request a rebuild to pick up Pygments 2.21.0

request a rebuild to pick up Pygments 2.21.0

[Trevor Woerner]

Hi,

Could someone build buildtools-docs against current master and publish
the result?

scripts/run-docs-build line 29 pins the docs toolchain to the
yocto-6.0_M2 tarball, which ships Pygments 2.19.2. The newest
published milestone, yocto-6.1_M2, has 2.20.0. OE-Core master carries
python3-pygments_2.21.0.bb.

2.21.0 is the release that added a BitBake lexer, so it is the first
version in which

	.. code-block:: bitbake

does anything at all. I have a conversion of the BitBake snippets
in yocto-docs and the BitBake user manual ready to send, but on the
current toolchain every block would publish unhighlighted - and since
run-docs-build drops the Makefile's -W, silently.

I've tested the change. run-docs-build feeds every BitBake docs branch
through $builddir/buildtools unconditionally, and yocto-docs from
scarthgap forward, so I built all of them twice - once with 2.19.2,
once with 2.21.0 - applying the same docs-build-patches and master-file
checkouts the script does:

	bitbake     1.46 through 2.18, and master          14/14 identical
	yocto-docs  scarthgap styhead walnascar whinlatter   4/4 identical

36 builds, same exit status, same warning count - zero warnings
throughout.

Best regards,
	Trevor

Re: request a rebuild to pick up Pygments 2.21.0

[Richard Purdie]

On Mon, 2026-08-24 at 18:49 -0400, Trevor Woerner wrote:
> Could someone build buildtools-docs against current master and publish
> the result?
> 
> scripts/run-docs-build line 29 pins the docs toolchain to the
> yocto-6.0_M2 tarball, which ships Pygments 2.19.2. The newest
> published milestone, yocto-6.1_M2, has 2.20.0. OE-Core master carries
> python3-pygments_2.21.0.bb.
> 
> 2.21.0 is the release that added a BitBake lexer, so it is the first
> version in which
> 
> 	.. code-block:: bitbake
> 
> does anything at all. I have a conversion of the BitBake snippets
> in yocto-docs and the BitBake user manual ready to send, but on the
> current toolchain every block would publish unhighlighted - and since
> run-docs-build drops the Makefile's -W, silently.
> 
> I've tested the change. run-docs-build feeds every BitBake docs branch
> through $builddir/buildtools unconditionally, and yocto-docs from
> scarthgap forward, so I built all of them twice - once with 2.19.2,
> once with 2.21.0 - applying the same docs-build-patches and master-file
> checkouts the script does:
> 
> 	bitbake     1.46 through 2.18, and master          14/14 identical
> 	yocto-docs  scarthgap styhead walnascar whinlatter   4/4 identical
> 
> 36 builds, same exit status, same warning count - zero warnings
> throughout.

How urgent is this? Could we do this as part of M3 and use the artefact
from that release build?

Cheers,

Richard

Re: request a rebuild to pick up Pygments 2.21.0

[Antonin Godard]

Hi,

On Tue Aug 25, 2026 at 8:36 AM CEST, Richard Purdie wrote:
> On Mon, 2026-08-24 at 18:49 -0400, Trevor Woerner wrote:
>> Could someone build buildtools-docs against current master and publish
>> the result?
>> 
>> scripts/run-docs-build line 29 pins the docs toolchain to the
>> yocto-6.0_M2 tarball, which ships Pygments 2.19.2. The newest
>> published milestone, yocto-6.1_M2, has 2.20.0. OE-Core master carries
>> python3-pygments_2.21.0.bb.
>> 
>> 2.21.0 is the release that added a BitBake lexer, so it is the first
>> version in which
>> 
>> 	.. code-block:: bitbake
>> 
>> does anything at all. I have a conversion of the BitBake snippets
>> in yocto-docs and the BitBake user manual ready to send, but on the
>> current toolchain every block would publish unhighlighted - and since
>> run-docs-build drops the Makefile's -W, silently.
>> 
>> I've tested the change. run-docs-build feeds every BitBake docs branch
>> through $builddir/buildtools unconditionally, and yocto-docs from
>> scarthgap forward, so I built all of them twice - once with 2.19.2,
>> once with 2.21.0 - applying the same docs-build-patches and master-file
>> checkouts the script does:
>> 
>> 	bitbake     1.46 through 2.18, and master          14/14 identical
>> 	yocto-docs  scarthgap styhead walnascar whinlatter   4/4 identical
>> 
>> 36 builds, same exit status, same warning count - zero warnings
>> throughout.
>
> How urgent is this? Could we do this as part of M3 and use the artefact
> from that release build?

