Jonathan Corbet wrote:
> Akira Yokosawa <akiyks@xxxxxxxxx> writes:
>
>> Jonathan Corbet wrote:
>> [...]
>>> Ah ... I should have thought of that ... I'm glad you did. I've just
>>> pushed the fix out to linux-next, will send it Linusward in the near
>>> future. Thanks for fixing this,
>>
>> Your docs-fixes is based on v6.13-rc1.
>> This fix needs to come after v6.13-rc2.
>
> ...which will happen once the fix hits mainline - the *fix* doesn't
> depend on -rc2.
>
Well...
The fix conflicts semantically against v6.13-rc1.
I know you hate full rebuild of htmldocs. But I'd like you to follow
the following steps to see the full scope of this fix.
- git checkout v6.13-rc2
- git checkout docs-fixes
- make htmldocs
You might be surprised, but you'll see the following:
./drivers/dma-buf/dma-buf.c:1: warning: no structured comments found
./drivers/iommu/iommufd/device.c:1: warning: no structured comments found
./drivers/iommu/iommufd/main.c:1: warning: no structured comments found
./drivers/counter/counter-core.c:1: warning: no structured comments found
./drivers/counter/counter-chrdev.c:1: warning: no structured comments found
Current docs-fixes has a semantic conflicts opposite to the one v6.13-rc2 has.
I'd really like you to send a pull request that resolves the semantic
conflict existing in v6.13-rc2, not the one that happens to negate the
conflict on merge.
That's the minimal expectation I have.
You might wonder whey you need to checkout v6.13-rc2 first and go back
to docs-fixes to see the conflict there.
That's because "make htmldocs" don't rerun kernel-doc conversion
when only the kernel-doc script is updated. To see the effect of
its change, source files with affected kernel-doc comments and
"EXPORT_SYMBOL*" need to be updated as well.
Of course you are well aware of this problematic behavior of
documentation build.
I don't mind docs-mw if you'd like to keep it v6.13-rc1 based.
Have I made my points clear enough for you?
Thanks, Akira
