[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <28c45f53-a3ff-428f-ba99-ebb09e0581d3@gmail.com>
Date: Fri, 12 Sep 2025 19:15:44 +0900
From: Akira Yokosawa <akiyks@...il.com>
To: mchehab+huawei@...nel.org
Cc: corbet@....net, jani.nikula@...ux.intel.com, linux-doc@...r.kernel.org,
linux-kernel@...r.kernel.org, tmgross@...ch.edu,
Akira Yokosawa <akiyks@...il.com>
Subject: Re: [PATCH v4 08/19] tools/docs: sphinx-build-wrapper: add a wrapper
for sphinx-build
[-CC: rust people and list]
OK, Looks like I have to bite.
On Wed, 10 Sep 2025 15:33:34 +0200, Mauro Carvalho Chehab wrote:
[...]
> The current approach of using LaTeX for PDF is dirty:
>
> - Sphinx can't produce a LaTeX file from the Kernel trees without
> hundreds of warnings;
> - latexmk hides some of them, but even it just one warning is reported,
> the return status is not zero.
This sounds interesting to me.
As far I remember, I have never seen such strange results of latexmk
under build envs where all the necessary packages are properly installed.
I think what you are trying here is to paper over whatever bug in latexmk/
xelatex by always ignoring their exit status. Am I guessing right?
If that is the case, I'd rather report such an issue at upstream
lagtexmk/xelatex.
Can you please provide a reproducer of such an issue?
Or you saw something while you were tackling issues you claimed in the
cover letter [1] of "Fix PDF doc builds on major distros" series?
[1]: https://lore.kernel.org/cover.1755763127.git.mchehab+huawei@kernel.org/
Thanks,
Akira
Powered by blists - more mailing lists