> Am 30.03.2026 um 15:43 schrieb Alexandre Oliva <[email protected]>:
>
> On Mar 29, 2026, Richard Biener <[email protected]> wrote:
>
>> It is not supposed to be used outside of vect.exp and it relies on
>> that passing -msse2 unconditionally on x86.
>
> It still seems to me that the documentation suggests it should return
> false if the target hardware under test doesn't support double vectors:
>
> @item vect_double
> Target supports hardware vectors of @code{double}.
>
> # Return 1 if the target supports hardware vectors of double, 0 otherwise.
> #
> # This won't change for different subtargets so cache the result.
>
> Should we amend any of that? Or is moving check_effective_target_vect_*
> to a separate file, or otherwise arranging for testsets other than
> vect.exp to not have access to them, enough to address the apparent (to
> me?) discrepancy?
I think we should move the subset of tests that clearly cannot work outside of
vect.exp setup and decide for the rest how/if to fix such issues. ‚Target
supports‘ is a blurry statement and we have multiple levels of support here
from compiling to assembling to running (which can require both OS and CPU
support).
Richard
> I'll try to keep that in mind for stage1, but no promises.
>
> --
> Alexandre Oliva, happy hacker https://blog.lx.oliva.nom.br/
> Free Software Activist FSFLA co-founder GNU Toolchain Engineer
> More tolerance and less prejudice are key for inclusion and diversity.
> Excluding neuro-others for not behaving ""normal"" is *not* inclusive!
