[RFC] implementation of revno:N:branch revisionspec (and general remark about diff implementation)

[Matthieu Moy]

Martin Pool <mbp at canonical.com> writes:

> It's a reasonable feature and the patch seems OK but it really needs a
> test before we can merge it.  Otherwise someone could break it in the
> future and we wouldn't know, and the revspec stuff is complex enough
> that could happen.

100% agree. I posted the patch as soon as it was ready to get a first
review, and didn't find time to write the test in the meantime, but
this will be done.

-- 
Matthieu