On 09.05.2010 14:31, Neal Becker wrote: > Adrian Buehlmann wrote: > >> On 09.05.2010 13:39, Neal Becker wrote: >>> I have a series of (in this case consecutive) patches I'd like to >>> collapse >>> into a single changeset. These are already committed and not tip. Any >>> ideas? >> >> I usually fold patches using the patch queue in hgtk log >> >> http://tortoisehg.bitbucket.org/manual/1.0/patches.html#patch-queues >> >> Enable the mq extension [1], import the patches into the queue using the >> context menu of the log viewer, then unapply the patch you want to fold >> and then select "Fold" from the context menu of the patch. >> >> http://bitbucket.org/abuehl/downloads/downloads/folding-a-patch-in-hgtk- > log.PNG >> >> [1] http://mercurial.selenic.com/wiki/MqExtension >> >> >> > > Unfortunately, I got stuck on step 1 - import patches to queue. I get: > > abort: revision 7 is the root of more than one branch > > I'm trying to combine changesets 7,8,9.
Given your history graph, folding csets will be pretty difficult. I never tried doing it with csets that far away from the tip. In your case, they are even part of a later merge. AFAICT merge csets can't be imported into mq. Folding csets with mq requires that you import all patches from the affected head back to the one you want to fold into the queue. If there is a merge in there, this won't work. Maybe you can import patches in a clone and redo all merges. IMHO I just wouldn't want to waste my time trying this. >From an purely academic angle, this might be an interesting exercise though. I just don't know of any tool that would automate what you are trying to do. For the academic exercise, you could start studying at: http://mercurial.selenic.com/wiki/ConcatenatingChangesets ------------------------------------------------------------------------------ _______________________________________________ Tortoisehg-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/tortoisehg-discuss
