Can someone help describe the workflow when using bzr-svn...

[John Szakmeister]

On Sat, May 2, 2009 at 12:58 AM, Jean-Francois Roy <bahamut at macstorm.org> wrote:
> I've been working for a while now with Bazaar as my main VCS system at work,
> where interfacing with Subversion repositories is absolutely critical.
> Here's the workflow I have more or less settled on:
[snip]

Thank you for outlining your workflow!  It's great to see how others
are using bzr-svn.

FWIW, I had something very similar set up for my testing.  The one
part where the workflow "fails" is when it comes to a long running
feature branch.  In Subversion, we use the strategy of periodically
pulling in the mainline, and when the feature is finished, we
essentially apply the difference between the branch and trunk back to
trunk.  If we take the same approach in Bazaar, it'll reorder commits
which is highly undesirable (it made cherry picking harder, and it
also introduces "spurious commits"... in that we've already changes
and now they're being un-applied and re-applied... but you can't
easily tell they had been reviewed already).

In the end, I think the way we use Subversion (at least for
long-running branches) is actually pretty close to rebasing.  However,
how we use short-term branches is very close to the workflow you've
described in Bazaar.  So I guess there is a good use for both
workflows.

Thanks again for taking the time to describe your workflow.  It's
definitely useful, and I'm sure others will find it useful too when
they begin their bzr-svn adventures.

-John