Re: [PATCH v2 0/7] KUnit integration for Rust doctests

[Miguel Ojeda]

On Tue, Jul 18, 2023 at 10:38 AM David Gow <davidgow@xxxxxxxxxx> wrote:
>
> I like "location" better, personally. The attributes work is still
> ongoing, and while there's some benefit to having "file" and "line"
> separate (it could potentially simplify some implementation on the C
> side), we'll cross that bridge when we come to it.

Yeah, I felt it looked a bit better, but if later on it ends up making
things too hard, then yeah, we can definitely simplify it.

> This seems to be working well on the existing cases under kunit.py
> here. I'll continue to play with it, but no worries on my end thus
> far.

Thanks for trying it out!

> Thanks: while we're still arguing a bit about exactly what the format
> of these will look like in the KUnit/KTAP attributes spec/patches,
> what you've used matches what we've been proposing so far.
>
> Let's stick with <test name>.location for now, and change it if needed
> when the attributes spec is finalised.

Sounds good.

> These are all (still) looking pretty good to me. If there are no
> objections, I'd like to take these into kselftest/kunit as-is and if
> we need to change anything (e.g. for consistency with attributes when
> they land), do that as a follow-up.
>
> Thanks again, Miguel, for all the work getting this going!

My pleasure -- and thanks for reviewing it so quickly and all your feedback!

Cheers,
Miguel