On Thu, Feb 1, 2024 at 8:58 AM Justin Forbes <jforbes@xxxxxxxxxxxxxxxxx> wrote:
>
> On Thu, Feb 1, 2024 at 8:41 AM Justin Forbes <jforbes@xxxxxxxxxxxxxxxxx> wrote:
> >
> > On Thu, Feb 1, 2024 at 8:25 AM Greg Kroah-Hartman
> > <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > >
> > > On Thu, Feb 01, 2024 at 06:43:46AM -0600, Justin Forbes wrote:
> > > > On Tue, Jan 30, 2024 at 10:21 AM Jonathan Corbet <corbet@xxxxxxx> wrote:
> > > > >
> > > > > Justin Forbes <jforbes@xxxxxxxxxxxxxxxxx> writes:
> > > > >
> > > > > > On Mon, Jan 29, 2024 at 09:01:07AM -0800, Greg Kroah-Hartman wrote:
> > > > > >> 6.6-stable review patch. If anyone has any objections, please let me know.
> > > > > >>
> > > > > >> ------------------
> > > > > >>
> > > > > >> From: Vegard Nossum <vegard.nossum@xxxxxxxxxx>
> > > > > >>
> > > > > >> [ Upstream commit c48a7c44a1d02516309015b6134c9bb982e17008 ]
> > > > > >>
> > > > > >> The kernel-feat directive passes its argument straight to the shell.
> > > > > >> This is unfortunate and unnecessary.
> > > > > >>
> > > > > >> Let's always use paths relative to $srctree/Documentation/ and use
> > > > > >> subprocess.check_call() instead of subprocess.Popen(shell=True).
> > > > > >>
> > > > > >> This also makes the code shorter.
> > > > > >>
> > > > > >> This is analogous to commit 3231dd586277 ("docs: kernel_abi.py: fix
> > > > > >> command injection") where we did exactly the same thing for
> > > > > >> kernel_abi.py, somehow I completely missed this one.
> > > > > >>
> > > > > >> Link: https://fosstodon.org/@jani/111676532203641247
> > > > > >> Reported-by: Jani Nikula <jani.nikula@xxxxxxxxx>
> > > > > >> Signed-off-by: Vegard Nossum <vegard.nossum@xxxxxxxxxx>
> > > > > >> Cc: stable@xxxxxxxxxxxxxxx
> > > > > >> Signed-off-by: Jonathan Corbet <corbet@xxxxxxx>
> > > > > >> Link: https://lore.kernel.org/r/20240110174758.3680506-1-vegard.nossum@xxxxxxxxxx
> > > > > >> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
> > > > > >
> > > > > > This patch seems to be missing something. In 6.6.15-rc1 I get a doc
> > > > > > build failure with:
> > > > > >
> > > > > > /builddir/build/BUILD/kernel-6.6.14-332-g1ff49073b88b/linux-6.6.15-0.rc1.1ff49073b88b.200.fc39.noarch/Documentation/sphinx/kerneldoc.py:133: SyntaxWarning: invalid escape sequence '\.'
> > > > > > line_regex = re.compile("^\.\. LINENO ([0-9]+)$")
> > > > >
> > > > > Ah ... you're missing 86a0adc029d3 (Documentation/sphinx: fix Python
> > > > > string escapes). That is not a problem with this patch, though; I would
> > > > > expect you to get the same error (with Python 3.12) without.
> > > >
> > > > Well, it appears that 6.6.15 shipped anyway, with this patch included,
> > > > but not with 86a0adc029d3. If anyone else builds docs, this thread
> > > > should at least show them the fix. Perhaps we can get the missing
> > > > patch into 6.6.16?
> > >
> > > Sure, but again, that should be independent of this change, right?
> >
> > I am not sure I would say independent. This particular change causes
> > docs to fail the build as I mentioned during rc1. There were no
> > issues building 6.6.14 or previous releases, and no problem building
> > 6.7.3.
>
> I can confirm that adding this patch to 6.6.15 makes docs build again.
I lied, it just fails slightly differently. Some of the noise is gone,
but we still have:
Sphinx parallel build error:
UnboundLocalError: cannot access local variable 'fname' where it is
not associated with a value
make[2]: *** [Documentation/Makefile:102: htmldocs] Error 2
make[1]: *** [/builddir/build/BUILD/kernel-6.6.15/linux-6.6.15-200.fc39.noarch/Makefile:1715:
htmldocs] Error 2
> Justin
>
> > Justin
> >
> > > thanks,
> > >
> > > greg k-h
> > >
> > >
> > > >
> > > > Jusitn
> > > >
> > > > > Thanks,
> > > > >
> > > > > jon
> > > > >
> > > >
> > >
