[darcs-users] annotate output

Dave Love d.love at liverpool.ac.uk
Tue Sep 14 10:00:11 UTC 2010


Eric Kow <kowey at darcs.net> writes:

> On Tue, Sep 07, 2010 at 14:51:16 +0200, Petr Rockai wrote:
>> > I just saw the disjointed thread about this in Gmane, but I can't find a
>> > description of what's actually proposed for annotate.  Could someone
>> > point me at it (e.g. as a message id, if it was posted here)?
>
> See also http://bugs.darcs.net/issue25

The timestamp in that git format looks longer than necessary, which
takes extra space but, more importantly, it's not in UTC.  Also it's not
clear to me that I can find the relevant patch name easily, as I can
with Peter's format.

> Unfortunately, I had not done a good job of making it clear that
> the proposed git-blame style UI was the proposed way of going
> forward.  I don't have particularly strong feelings anyway, just
> as long as we have good reasons for picking what we finally go
> with.

Right.  I hope a real example of practice is useful.

> Folks who would like to weigh in are more than welcome -- we need your
> perspective -- but please and take take stock of what has already been
> said to help us avoid going around in circles.

>From the point of view of using this in Emacs VC, I think I want:

* A format that's trivial to parse, probably in a fixed format like the
  original one from CVS -- don't make me parse XML or YAML;

* Something that's easily distinguishable from the previous format,
  though doing an initial `darcs --version' is probably OK for that;

* Information that gives me the patch timestamp, author, and name
  without having to run darcs for each one in the file.  I think Peter's
  initial table in the human-readable version is OK for that, but I
  haven't looked at implementing it yet.

By the way, the current implementation of the Emacs support is at
<http://www.loveshack.ukfsn.org/emacs/vc-darcs.el>, but probably moving
soon.


More information about the darcs-users mailing list