I said last week that I would drop an email to the list with details of the plan for rawhide building. Current blocking items are follows: 1) Problems building core build root/chain packages in koji (gcc, python, coreutils, findutils, gettext, libssh2) 2) gcc upstream patches needed for build 3) other unknown core packages with non upstreamed components 1) The problems building the core packages mentioned about were primarily either unknown or pending an upgrade on the koji builders. Having done some testing (thanks Dennis) we're now fairly confident that an upgrade of the builders to a consistent image/kernel based on F-15 should fix most (and likely all) our current problems. Not to say there's not other issues we'll run into but I'm quietly confident we'll be able to move forward 2) ggc: blocking on PR 51442 [1], would be good if we could get forward movement upstream on this. 3). Other packages. I'm aware of a few other problematic packages that will need fixes but none are really what I would call core that should stop a mass rebuild from kicking off. I plan to deal with these break/fix style. There have already been a number of fixes I'm committed to rawhide during the F-15 initial upstreaming and subsequent building. [1] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51442 Moving forward for the plan to build rawhide. The process will be simply build the core buildroot packages and then run koji-shadow. Along side koji-shadow I'll likely do some ordered/targeted dep chain building. The problem with rawhide over the current F-15 building is that we really need to build things in order as we don't have a pre-built F-15 that we're just rebuilding in koji so we need to be aware of and take into account ABI and soname changes to ensure we don't end up with a broken infra. The basics is: 1) Import f16 noarch packages (alot of f17 is still f16 and even f15) - done 2) Import a snapshot of f17 noarch packages - done 3) build an ordered build root - reviewed and in progress (problems 1 and 2 above blocking) 4) run koji-shadow 5) break / fix 6) get upstream teams that are interested to build their stacks as we move further into the build (haskell, kde, xfce etc) 7) target initially a core minimal rootfs and move out from there. 8) onwards an upwards 9) f17 branch etc.... All as fast as possible! Currently we're awaiting the Seneca guys on the much anticipated infra upgrade and the gcc upstream bug. My thoughts are to do a local arm gcc build with the patch so as not to block short term progress. I look forward to thoughts and feedback. Peter _______________________________________________ arm mailing list arm@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/arm
