BzrVsMercurial, choosing between the two

[David Cournapeau]

Matt Nordhoff wrote:
> <snip>
>
> Since there's a link to the discussion on one of Mercurial's mailing
> lists, I thought I should put it here too:
>
> <http://thread.gmane.org/gmane.comp.python.numeric.general/18130>
>   
Thanks, I should have done so myself in my email.
> I haven't read much of the thread or thought about this much. I will say
> that I think bzr and hg are both good choices. If more of your project's
> contributors are experienced with hg, it might be better to go with it.
>   
Yes, this is certainly a good argument. But maybe mercurial is missing 
something from bzr I am not aware, because I do not know much mercurial. 
IOW, I would hope for a more rationale comparison of the two, which is 
the reason why I sent my email.
> On your comment in the thread about bzr-svn using a different format,
> you can 'bzr upgrade --pack-0.92-subtree' (in the repo and all of the
> branches). That's the same as the new pack-0.92 format, plus subtree
> support. You can also use the 'rich-root-pack' format, which doesn't
> support subtrees, but is enough for bzr-svn.
>   
I knew about rich-root-pack (I think I mentioned that my discussion 
about formats were an oversimplification, though), but did not know 
about --pack-0.92-subtree. I do not see any mention of it in bzr help 
format (with a recent bzr.dev checkout) ?

cheers,

David