Distributions going smart & bugfixes?

[Gustavo Niemeyer]

Hello Alex,

I hope you've had a terrific year crossover.

(...)
> Can we find a way in between? A branch were critical bugfixes are
> allowed to be made on stable code (e.g. non trunk)? This service
> is/was invaluable to us simple distribution supporters. Otherwise we
> would need to wade through all the known bugs to see whether there is
> already a patch attached or not.

Of course I can't prevent people from fixing their problems on their
own branches, and publishing that for other users.  Actually, that's
the beauty of open source.  The only comment I made is that this
branch is *not* what will become Smart version 1.X.Y.  The next
version of Smart will come out of trunk, and will contain changes
reviewed and unittested.

I want to speed up the integration of changes, but we need unittest
for pretty much anything touched (with rare exceptions).  Smart is
used in several different environments, and it's hard today to ensure
that new changes do not break the other environments in which the
developer isn't considering when he performs such changes.   Good
unittesting is one way to reduce this issue in the future.

Anders has been doing an amazing job on these branches and on pushing
these fixes forward, and I want to fully support him on integrating
these changes on trunk, even if I have to write the tests myself (as I
have been doing with the little time I'm able to dedicate).  I'm
hoping to work with Anders so that he can produce more changes which
are properly unittested and ready to go, so that I can get out of the
way and become less of a bottleneck.

-- 
Gustavo Niemeyer
http://niemeyer.net