Re: arch patch queue manager

> I have a question about branching with tla. I need to branch from
> gephex--main--0.4 for the v4l refactoring stuff. If i do this with
>
> "tla tag -S [email protected]/gephex--main--0.4
> [email protected]/gephex--refactor-v4l--0.4"
>
> tla fetches the base-0 revision and the 1726 patches from the
> [email protected] archive. Creating a microbranch is a quite time
> intensive operation if tla does it this way. Why doesn't tla use the
> cacherev at patch-1722 and patches it up to the newest revision?
>

Why the -S flag? I think I created the branch by first doing "tla
setup-archive gephex--georg--0.4"
followed by a "tla tag" (without flags).
(Cannot verify this at the moment, maybe the syntax is wrong)

> | The remaining TODOs are:
> |
> |     - cron job that calls arch-pqm --run
> |     - create one branch of gephex--main--0.4 and gephex--main--0.5 for
> |       every developer and tell arch-pqm where it is
> |       (those branches can be on different archives, so for example
martin
> |        could have his archive on tyr, I could mirror my offline archive
> |        to gephex.org, ...)
>
> My tla archive is at http://arch.gephex.org/martin and my current
> development branch is gephex--refactor-v4l1--0.4. I'm not sure if it
> would be better to create an extra "ready-to-merge" branch.  How do you
> plan to organize your microbranches?

I keep one "integration branch" gephex--georg--0.4 and create branches for
different tasks
(bug fixes, new features) from that branch.

Advantage: my changes are always merged into gephex--main--0.4 from the same
branch.

Georg