[p4] rolling back

[David Weintraub]

Here's the CVS command to "rollback" a change:

$ cvs update -j <checkedout revision> -j <rollbackRevision>

for example:

$ cvs update -j REL_1_3 -j REL_1_2

It calculates the diff between what you currently have and what you
want to roll back to, and merges those changes into each other. It
would be equivalent to being able to do this:

$ p4 integrate ...@<oldChangeList> ...#head
$ p4 resolve -at

The problem is that this doesn't doesn't work if the files involved
are on the same branch.

On 2/6/07, Ken Williams <ken.williams at thomson.com> wrote:
>
>
>
> On 2/5/07 9:36 PM, "Steve M. Robbins" <steve at sumost.ca> wrote:
>
> > On Mon, Feb 05, 2007 at 07:27:30PM -0500, David Weintraub wrote:
> >> Not too much of an issue if these are indeed the most recent and
> >> you've done no integration. The full directions can be found in Tech
> >> Note #14 <http://www.perforce.com/perforce/technotes/note014.html>. It
> >> basically involves sync'ing back to the earlier version, checking the
> >> files that were changed for edit, sync'ing to the head, and resolving
> >> the changes with a "p4 resolve -ay".
> >
> > I must say that I'm a little surprised by this.  Why isn't there a
> > 1-liner to apply the changelist in reverse?  Even the venerable CVS
> > offers (equivalent to) this functionality.
>
> If you can show me a CVS command that will undo the changes made during a
> certain timeframe, possibly including adds and deletes, respecting any
> changes that have been made since, I'll eat my [chocolate] hat. =)
>
> I think CVS doesn't even have enough information to start - the first step
> would probably be something like "have a meeting with your developers and
> figure out which revisions of which files were changes you want to get rid
> of".  In Zoltan's case, they're nicely packaged up into 3 atomic
> changelists.
>
> That said, I also figure lots of people probably write their own local
> "p4_undo" wrapper scripts, which is probably a sign that a "p4 undo" command
> built right into perforce could be a good idea.
>
>  -Ken
>
> _______________________________________________
> perforce-user mailing list  -  perforce-user at perforce.com
> http://maillist.perforce.com/mailman/listinfo/perforce-user
>


-- 
--
David Weintraub
qazwart at gmail.com