Moving Bugzilla to git?

[Frédéric Buclin]

Le 22. 10. 13 01:55, David Marshall a écrit :
> I recall that Max once gave me a lot of reasons why bzr was preferred
> to git. That was a long time ago, so some of his objections may have
> been mitigated.

For the record, here are the discussions about bzr vs Hg vs git:

http://blog.gmane.org/gmane.comp.bug-tracking.bugzilla.devel/day=20090904

http://blog.gmane.org/gmane.comp.bug-tracking.bugzilla.devel/day=20081220

Do not forget to read comments posted the following days as there are
much more replies than those displayed on a single page.


About killing bzr in 9 months, I'm opposed to that. I'm not speaking as
a former Bugzilla approver/core developer but as a Bugzilla installation
maintainer (GCC Bugzilla). It's way too short, and it's a pain to move
to another VCS when doing minor upgrades (i.e. on the same branch). Just
for your information, we stopped using CVS and moved to bzr for GCC
Bugzilla only *5 months* ago, see
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52239. It wasn't because I
didn't want to move to bzr earlier, but because I *technically*
couldn't. Now the migration is done, GCC Bugzilla is running 4.4.1, and
I don't expect to have to move to git to upgrade to another 4.4.x
release. So we should stop running the bzr server only when the 4.4
branch reaches EOL, which means at the end of 2015 or 2016 at the
earliest. Sorry if that seems too long for you.


And about attracting more contributors thanks to a migration to git, I
absolutely don't believe in this at all. I spent enough time on the
Bugzilla project to know that it's irrelevant. What matters is the
language used by Bugzilla, i.e. Perl, and developers coding in this
language.


LpSolit