ppc64 builds

[Michael Schwendt]

On Sun, 18 Mar 2007 18:43:18 +0000, David Woodhouse wrote:

> On Sun, 2007-03-18 at 11:30 -0400, Jesse Keating wrote:
> > On Sunday 18 March 2007 08:05:37 David Woodhouse wrote:
> > > We've never bothered shipping 64-bit versions of Extras packages before
> > > -- unless you suddenly find an overriding reason to do so, I don't see
> > > any reason to rebuild for F7 just to add a 64-bit binary package which
> > > we don't need to ship anyway.
> > 
> > By merging all the packages into one big collection we can't 
> > segregate "Extras" and "Core" anymore for decisions such as build for ppc64 
> > or not.  Every package will build for every arch unless explicitly told not 
> > to, and if told, there is supposed to be bug regarding this according to our 
> > guidelines (which you wanted IIRC).  This means we need to turn on ppc64 in 
> > the new build system to keep the current "Core" packages building there, and 
> > we need to bootstrap the rest of the packages so that they can start building 
> > ppc64 without causing failures all over the place. 
> 
> This is true. I was just suggesting that we don't necessarily need to
> rush out and do a mass rebuild of all Extras packages before F7 just to
> create ppc64 versions of them, since those _wouldn't_ be likely to end
> up in the "ppc" compose; they'd only be in the pure ppc64 tree which
> isn't a product we release; it's just the same as the unshipped ia64,
> s390 rawhide trees.

 From Push.py:

        # rpmUtils.arch.getBaseArch(a)
        elif a in ['i386', 'i486', 'i586', 'i686', 'athlon']:
            basearch = 'i386'
        
        elif a in ['x86_64', 'ia32e', 'amd64']:
            basearch = 'x86_64'
        
        elif a in ['ppc', 'ppc64', 'ppc32']:
            basearch = 'ppc'

        else:
            print 'Unknown arch %s' % a
            continue  # with next package