Re: On the subject of Git, Bazaar, and the future of Emacs development

[Stephen J. Turnbull]

Eli Zaretskii writes:

 > Then why did XEmacs choose Mercurial, and did not switch even now?

The main reason was that Mike (who had used Mercurial heavily on some
other projects) beat me to getting a reasonably complete conversion
(which is quite broken in some ways, but it almost never matters even
for exercising an idle curiosity, and it's never been a hindrance to
real work).  Several people expressed a a preference for the Mercurial
CLI, and at least one guy worked for Sun where Mercurial was the
"official" VCS at that time.  OTOH, at that time, it wasn't clear to
me that git was going to be more featureful than Mercurial so I didn't
fight it.

Right now Mercurial is a well-maintained tool with some ongoing
development whose only real downside[1] is that it isn't the market
leader.  So we stick with what we've got.

If a project is going to change (but for Emacs, my money is on "not
this year", Richard seems pretty adamant), what I wrote about git
being the market leader would matter in choosing a successor.


Footnotes: 
[1]  I hate the way "named branches" are implemented in Mercurial, but
I find Mercurial queues to be an adequate substitute for git-style
branching for most work.