Re: [PATCH] rust: id_pool: fix example

From: Alice Ryhl

Date: Tue Dec 02 2025 - 07:23:19 EST


On Tue, Dec 02, 2025 at 12:13:33AM +0100, Miguel Ojeda wrote:
> On Mon, Dec 1, 2025 at 8:29 PM Yury Norov <yury.norov@xxxxxxxxx> wrote:
> >
> > Thanks Miguel,
> >
> > I applied this, but the fact that you've sent a second fix to
> > documentation that actually is a build fix, raises the questions.
> >
> > Because Rust documentation bears compilable chunks of code, I think
> > we need to enable rustdoc tests target by default, so that developers
> > will not send broken tests.
>
> You're welcome!
>
> This one is a Kconfig in the normal build, i.e.
> `CONFIG_RUST_KERNEL_DOCTESTS` (it is true that the other patch was for
> a different target).
>
> If you mean enabling that by Kconfig default, then I don't think we
> can do that -- KUnit on its own (which this uses) is not meant for
> normal builds. Perhaps we could have something that just checks the
> build, but people should really run the doctests anyway, since they
> typically contain `assert!`s and so on that can be wrong.
>
> If you mean enabling it in Intel's kernel test robot, then I think
> they do it (or at least I told them about it long ago). But maybe
> something is missing.
>
> In our P entry in `MAINTAINERS` ("Subsystem Profile document") we ask
> people to run them among other things, so I would perhaps suggest
> having something like that too (or linking to ours if you prefer):
>
> https://rust-for-linux.com/contributing#submit-checklist-addendum
>
> Of course new contributors will miss that initially, but actually
> people find the doctests quite useful, so generally people get to run
> them. At the end of the day, maintainers should test these too before
> applying and otherwise someone will notice sooner or later.

Yeah sorry I forgot to check the docs this time. I usually do run them,
but there are so many targets to check that sometimes I forget some of
them.

Alice