[fab] Architecture Policy.

[David Woodhouse]

On Sat, 2006-11-18 at 11:59 -0500, Christopher Blizzard wrote:
> Please tell us how we can help enable you and the rest of the PPC 
> community into better partners.  It was my hope that by enabling the PPC 
> folks to work with freedom to walk all over the tree as part of a 
> specific arch team that they would get more involved, not less. 

We _already_ do that -- and quite effectively too. We pick up with the
hairier arch-specific issues where individual package maintainers don't
have the skills required (like the Modula-3 port which was required for
cvsup), and we poke maintainers to fix it themselves when they (or their
package) are just being dim (like assuming char is signed). We provide
testing and/or access to machines as required.

I'm pleased with FC6 -- I think it's probably our best release on
PowerPC so far. As I said, when I managed to snatch a few moments for
testing FC6 in the run-up to the release (don't tell my boss), what I
found myself working on wasn't even PowerPC-specific stuff; it was
Bluetooth instead.

>  Do you not think that this is the case?  Please tell me if you think
> I'm wrong about this. 

I think you're wrong that it's going to be helpful for PPC. The best
that we can hope for is that we can regain the current situation -- the
healthy balance between tasks accomplished by package maintainers and by
the de facto 'arch maintainers' that we have at the moment.

I also think it's massively naïve to think that the new 'secondary'
build and release process could be put in place in time for a
simultaneous release of FC7 -- we have to be realistic and accept that
it's _very_ unlikely to happen in that timescale. We really shouldn't
hold PowerPC hostage to that; we should get the new process working with
a _new_ architecture, like IA64, SPARC or Alpha, before we make the
already-supported PowerPC architecture use it.

Once it's _working_, move PowerPC to it by all means. But we've spent a
lot of time and effort on Fedora/PPC and it isn't right to hold it
hostage to the implementation of a hazily-defined new process.

I'm just asking that we change one thing at a time. Get the new process
working, and _then_ we can move PowerPC to it. After we've done a proper
release for IA64 or SPARC or Alpha.

-- 
dwmw2