Re: [PATCH v4 2/4] iio: light: add support for veml6031x00 ALS series

["Javier Carrasco" <javier.carrasco.cruz@gmail.com>]

On Tue Jun 2, 2026 at 2:33 PM CEST, Joshua Crofts wrote:
> On Tue, 2 Jun 2026 at 14:22, Javier Carrasco
> <javier.carrasco.cruz@gmail.com> wrote:
>>
>> On Tue Jun 2, 2026 at 1:47 PM CEST, Joshua Crofts wrote:
>> > On Tue, 2 Jun 2026 at 13:35, Javier Carrasco
>> > <javier.carrasco.cruz@gmail.com> wrote:
>> >> Thank you, Joshua. One question, though: maybe you remember that your
>> >> last driver was missing one header (which I am not criticizing at all,
>> >> as you can see I missed some too!), and I am wondering if it was because
>> >> iwyu missed it and if so, how that could be avoided.
>> >
>> > About that - I didn't actually run iwyu on my driver, I recently migrated to
>> > Fedora and forgot to set up LLVM and iywu-tool :-)
>>
>> I installed IWYU and added your mapping for a quick test, and it wants
>> to add a million headers or so:
>> If that is correct, I bet there is almost no driver prior to IWYU that
>> has all the headers :D
>
> Well, as Andy said - it's still far from ideal. It generates unnecessary noise.
>
> Question: did you run `make compile_commands.json` before running iwyu?
>
> iwyu_tool -p compile_commands.json drivers/iio/magnetometer/ak8975.c -- \
> -Xiwyu --no_default_mappings -Xiwyu --mapping_file=/home/josh/iio.imp
>
> This was the command I used when I did some cleaning up on the ak8975.

I used the compile_commands.json generated by
scripts/clang-tools/gen_compile_commands.py, which I always use to have
a better experience with clangd and Neovim.

Should it make any difference?