[darcs-users] reviewing and testing pulls

Andrew Pimlott andrew at pimlott.net
Tue Apr 27 15:47:10 UTC 2004


On Tue, Apr 27, 2004 at 08:53:51AM -0400, David Roundy wrote:
> On Mon, Apr 26, 2004 at 05:38:30PM -0400, Andrew Pimlott wrote:
> > 1b. Why isn't there a nice human-readable form of a patch-set?
> > 
> > Has it just not been done yet?  I realize that a darcs patch can't be
> > represented as a simple diff, but is it too much to hope for something
> > diff-like?  It doesn't have to be usable for applying the patches, just
> > for reviewing them.  "darcs send" could optionally include both forms in
> > an email.
> 
> Would darcs whatsnew -u be similar to what you'd want? It's not perfect
> when there are hunks close to one another (the context gets weird in that
> case), but isn't bad either.

Well, since I'm thinking of the differences from a set of patches, it
would be "darcs diff -u".  Actually, tagging the tree, doing a pull, and
then diffing to the tag works pretty well.  Somehow, I didn't stumble on
this before.  However, it would still be useful to see something like
this in a send.

> It's a bit more of a pain to output patches
> with context, but is definitely doable, and patches with context are still
> valid darcs patches (darcs currently just ignores the context, although it
> could be used for redundant checks).  Adding -u support to send is
> something I've wanted to do, but never got around to.

If I understand what you're proposing, the problem with this is that the
hunks would still be divided among multiple patches.  If multiple
patches touch one file, I'd like to see the cumulative change.  This is
why I said that the human-readable form of send doesn't have to be
suitable for applying the patches.

> > Although I could use a separate tree, and throw it away to undo the
> > pull, it would be nice for darcs to remember which patches were added in
> > the same batch, and let me unpull the whole batch.
> 
> To me, this doesn't seem like it would be a very useful feature.  I'd say
> that if you are likely to want to unpull a bunch of patches that you pulled
> all at once, you should pull into a temporary repo.  Of course, you may not
> know beforehand that you won't want the changes, but it doesn't seem *that*
> hard to just run unpull a bunch of times.

Would it be feasible to unpull back to a tag?  Right now, unpull
--tag-name removes the tag patch, so I guess it would need a different
option.  But if I could unpull to a tag, I'd just tag, pull, and unpull
to the tag to reject.

Andrew




More information about the darcs-users mailing list