal
> fix that makes the result quite a bit more readable:
>
>       line_regex = re.compile(r"^\.\. LINENO (\S+)\#([0-9]+)$")
>                               ^
>                               |
>    ---------------------------+
>
> That, I think, is how these should be fixed.

Yup, I mentioned that at the end of the cover letter. I can automate and 
verify the conversion, but automating what _should_ be treated as a 
'regex' string is fuzzier. Checking if there's a `re.*(` prefix on the 
string should work for most though. I'll give it a shot.

> Thanks,
>
> jon