On Tue, Oct 25, 2011 at 7:35 PM, Chris Tyler <chris@xxxxxxxxxxx> wrote: > On Tue, 2011-10-25 at 19:17 +0100, Peter Robinson wrote: >> On Tue, Oct 25, 2011 at 6:09 PM, Chris Tyler <chris@xxxxxxxxxxx> wrote: >> > On Tue, 2011-10-25 at 17:49 +0100, Peter Robinson wrote: >> >> So what's the status of this? Are we using the F-15 as the external >> >> repo base for building F-17? When is it going to happen? My >> >> understanding is we have enough of both arches built that we can start >> >> this rolling. It would be good to see some public movement on this now >> >> and to get this rolling rather than delaying further. >> > >> > No one's delaying, but there is a sequence to this that we shouldn't >> > short-circuit: >> > >> > * Sync up the armv5tel and armv7hl pre-koji package sets <= We are here >> > * Do final F15 build in Koji >> >> It was my understanding from dgilmore's proposal above that we were >> going to skip F-15 and just go straight to F-17. Ultimately whether we >> do a final F-15 or F-17 it doesn't really matter, the current build >> tool chain for them both is derived from gcc 4.6.1 > > The proposal was to skip F16. > > We're close to F15 and we need it. If we don't ship it, we have a > substantial time gap where we have no supported Fedora release. Shipping > F15 also gives us experience with systemd on ARM (and other pieces of a > fairly complete release, such as haskell and ocaml) which can feed into > the F17 development cycle. > >> > * Release F15 (& start F15-updates builds) >> > * Finalize koji changes for rawhide (rpm, rpmbuild, yum, koji, armv7hl >> > rootfs) >> > * Do necessary prebuilds for rawhide (build-previous or other) >> > * Build for rawhide, shadow PA as closely as possible >> > * Branch F17 when PA does >> > >> > At the end of this process we'll be building for at least three releases >> > (F15updates, F17, rawhide) across two archs - 6 builds total. >> >> Why? I thought the above process was with the idea of just using the >> F-15 work that was done as a base to go straight to F-17 and forget >> the rest. Clearly people need to start documenting the exact process >> for all to see. > > The goal is to get as close to PA as possible, as quickly as possible. > If we follow only the F17 branch, we're in a catch-up position *again* > for F18. If we follow both rawhide and F17, we're in much better shape > for F18. And if we ship F15 for ARM, we need to ship updates (and prove > that we can ship timely updates, which will be important when making the > case for primary arch status). That's what I mean. F-17 doesn't fork from rawhide until alpha, so at the moment F-17 == rawhide for a couple of months yet and hence there is no F17 branch to discuss. I'm sort of hoping we have a lot of that build by branch time and its hence not a problem. Peter _______________________________________________ arm mailing list arm@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/arm
