[PATCH 02/10] compiler.h: add is_const() as a replacement of __is_constexpr()

Vincent Mailhol vincent.mailhol at gmail.com
Sat Dec 7 15:23:24 UTC 2024


On Sun. 8 Dec. 2024 at 00:10, Martin Uecker <muecker at gwdg.de> wrote:
> Am Samstag, dem 07.12.2024 um 22:50 +0900 schrieb Vincent Mailhol:
> > On Sat. 7 Dec. 2024 à 22:19, Martin Uecker <muecker at gwdg.de> wrote:
> > >
> ...
>
> >
> > I was invited to WG14 this September. For now, I am only lurking. The
> > thing I have in mind right now is to write a paper to allow the use of
> > static_assert() in expressions (i.e. make it return 0 on success).
> > That should be a relatively small change, but would bring a nice
> > quality of life improvement.
> >
> > For context, look at this:
> >
> >   https://lore.kernel.org/all/CAHk-=wjLSEcZ5LdW+3C+9rtjvNPHZT6zdk0POj67T5k2ZpDbgA@mail.gmail.com/T/#m1ba33a804b4041154b72a1d0333f90ec7204c461
>
> What one can do is put it into a structure.
>
> #define const_assert(x) \
>     (sizeof(struct { _Static_assert(x, ""); }))
>
> but yeah, also a hack to work around a limitation of the standard
> feature.

If you scroll down a couple more messages, you can see that Linus came
up with that exact same hack :-)

It is now upstreamed in:

  https://git.kernel.org/torvalds/c/d7a516c6eeae

And yes, this solves the problem for the kernel, but I would still
like to change the standard to solve it for everyone else.


Yours sincerely,
Vincent Mailhol


More information about the Intel-gfx mailing list