ons 2011-10-26 klockan 13:34 -0400 skrev omalleys(a)msu.edu:
My concerns still are we need to make sure the current patches are
getting upstream to f16+ and rawhide.
And we are actively working on that together with mainline package
maintainers. The goal is even to get them into F15+ with as little
modifications as possible (hopefully none) in the ARM repository
compared to mainline F15, and as always for mainline to have as few
patches as possible relative to upstream.
But would be good with an activity day with some proven packager(s) to
speed up some of this.
I hate to see time wasted trying to track down already fixed bugs.
There are other people working on ARM support ie Ubuntu, and upgraded
versions of the software may actually fix some issues. (yes there
still will be issues.)
Yes, and there is at least on GCC a bit of duplicate work going on there
cross the distributions from needed bugfixes only being available in GCC
trunk and not the stable gcc release, and this for a very long time.
Given the decision has been made to forgo the F16 release.
Can we get away with taking a snapshot of the F15 base and F15 Updates
trees, and merge them and call that F15 base?
I proposed quite some time back that a such snapshot shold be the
startingpoint for F15 ARM, but it's decided the goal is to start as
close to GA as possible and then try to wiggle forward with updates
until things is in a shape that we can spin a release.
All the work that was done for F14 is really all in F16. It may not
appear in F15 base, but may appear in F15 updates. Depending on
when/if the patches were applied.
Yes, that's the normal rule, with ARM fixes from F14 being available in
F15 updates, or at least F15 git + koji build (but now garbage collected
There is a few cases where the change have only been applied to rawhide
(now F16), but those are more of an exception.
I have so far not seen any case where some needed ARM change have only
been applied to F14 or F15 and not later releases.