So after some bullying at lunch the other day, I got off my arse and
updated my private copy of webrev which implements the ability to spit out
a webrev based on a single specified changeset or a range of changesets,
rather than the changesets that are in your workspace but not in the
parent. It now addresses the following bugs:
6793444 webrev could support specifying revision range for mercurial
workspaces
6793446 webrev could support "hg copy" better
6793448 webrev should automatically link to defect.opensolaris.org bugzilla
bugs
You can find it (not a webrev for it) at
http://cr.opensolaris.org/~dduvall/webrev
I've sat on it for a day or two thinking about its current issues, and I'm
beginning to think that I really need to address them. In particular, the
-r support is implemented using the output from various hg commands, rather
than putting the support into hg-active, which is probably where it
belongs. In addition, copy support only works with -r, which is likely to
be at least occasionally frustrating for some people.
Unfortunately, I looked at hg-active, and the module code underlying it,
but I've no idea where to start trying to add this support. Rich mentioned
to me some time ago that getting this right was next to impossible, which
was one of the reasons I didn't move forward, but I've lost the email that
detailed the issues involved, so I'm not sure how much of the argument was
perfectionism and how much was related to solid reasons it couldn't be
done.
Anyway, if anyone could help me get started with what I'd need to do to
reimplement this in Python, that'd be awesome. If Rich remembers what his
concerns were, I'm more than happy to try to make a decision based on that.
And if anyone wants to try this out, please do so.
Also, if anyone has played with the pbranch extension and has a good idea
of how webrev should interact with it, that'd be awesome. I mostly use MQ,
and webrev -r <tag> works great with that, but since pbranch does its work
on a much more complicated change graph, it's not clear what the interface
should be. Perhaps we'll want a webrev -p <branchname> for that, which
could be used for both pbranch branchnames and applied MQ patch names?
Thanks,
Danek
_______________________________________________
tools-discuss mailing list
tools-discuss@opensolaris.org
