[PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[Akira Yokosawa]

The option list to sphinx-build via SPHINXOPTS should have higher
priority than those the wrapper comes up with.
sphinx-build will choose the latest one if there are duplicates.

To restore the behavior of Makefile era, when the documentation builds
at https://www.kernel.org/doc/html/next/ had been depending on it,
reorder the flag list.

Reported-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
Closes: https://lore.kernel.org/20251007-awesome-guan-of-greatness-e6ec75@lemur/
Reported-by: Akira Yokosawa <akiyks@gmail.com>
Closes: https://lore.kernel.org/c35e690f-0579-49cb-850c-07af98e5253a@gmail.com/
Signed-off-by: Akira Yokosawa <akiyks@gmail.com>
---
Hi,

Not having hearing from Mauro on this _minor_ regression, I went
forward and made this fix.

Konstantin, could you test this against your doc build environment
for linux-next?

Regards,
Akira
--
 tools/docs/sphinx-build-wrapper | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/docs/sphinx-build-wrapper b/tools/docs/sphinx-build-wrapper
index 3e6d166d4102..1efaca3d16aa 100755
--- a/tools/docs/sphinx-build-wrapper
+++ b/tools/docs/sphinx-build-wrapper
@@ -298,8 +298,8 @@ class SphinxBuilder:
 
             cmd += [sphinx_build]
             cmd += [f"-j{n_jobs}"]
-            cmd += self.sphinxopts
             cmd += build_args
+            cmd += self.sphinxopts
 
             if self.verbose:
                 print(" ".join(cmd))

base-commit: 96b546c241b11a97ba1247580208c554458e7866
-- 
2.43.0

Re: [PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[Akira Yokosawa]

On Sun, 19 Oct 2025 23:24:51 +0900, Akira Yokosawa wrote:
> The option list to sphinx-build via SPHINXOPTS should have higher
> priority than those the wrapper comes up with.
> sphinx-build will choose the latest one if there are duplicates.
> 
> To restore the behavior of Makefile era, when the documentation builds
> at https://www.kernel.org/doc/html/next/ had been depending on it,
> reorder the flag list.
> 
> Reported-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
> Closes: https://lore.kernel.org/20251007-awesome-guan-of-greatness-e6ec75@lemur/
> Reported-by: Akira Yokosawa <akiyks@gmail.com>
> Closes: https://lore.kernel.org/c35e690f-0579-49cb-850c-07af98e5253a@gmail.com/
> Signed-off-by: Akira Yokosawa <akiyks@gmail.com>
> ---
> Hi,
> 
> Not having hearing from Mauro on this _minor_ regression, I went
> forward and made this fix.
> 
> Konstantin, could you test this against your doc build environment
> for linux-next?

Jon, could you consider applying this fix and making it be tested
directly at https://www.kernel.org/doc/html/next/ ?

The custom "next-2025mmdd" tag there would be helpful for us to catch
regressions in linux-next docs builds.

        Thanks, Akira

> 
> Regards,
> Akira
> --
>  tools/docs/sphinx-build-wrapper | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tools/docs/sphinx-build-wrapper b/tools/docs/sphinx-build-wrapper
> index 3e6d166d4102..1efaca3d16aa 100755
> --- a/tools/docs/sphinx-build-wrapper
> +++ b/tools/docs/sphinx-build-wrapper
> @@ -298,8 +298,8 @@ class SphinxBuilder:
>  
>              cmd += [sphinx_build]
>              cmd += [f"-j{n_jobs}"]
> -            cmd += self.sphinxopts
>              cmd += build_args
> +            cmd += self.sphinxopts
>  
>              if self.verbose:
>                  print(" ".join(cmd))
> 
> base-commit: 96b546c241b11a97ba1247580208c554458e7866

Re: [PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[Jonathan Corbet]

Akira Yokosawa <akiyks@gmail.com> writes:

