On Jun 2, 2014 5:47 PM, "PulkoMandy" <pulkomandy@xxxxxxxxx> wrote: > To make things clear, I am willing to be the release coordinator for > alpha5. The short timeline set up by Kallisti5 and my vacations during > last week made this difficult, but now that I'm back from vacations, I > can handle this fine. Oh, OK -- I was all set up and had already started planning on doing it (actually I may be able to do it full time over the summer as volunteer work, and I already sent this to the mailing list a few times) but I suppose you're more qualified... I had started preliminary planning for an August release, if you want to go with that. > This morning I filtered the trac ticket and moved away things that > shouldn't delay alpha5 from the milestone. This doesn't mean these > issues won't be fixed for alpha5, if someone does the work, it can be > integrated into the release. However, if we keep adding stuff to the > next release, it will simply never happen. Most of those I assigned to myself, and yes, if I'm doing it full-time I will get to it. > Following this, there are things that are out of scope: > - ffmpeg update GCC4 completed. > - Qt update Already competed, just needs packaging. > - USB3 support Probably not happening since Jerome is busy. > - several hardware specific issues that can be worked around by > blacklisting drivers > I have moved matching tickets out of the alpha5 milestone. Some I approve of, some in don't. But since I appear to have been replaced, it's your call. > Several issues were raised on this ML thread. I'll create issues for > them if no one does. This includes: > * The haikuports repo is in inconsistent state, so we can't build a > full set of packages > * The upgrade path for alpha5 should be defined (do we want to have a > separate repo for it? should all packages require haiku "alpha5" > instead of a specific hrev?) > > Once these issues (and probably I missed some more) are solved, we can > see how far we are and setup a timeline. > > On the release process, I would like to lock commit access to the > trunk, and have all changes go through review on gerrit before they > are merged. This would allow for a simpler workflow than what we used > for previous releases, and could be made permanent to keep the trunk > in "releasable" state at all times. However, we don't have Gerrit > running yet, so if needed we will go with the traditional branch+merge > way once more. In that case, we need the branch to be more > short-lived, so it would be better to solve the major issues before > creating it. +1. -Augustin