So, how shall we use the stable branches now? Do we just commit everything into default and then $someone comes along for the next release and cherry-picks the stable commits? Do we commit twice? Do we commit once into stable and merge stable into default occasionally?
I noticed that e.g. 0e865f7d8bfd (Rewrote PropList output string sorting to get rid of C4Set<T>::Sort()) is in stable, although it's really nothing important for a stable release.
I noticed that e.g. 0e865f7d8bfd (Rewrote PropList output string sorting to get rid of C4Set<T>::Sort()) is in stable, although it's really nothing important for a stable release.
0e865f7d8bfd is not in stable. See the graph here. I branched right after the release commit.
The way I envision the branch usage is that every commit goes to default, and if the commit is a bugfix commit then it should also be cherry-picked (hg transplant?) to the stable branch as well immediately. This is not how we did it in the past where $someone (mostly Günther) cherry-picked the stable commits right before the release, but I think it simplifies things since the author of the commit knows best whether it is something for the stable branch or not.
The way I envision the branch usage is that every commit goes to default, and if the commit is a bugfix commit then it should also be cherry-picked (hg transplant?) to the stable branch as well immediately. This is not how we did it in the past where $someone (mostly Günther) cherry-picked the stable commits right before the release, but I think it simplifies things since the author of the commit knows best whether it is something for the stable branch or not.
Powered by mwForum 2.29.7 © 1999-2015 Markus Wichitill