Pavel Zhukov <pzhukov(a)redhat.com> wrote:
As x86 (32-bit for i686->) is not primary architecture anymore
 and we
have at least one bug which blocks gprbuild  (and as the result
almost whole stack) I start to think about dropping i686 from the list
of GNAT_arches since F26 onward.
First, did you mean to write "F27"? Because I thought we had a working
GPRbuild in F26. Or is it necessary to rebuild it now, right after the
Have you reported this to GCC upstream? I don't think a report in Red
Hat Bugzilla accomplishes anything more than tracking the bug. A report
in GCC Bugzilla should reach some GNAT developers. We might not get a
speedy fix for F27 though.
If only GPRbuild triggers the bug, and no fix arrives before the next
soname change in libgnat, then we'll probably need to drop x86 from
GPRbuild_arches at least temporarily, but we could keep it in
GNAT_arches. If it turns out to be so bad that many things can't be
compiled on x86, then we may need to drop it from GNAT_arches too.