https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124547
--- Comment #13 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot
Uni-Bielefeld.DE> ---
> --- Comment #11 from Richard Biener <rguenth at gcc dot gnu.org> ---
> And to amend the above - config.gcc should end with gas/gnu_ld that we
> assume unless overridden by flags provided to configure. For the
> solaris case that might mean they are empty, aka unknown, or we still
> assume they are used and fixup during configure.
The Solaris case is easy and already dealt with: we explicitly check for
either Solaris as/ld or GNU as/ld on Solaris, setting
solaris_as/solaris_ld and gas_flag/gnu_ld_flag accordingly.
We could make this even more robust by assuming that all
assemblers/linker that are not known to be Solaris as/ld are gas/gld
compatible. Maybe there are nascent ports of lld or wild that would
benefit.
As a follow-up, we should check solaris_as/solaris_ld in cases where
that's the actual issue, not !gas_flag or !gnu_ld_flag.
This was included in the original versions of the
Switch to HAVE_SOLARIS_AS
https://gcc.gnu.org/pipermail/gcc-patches/2026-February/708352.html
and
Switch to HAVE_SOLARIS_LD
https://gcc.gnu.org/pipermail/gcc-patches/2026-February/708354.html
patches, but split off for GCC 17 stage 1 considering all the breakage
I've already caused ;-(