I was planning on doing that.

I have a local copy and build setup that has an updated buildtools tarball with
the new version of Pygments. Trevor, if you want to send another version of your
patch, I can review and test it locally.

Antonin

Re: request a rebuild to pick up Pygments 2.21.0

[Trevor Woerner]

On Tue 2026-08-25 @ 07:36:48 AM, Richard Purdie wrote:
> On Mon, 2026-08-24 at 18:49 -0400, Trevor Woerner wrote:
> > Could someone build buildtools-docs against current master and publish
> > the result?
> > 
> > scripts/run-docs-build line 29 pins the docs toolchain to the
> > yocto-6.0_M2 tarball, which ships Pygments 2.19.2. The newest
> > published milestone, yocto-6.1_M2, has 2.20.0. OE-Core master carries
> > python3-pygments_2.21.0.bb.
> > 
> > 2.21.0 is the release that added a BitBake lexer, so it is the first
> > version in which
> > 
> > 	.. code-block:: bitbake
> > 
> > does anything at all. I have a conversion of the BitBake snippets
> > in yocto-docs and the BitBake user manual ready to send, but on the
> > current toolchain every block would publish unhighlighted - and since
> > run-docs-build drops the Makefile's -W, silently.
> > 
> > I've tested the change. run-docs-build feeds every BitBake docs branch
> > through $builddir/buildtools unconditionally, and yocto-docs from
> > scarthgap forward, so I built all of them twice - once with 2.19.2,
> > once with 2.21.0 - applying the same docs-build-patches and master-file
> > checkouts the script does:
> > 
> > 	bitbake     1.46 through 2.18, and master          14/14 identical
> > 	yocto-docs  scarthgap styhead walnascar whinlatter   4/4 identical
> > 
> > 36 builds, same exit status, same warning count - zero warnings
> > throughout.
> 
> How urgent is this? Could we do this as part of M3 and use the artefact
> from that release build?

Not urgent at all. Adding a bitbake lexer will just make the docs
prettier.

Re: request a rebuild to pick up Pygments 2.21.0

[Trevor Woerner]

On Tue 2026-08-25 @ 09:07:41 AM, Antonin Godard wrote:
> Hi,
> 
> On Tue Aug 25, 2026 at 8:36 AM CEST, Richard Purdie wrote:
> > On Mon, 2026-08-24 at 18:49 -0400, Trevor Woerner wrote:
> >> Could someone build buildtools-docs against current master and publish
> >> the result?
> >> 
> >> scripts/run-docs-build line 29 pins the docs toolchain to the
> >> yocto-6.0_M2 tarball, which ships Pygments 2.19.2. The newest
> >> published milestone, yocto-6.1_M2, has 2.20.0. OE-Core master carries
> >> python3-pygments_2.21.0.bb.
> >> 
> >> 2.21.0 is the release that added a BitBake lexer, so it is the first
> >> version in which
> >> 
> >> 	.. code-block:: bitbake
> >> 
> >> does anything at all. I have a conversion of the BitBake snippets
> >> in yocto-docs and the BitBake user manual ready to send, but on the
> >> current toolchain every block would publish unhighlighted - and since
> >> run-docs-build drops the Makefile's -W, silently.
> >> 
> >> I've tested the change. run-docs-build feeds every BitBake docs branch
> >> through $builddir/buildtools unconditionally, and yocto-docs from
> >> scarthgap forward, so I built all of them twice - once with 2.19.2,
> >> once with 2.21.0 - applying the same docs-build-patches and master-file
> >> checkouts the script does:
> >> 
> >> 	bitbake     1.46 through 2.18, and master          14/14 identical
> >> 	yocto-docs  scarthgap styhead walnascar whinlatter   4/4 identical
> >> 
> >> 36 builds, same exit status, same warning count - zero warnings
> >> throughout.
> >
> > How urgent is this? Could we do this as part of M3 and use the artefact
> > from that release build?
> 
> I was planning on doing that.
> 
> I have a local copy and build setup that has an updated buildtools tarball with
> the new version of Pygments. Trevor, if you want to send another version of your
> patch, I can review and test it locally.

Sounds good, I'll start posting them to the list.