[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Make ppc64 secondary arch - don't block builds (was: Dealing with ppc64 BRs)

On Mon, Aug 06, 2007 at 11:48:34AM -0400, Tom Lane wrote:
> Jesse Keating <jkeating redhat com> writes:
> > But we're not talking about a secondary arch, we're talking about a
> > primary arch.  Like it or not, ppc (and thus ppc64) is a primary arch
> > set and must addressed just like the other primary arches.
> I really find it troublesome that anyone thinks they shouldn't be
> primary.  If the only primary arches are x86/x86_64, then packagers will
> basically not have any forcing function to make them worry about whether
> the code is portable to any non-Intel platform.

So by forcing packagers to rebuilding with ExclusiveArchs and users to
redownload artificially bumped packages w/o any content changes you
make anyonw worry about ppc support? I for one felt like this is a big
PITA and was not motivated at all to find the ppc specific issues.

> I think that at minimum we need a bigendian arch or two in the
> primary set, just so that there's at least a token requirement for
> portability.

Why not add some 8bit arch to ensure even more portability? Or 31bits
to make it at least useful?

Let's worry about portability when the arch we want to port to appears
and not use some archs to *practice* portability.

> Else the secondary arches are *all* doomed to failure in the long
> run.

Axel.Thimm at ATrpms.net

Attachment: pgpHV5Gma8MyS.pgp
Description: PGP signature

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]