Roadmap for Bazaar...

Martin Pool mbp at canonical.com
Wed Sep 22 02:24:54 BST 2010


On 21 September 2010 21:05, John Szakmeister <john at szakmeister.net> wrote:
> It's been quiet on the list lately, so I figured I'd ask a question
> that has been lurking in the back of my mind for a while.  What's the
> goal for Bazaar over the next couple of releases?  And while I'm
> asking, what's the real aim of Bazaar?  Are you looking to be as
> wide-spread as SVN?  To appeal to similar kinds of people?  Just
> developers?  I ask, because it kind of sets the tone for how easy
> Bazaar needs to be to get started and scale with its intended
> audience.

That's a pretty good question.  Beyond what you mentioned on the list,
for the Bazaar team at Canonical we have two complementary
motivations: to make bzr a good tool that's both approachable and
powerful; and in particular to use it for vcs-based development of
Ubuntu.  The second builds on the first through things like
bzr-builddeb, bzr builder, soyuz build-from-recipe, and our priorities
with things like foreign branches.  I want bzr to continue to be
useful to the full range of community/open-source through to very
corporate users.

On the qualitative side I'd like to make it easier and more fun to
make any change.  Patch pilot has, I think, helped a lot, as have some
of the testing improvements.  To me this is mostly about people having
enough time to remove sharp edges when they see them.

Of the things you name, colocated branches are next on the Canonical
team's todo list, including making them work properly with bzr-git and
integrating them with looms.  Some work towards this has been done by
Jelmer and Parth, then nested trees.

Disabling mainline-changing push by default seems reasonable to me and
will avoid some confusion.  It will not be a very big code change
because the policy switch already exists, and we can mentor someone to
do it.  It mostly needs someone to consider the cases and update the
documentation to make sure it is explained to users and won't block
anything important.

http hosting is probably one place where people outside of Canonical
can help a lot, because of course we mostly encounter Launchpad
hosting.  Someone who really feels the itch is best placed to drive
it, but we will help.

Nested trees would come after that. I'd like to see nested trees done
in a hook-like way that is not intrusive when they're not used or not
being touched.

I'll also mention that we are looking to hire a very good software
engineer (to replace Robert, who is now working on Launchpad), and if
you're keen to work on any of these issues it would be a good way to
get lots of time to do it:
http://webapps.ubuntu.com/employment/canonical_BSE/

-- 
Martin



More information about the bazaar mailing list