On Wed, Apr 3, 2013 at 9:13 PM, Andreas Gal <[email protected]> wrote:

> What exactly is the problem we are trying to solve here though. How often
> do people still have to double-land themselves? The feedback I was getting
> is that the vast majority of landings is done by our awesome uplifting
> mini-team. In rare instances people have to help directly (thats what we
> should try to reduce by keeping trunk and v1.1 close as long v1.1 is still
> changing). Is this accurate or do you feel that uplifting is still a pain?
>

I suspect you're really asking about gecko here, but I'd like to share an
experience in gaia where the magic of uplifting didn't work so well.

Bug 836199 was blocked by bug 836647 so bug 836647 was uplifted. But this
patch caused a regression which was later fixed in bug 849280 so that was
uplifted too. But bug 849280 turned out to be subtly dependent on
bug 830644 which wasn't uplifted, thereby causing bug 855021. Bug 855021
could be fixed by uplifting bug 830644, but that bug has 25+ patches
attached to it, so that wouldn't be very smart. I ended up writing a patch
just for 1.0.1 and 1.1 to stop the dependency madness.

I was asked to uplift most of these myself because there were serious merge
conflicts caused by the divergence between v1-train and master over a long
period of time.

I guess what I'm trying to say is that the longer we keep those branches
open the more crazy things get, the more weird regressions there
are (pushing out release dates) and the more it hits our productivity
having to spend time resolving these things.

It's perhaps worth mentioning that gaia and gecko don't necessarily have to
follow the same release/branching schedule as they should be relatively
independent from each other, in theory.

Ben

-- 
Ben Francis
http://tola.me.uk
_______________________________________________
dev-b2g mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-b2g

Reply via email to