Re: linux-next: build failure after merge all the trees

[Mauro Carvalho Chehab <mchehab+huawei@kernel.org>]

Em Tue, 20 May 2025 07:18:02 +0200
Mauro Carvalho Chehab <mchehab+huawei@kernel.org> escreveu:

> Em Mon, 19 May 2025 12:33:03 -0700
> Randy Dunlap <rdunlap@infradead.org> escreveu:
> 
> > On 5/14/25 7:33 PM, Akira Yokosawa wrote:  
> > > [+CC linux-doc]
> > > 
> > > Hi,
> > > 
> > > On Thu, 8 May 2025 14:39:11 +0200, Mauro Carvalho Chehab wrote:    
> > >> Em Thu, 8 May 2025 22:25:31 +1000
> > >> Stephen Rothwell <sfr@canb.auug.org.au> escreveu:    
> > > [...]
> > >     
> > >>>
> > >>> So, I used "KERNELDOC=$(pwd)/scripts/kernel-doc.pl" and tried again.
> > >>>
> > >>> I got these (new) messages:
> > >>>
> > >>> Error: Cannot open file drivers/virt/coco/tsm-mr.c
> > >>> Error: Cannot open file drivers/virt/coco/tsm-mr.c
> > >>> WARNING: kernel-doc 'scripts/kernel-doc.pl -rst -enable-lineno -export drivers/virt/coco/tsm-mr.c' failed with return code 2
> > >>>
> > >>> (and a few other innocuous ones)
> > >>>
> > >>> So your guess is good.
> > >>>
> > >>> It would be nice to have the Python kernel-doc fixed as well as the
> > >>> devsec-tsm tree.    
> > >>
> > >> With regards to kernel-doc, failing to build if a file is missing
> > >> is the right thing to do.    
> > > 
> > > Mauro, I don't agree here.
> > > 
> > > With the perl version of kernel-doc, a typo in a file path doesn't cause
> > > a fatal error of docs build.
> > > 
> > > kernel-doc as python class libs ends up in a fatal error.
> > > 
> > > Here is a log of such a fatal error (on top of current docs-next with
> > > intentional typo made in a pathname in one of .. kernel-doc::
> > > 
> > > -----------------------------------------------------------------
> > > Sphinx parallel build error!
> > > 
> > > Versions
> > > ========
> > > 
> > > * Platform:         linux; (Linux-6.8.0-59-generic-x86_64-with-glibc2.39)
> > > * Python version:   3.12.3 (CPython)
> > > * Sphinx version:   8.2.3
> > > * Docutils version: 0.21.2
> > > * Jinja2 version:   3.1.6
> > > * Pygments version: 2.19.1
> > > 
> > > Last Messages
> > > =============
> > > 
> > >     userspace-api/gpio/gpio-get-chipinfo-ioctl .. userspace-api/media/dvb/dmx-fclose
> > > 
> > > 
> > >     reading sources... [ 90%]
> > >     userspace-api/media/dvb/dmx-fopen .. userspace-api/media/mediactl/media-controller-model
> > > 
> > > 
> > >     reading sources... [ 92%]
> > >     userspace-api/media/mediactl/media-func-close .. userspace-api/media/v4l/diff-v4l
> > > 
> > > Loaded Extensions
> > > =================
> > > 
> > > * sphinx.ext.mathjax (8.2.3)
> > > * alabaster (1.0.0)
> > > * sphinxcontrib.applehelp (2.0.0)
> > > * sphinxcontrib.devhelp (2.0.0)
> > > * sphinxcontrib.htmlhelp (2.1.0)
> > > * sphinxcontrib.serializinghtml (2.0.0)
> > > * sphinxcontrib.qthelp (2.0.0)
> > > * kerneldoc (1.0)
> > > * rstFlatTable (1.0)
> > > * kernel_include (1.0)
> > > * kfigure (1.0.0)
> > > * sphinx.ext.ifconfig (8.2.3)
> > > * automarkup (unknown version)
> > > * maintainers_include (1.0)
> > > * sphinx.ext.autosectionlabel (8.2.3)
> > > * kernel_abi (1.0)
> > > * kernel_feat (1.0)
> > > * translations (unknown version)
> > > * sphinx.ext.imgmath (8.2.3)
> > > 
> > > Traceback
> > > =========
> > > 
> > >       File "/<...>/sphinx-8.2.3/lib/python3.12/site-packages/sphinx/util/parallel.py", line 137, in _join_one
> > >         raise SphinxParallelError(*result)
> > >     sphinx.errors.SphinxParallelError: KeyError: '/<...>/lib/bitmap-bad.c'
> > > 
> > > 
> > > The full traceback has been saved in:
> > > /tmp/sphinx-err-8jzxndsr.log
> > > 
> > > To report this error to the developers, please open an issue at <https://github.com/sphinx-doc/sphinx/issues/>. Thanks!
> > > Please also report this if it was a user error, so that a better error message can be provided next time.
> > > make[3]: *** [/<...>/Documentation/Makefile:123: htmldocs] Error 2
> > > make[2]: *** [/<...>/Makefile:1806: htmldocs] Error 2
> > > make[1]: *** [/<...>/Makefile:248: __sub-make] Error 2
> > > make[1]: Leaving directory '/<...>/my-output'
> > > make: *** [Makefile:248: __sub-make] Error 2
> > > 
> > > -----------------------------------------------------------------
> > > 
> > > This would surprise innocent devs who are kindly willing to test docs build.
> > > 
> > > I think you need to tame its behavior and make it emit a proper warning and
> > > continue building docs in case of such predictable user errors.    
> > 
> > Totally agree.  
> 
> I also agree. 
> 
> The main difference between calling kernel-doc via a shell script or
> via a Python class is that now errors flow via Sphinx logger class,
> so they are subject to Sphinx filtering rules.
> 
> I double-checked: the logs are produced, and you can see them with "V=1",
> but Sphinx is hiding them, perhaps because of some options passed through
> sphinx-build call, or because they require them to have certain types.
> 
> A quick workaround would be to not use Sphinx logger anymore (see
> enclosed). It has a side effect, though: we lose control of setting
> it via V= variable, which is not good.

Heh, V=1 is not actually affected by not using Sphinx logger inside
the class. So, I can't see a side effect of letting the kernel-doc
use directly Python logger instead of the Sphinx variant.

So, I submitted the fix at:
	https://lore.kernel.org/linux-doc/cover.1747719873.git.mchehab+huawei@kernel.org/T/#t

As we may want to revisit it later in the future, in case Sphinx makes
something more fancy there, I added a comment at the patch.

Thanks,
Mauro