Hi

I noticed "darcs rollback" lacks a --last=NUMBER optional argument,
the sort that unrecord et. al. have.

Could have used it today, and I think it's generally a nice argument
and useful for rollback in particular.

Regrads,
Adam

darcs rollback is designed to select a single patch, which is why "--last" is
not an option. 

I'm fine with leaving it that way and resolving this as "wont-fix". 

David, Eric: You have both touched this ticket. What do you think about rollback
selecting more than one patch?

> David, Eric: You have both touched this ticket. What do you think about rollback
> selecting more than one patch?

Well, I don't have a very clear picture in my head of how --last is
really supposed to behave, but I figured that since rollback accepts
--match and --patch, --last would be a pretty reasonable addition as
well.

The only thing that is affected is which patches darcs proposes to
you.  It is still only one patch that gets rolled back, but having
these matchers might help to narrow down the set of patches to choose
from.

darcs rollback --last=1, for instance, might be a good use for it.

P.S. welcome back! and good luck/happy hacking with the Haskell

On Tue, Jan 01, 2008 at 10:56:30PM -0000, Eric Kow wrote:
> > David, Eric: You have both touched this ticket. What do you think about
> > rollback selecting more than one patch?
> 
> Well, I don't have a very clear picture in my head of how --last is
> really supposed to behave, but I figured that since rollback accepts
> --match and --patch, --last would be a pretty reasonable addition as
> well.
> 
> The only thing that is affected is which patches darcs proposes to
> you.  It is still only one patch that gets rolled back, but having
> these matchers might help to narrow down the set of patches to choose
> from.
> 
> darcs rollback --last=1, for instance, might be a good use for it.

Yeah, that makes sense.  Internally, it'd be a bit of a pain, and I think
if we added --last, it'd only be consistent to add --matches and --patches
as well.  And I suspect that those might be confusing.

Incidentally, I'm seriously considering removing rollback for darcs 2.
With the way the conflict handling ended up turning out, it's no more
useful than in darcs 1, and it's always been a bad command.  :(
-- 
David Roundy
Department of Physics
Oregon State University

David is right that it would be consistent to add --matches and --patches along
with --last, and I don't think appropriate in this case. 

I don't think the "--last=1" example is a worthwhile use-case, since the very
last patch is the first one you are offered to rollback in interactive mode. 

So, I'm being assertive and marking this as "wont-fix", but anyone who disagrees
is a welcome to re-open the wihlist item. 

( Of course, the whole thing will be moot if the rollback is removed. )