On Thu, Sep 04, 2025 at 02:24:57AM +0100, Adriano Carvalho wrote: > On Wed, 3 Sept 2025 at 02:14, Adriano Carvalho > <[email protected]> wrote: > > [...] > > So, in principle, we "only" need to replace all ".. image" with ".. > > kernel-image" (65 occurrences / 17 files). However, I did that, and > > got: > > > > sphinx.errors.ExtensionError: Handler <bound method > > ImageCollector.process_doc of > > <sphinx.environment.collectors.asset.ImageCollector object at > > 0x79380051bbe0>> for event 'doctree-read' threw an exception > > (exception: 'uri') > > [...] > > Actually, the right approach is to use ".. kernel-figure" not ".. > kernel-image". The latter is never used in the kernel. > > Using ".. kernel-figure" it works fine. The remaining patches are > still required though. > > So, if you intend to remain aligned with Linux in the way the > Documentation is built, instead of this patch a new one is required > that replaces all instances of ".. image" with ".. kernel-image" (65 > ccurrences / 17 files). > > Please let me know how you'd like me to proceed.
We should switch to "kernel-image" and update our checkpatch.pl to complain (and this might be acceptable to the linux kernel too, are there instances of ".. image" ? I assume they too should be kernel-image) so that we don't regress in easy to catch ways. And then yes, getting the correct non-zero SVG for the TI example needs to be done. > Moreover, the infrastructure to build the Documentation is a bit > behind that of Linux. Not a lot but quite a few changes. Should I try > to issue a patch series to place U-Boot as in sync as possible with > Linux? This is a question for Heinrich. Thanks again for digging here! -- Tom
signature.asc
Description: PGP signature
