On Wed, Nov 07, 2018 at 01:40:11PM +0000, Joseph Myers wrote:
> On Wed, 7 Nov 2018, Jakub Jelinek wrote:
>
> > Don't know about the dropping of HAVE_FENV_H/USE_FENV_H stuff, don't we
> > support libquadmath on targets that don't have fenv.h?
> > In other sources, like e.g. expq.c, the USE_FENV_H guards are still kept.
>
> All those conditionals are now meant to be handled through quadmath-imp.h,
> which defines various functions as macros in the no-fenv.h case (macros
> that ignore their arguments, where there might be issues with calls using
> FE_* macros that themselves aren't defined).
>
> I haven't tested any systems without fenv.h, but any issues in that regard
> are intended to be addressed through further quadmath-imp.h, to avoid
> needing to insert conditionals in individual source files in an automated
> way.
Ah, ok then.
Jakub
