Make ppc64 secondary arch - don't block builds

Oliver Falk oliver at linux-kernel.at
Mon Aug 6 17:33:30 UTC 2007


Jesse Keating schrieb:
> On Mon, 6 Aug 2007 17:16:06 +0100
> Andrew Haley <aph at redhat.com> wrote:
> 
>> Bugs found when chasing down portability problems are often real bugs,
>> not just non-portable code.  For that reason, 100% of our users
>> benefit from portable code, not just the 1% who use minor
>> architectures.  It's in everyone's interest that Fedora code is real C
>> (or C++).
> 
> But when that bug doesn't effect the majority of our user base, do
> those users really care?  Should the?  Yes.  Do they?  No.  "It works
> for me and my major arch, why should I care that it broke on your silly
> arch?"

If we think this way... Well then we must stop the pedantic check of
open from latest glibc (I think it's already disabled however.)... Seems
100% of our userbase could have lived without that check... And it also
seems that *many* of our packagers could as well - they now need to
write patches and send 'em upstream. I don't want to imagine what some
upstream maintainers will say.....

We then could also ignore any kind of -Werror :-P

OK, /me comes back from sarkasm. Jesse, we had this topic a few times
already - on this list.

I can understand if someone says: Ignore it, it's ARM or ignore it, it's
Alpha (well, as Alpha-maintainer I don't really *want* to hear that).
But for ppc - I guess there are already quite a lot people out there
using Fedora on ppc(64). Isn't it?

For the "blocking builds" topic. Well, if it doesn't work on ppc64 - the
only primary-non-intel-64bit arch in Fedora (as far as I know), we
should really block it and have a look... If it's an security-update,
push it ASAP... If not, try to fix it or find someone who can fix it.
There are usually people out there willing to help.

-of




More information about the Fedora-maintainers mailing list