Hg sure is weird!

[mike.kupfer at xemacs.org]

>>>>> "MikeS" == Michael Sperber <sperber at deinprogramm.de> writes:

MikeK> I'm wondering if a changegroup hook like
[...]
MikeK> would produce more useful results.

MikeS> I don't think so: Just because I'm pushing several changesets at
MikeS> once doesn't mean they're related.

True.  OTOH, if the changesets are intermediate steps, having a separate
set of diffs for each changeset doesn't seem all that useful, either.
And it seems like with "incoming", you'd be more likely to trip over the
wrong-parent problem.

*shrug*

I guess it depends on which scenario (unrelated changes or related
changes) is more common.

mike