Re: [Gnu-arch-users] Future of GNU Arch, bazaar and bazaar-ng ... ?

[Martin Langhoff]

On 8/21/05, Matthieu Moy <address@hidden> wrote:
> What's the future of tla 1.x? Same for 2.0. Given that the main
> contributors of GNU Arch--except Tom--are mostly people working now on
> Bazaar-NG, I _guess_ Bazaar is the way of the future.

I apologize if this is a bit tactless and offtopic. After using Arch
for a long time, and playing with other SCMs ocassionally, I am
preparing to transition all my projects to GIT. I have been polishing
the cvs import that git includes, and I have an extremely draft Arch
importer I've been working on this weekend on a ferry trip across the
Cook Strait. I will be polishing it through the week.

This transition is tainted by the fact that patch-centric SCMs have
disappointed me a bit. GIT (I am actually using cogito, which provides
nice and easy shell wrappers) is patch-smart but not patch-centric,
and the more I use it, the more apparent it is that is a good design
decision.

YMMV. Arch and other patch-centric SCMs are forever-diverging: there
is nil support for identifying when two branches are identical. If a
small group of developers work on their own branches, and exchange
patches, most if the time you have the same tree, just different
"record" of patches. GIT knows that instantaneously, and marks it as
an un-branching: convergence. The trick of constantly hashing files
and trees pays of handsomely.

GIT doesn't natively do cherry-picking. It tries too hard to merge
branches fully to be good at that. But you can use Stacked GIT (StGIT)
which does cherry picking and many patch tricks on top of GIT. As git
is doing the 'formal' SCM, StGIT stacks patches on top of a formally
committed history. Patches in the stack are extremely malleable - a
weirdly nice concept of being able to "edit the patch". One of git's
GUIs, qgit, is poised to start doing cherrypicking, possibly based on
StGIT.

OTOH, Canonical people are doing some really interesting things with
bzr and hct. hct  is tied to their launchpad project (there was a good
talk at Debconf5 about it); I think it's interesting, specially if you
are looking for patch-centric tools. Canonical has a strong driver:
they need strong SCM tools to manage Ubuntu efficiently. It's a bunch
to watch, even if I don't agree with the technical decisions at the
core of their SCMs.

Sorry again for flogging a different scm. It's strange times for Arch
users, as tla is orphan and baz will probably be orphaned by Canonical
at some point not too far away.

cheers,



martin
pd: send flames privately if you must. please ;-)