> On Sun, 19 Oct 2025 23:24:51 +0900, Akira Yokosawa wrote:
>> The option list to sphinx-build via SPHINXOPTS should have higher
>> priority than those the wrapper comes up with.
>> sphinx-build will choose the latest one if there are duplicates.
>> 
>> To restore the behavior of Makefile era, when the documentation builds
>> at https://www.kernel.org/doc/html/next/ had been depending on it,
>> reorder the flag list.
>> 
>> Reported-by: Konstantin Ryabitsev <konstantin@linuxfoundation.org>
>> Closes: https://lore.kernel.org/20251007-awesome-guan-of-greatness-e6ec75@lemur/
>> Reported-by: Akira Yokosawa <akiyks@gmail.com>
>> Closes: https://lore.kernel.org/c35e690f-0579-49cb-850c-07af98e5253a@gmail.com/
>> Signed-off-by: Akira Yokosawa <akiyks@gmail.com>
>> ---
>> Hi,
>> 
>> Not having hearing from Mauro on this _minor_ regression, I went
>> forward and made this fix.
>> 
>> Konstantin, could you test this against your doc build environment
>> for linux-next?
>
> Jon, could you consider applying this fix and making it be tested
> directly at https://www.kernel.org/doc/html/next/ ?
>
> The custom "next-2025mmdd" tag there would be helpful for us to catch
> regressions in linux-next docs builds.

It doesn't seem like there should be any regressions ... I've applied
it, thanks.

jon

Re: [PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[Akira Yokosawa]

[-CC: Mauro and Randy]

Hi,

On Wed, 29 Oct 2025 09:52:30 -0600, Jonathan Corbet wrote:
> Akira Yokosawa <akiyks@gmail.com> writes:
[...]

>> Jon, could you consider applying this fix and making it be tested
>> directly at https://www.kernel.org/doc/html/next/ ?
>>
>> The custom "next-2025mmdd" tag there would be helpful for us to catch
>> regressions in linux-next docs builds.
> 
> It doesn't seem like there should be any regressions ... I've applied
> it, thanks.

Today's build at https://www.kernel.org/doc/html/next/ still has
"6.18.0-rc3" as the version tag.  Other new changes in docs-next are
reflected there.

Konstantin, I'm not sure but have you dropped the

    SPHINXOPTS="-D version=${TAGVER} -q"

flag from the command line of next docs build?

If that's the case, can you restore it?

Thanks, Akira

Re: [PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[Akira Yokosawa]

On Thu, 30 Oct 2025 18:55:04 +0900, Akira Yokosawa wrote:
> [-CC: Mauro and Randy]
> 
> Hi,
> 
> On Wed, 29 Oct 2025 09:52:30 -0600, Jonathan Corbet wrote:
>> Akira Yokosawa <akiyks@gmail.com> writes:
> [...]
> 
>>> Jon, could you consider applying this fix and making it be tested
>>> directly at https://www.kernel.org/doc/html/next/ ?
>>>
>>> The custom "next-2025mmdd" tag there would be helpful for us to catch
>>> regressions in linux-next docs builds.
>>
>> It doesn't seem like there should be any regressions ... I've applied
>> it, thanks.
> 
> Today's build at https://www.kernel.org/doc/html/next/ still has
> "6.18.0-rc3" as the version tag.  Other new changes in docs-next are
> reflected there.
> 

Now, the version tag reads "next-20251030".

Nice to see it working again!

Thanks, Akira

Re: [PATCH] tools/docs/sphinx-build-wrapper: Emit $SPHINXOPTS later in args list

[Konstantin Ryabitsev]

On Fri, Oct 31, 2025 at 10:25:56AM +0900, Akira Yokosawa wrote:
> > Today's build at https://www.kernel.org/doc/html/next/ still has
> > "6.18.0-rc3" as the version tag.  Other new changes in docs-next are
> > reflected there.
> > 
> 
> Now, the version tag reads "next-20251030".

Yes, this is now working properly, thanks everyone!

-K