Mike Jaques wrote:
>
> Ok, I know its early, and I don't want to too much open up a can of worms,
> but I was wondering if any thought has gone into a post 1.0
> release/development scheme. Namely creating a stable branch, and a
> development trunk, of which minor bug-fixes and regression-tested
> enhancements will go, and the trunk for the mainstream melee of regression
> and development work? A branch-point at version. 1.0, leading to 1.0.1
> 1.0.2 and so on, while development on the trunk will be deemed 1.1.1, 1.1.2
> and so on.
On a related issue:
Its often been stated that what Mozilla really needs is to stop
developing on the tip of the trunk.
Instead of pulling the very latest everything, there should be a well
known tag people can pull from, something akin to "last known good".
Then when things land which cause flames, be they hour long or week
long, most people are insulated because they pull last known good.
Given the way things are currently run, it would make sense for sherif
of the day to move the last known good tag when the people on IRC report
things are at least starting up.
The risk is that bad landings could cause the last known good to freeze
in place for some time. Actually, this might be a good thing...
Its going to get more and more important to avoid screwing completely
unrelated modules as things mature.
And, well, if not when 1.0 branches, when else should this begin to happen?
Thoughts?
--
AndyT (lordpixel)
