atom feed69 messages in com.selenic.mercurial-develRe: RFC: dealing with dead, anonymous...
FromSent OnAttachments
Martin GeislerMay 7, 2010 10:33 am 
Dirkjan OchtmanMay 7, 2010 10:37 am 
Matt MackallMay 7, 2010 10:59 am 
Martin GeislerMay 7, 2010 11:14 am 
Matt MackallMay 7, 2010 11:46 am 
Dirkjan OchtmanMay 7, 2010 1:14 pm 
Gilles MorisMay 8, 2010 12:10 am 
Martin GeislerMay 8, 2010 1:31 am 
Martin GeislerMay 8, 2010 1:34 am 
Martin GeislerMay 8, 2010 1:48 am 
Martin GeislerMay 8, 2010 2:09 am 
Gilles MorisMay 8, 2010 3:50 am 
Martin GeislerMay 8, 2010 1:10 pm 
Gilles MorisMay 9, 2010 9:33 am 
Gilles MorisMay 9, 2010 1:49 pm 
Gilles MorisMay 9, 2010 1:49 pm 
Gilles MorisMay 9, 2010 1:49 pm 
Gilles MorisMay 9, 2010 1:49 pm 
Gilles MorisMay 9, 2010 1:49 pm 
Matt MackallMay 9, 2010 2:36 pm 
Matt MackallMay 9, 2010 2:43 pm 
Matt MackallMay 9, 2010 2:45 pm 
Matt MackallMay 9, 2010 2:47 pm 
Martin GeislerMay 9, 2010 4:11 pm 
Dirkjan OchtmanMay 9, 2010 10:50 pm 
Gilles MorisMay 9, 2010 11:31 pm 
Gilles MorisMay 9, 2010 11:49 pm 
Matt MackallMay 10, 2010 12:04 am 
Gilles MorisMay 10, 2010 12:37 am 
Gilles MorisMay 10, 2010 12:37 am 
Matt MackallMay 10, 2010 12:38 am 
Matt MackallMay 10, 2010 12:58 am 
Sune FoldagerMay 11, 2010 2:02 am 
Sune FoldagerMay 11, 2010 2:08 am 
Sune FoldagerMay 11, 2010 2:14 am 
Sune FoldagerMay 11, 2010 2:20 am 
Matt MackallMay 11, 2010 3:53 pm 
Matt MackallMay 11, 2010 3:56 pm 
Gilles MorisMay 12, 2010 12:52 am 
Gilles MorisMay 12, 2010 1:15 am 
Gilles MorisMay 12, 2010 1:15 am 
Gilles MorisMay 12, 2010 1:15 am 
Gilles MorisMay 12, 2010 1:15 am 
Gilles MorisMay 12, 2010 1:15 am 
Gilles MorisMay 12, 2010 5:17 am 
Augie FacklerMay 12, 2010 7:35 am 
Matt MackallMay 12, 2010 8:33 am 
Matt MackallMay 12, 2010 8:34 am 
Matt MackallMay 12, 2010 8:34 am 
Matt MackallMay 12, 2010 8:55 am 
Gilles MorisMay 12, 2010 2:33 pm 
Gilles MorisMay 13, 2010 7:31 am 
Matt MackallMay 13, 2010 1:48 pm 
Martin GeislerMay 17, 2010 1:09 am 
Martin GeislerMay 17, 2010 1:22 am 
Rafael Villar Burke (Pachi)May 17, 2010 2:07 am 
Augie FacklerMay 17, 2010 5:57 am 
Martin GeislerMay 17, 2010 6:56 am 
Augie FacklerMay 17, 2010 7:46 am 
Martin GeislerMay 17, 2010 8:13 am 
Augie FacklerMay 17, 2010 8:41 am 
Gilles MorisMay 19, 2010 8:00 am 
Martin GeislerMay 19, 2010 3:53 pm 
Martin GeislerMay 19, 2010 3:58 pm 
Augie FacklerMay 24, 2010 5:15 am 
Peter ArrenbrechtMay 26, 2010 6:23 am 
Rafael Fernández LópezJun 3, 2010 5:35 am 
Peter ArrenbrechtJun 3, 2010 6:14 am 
Martin GeislerJun 4, 2010 12:18 am 
Subject:Re: RFC: dealing with dead, anonymous feature branches
From:Rafael Fernández López (eres@kde.org)
Date:Jun 3, 2010 5:35:32 am
List:com.selenic.mercurial-devel

Hello all,

I first want to point out that git's way of deleting a remote branch is "sane", since their syntax is:

git push <remote> <localbranch>:<remotebranch>

so doing "git push origin :mycrazyfeature" means "push on origin at remote branch mycrazyfeature the empty local branch", which means "delete mycrazyfeature branch".

They later added (I think in 1.7.0) the option to delete a branch even in a more intuitive fashion:

git push <remote> --delete <remotebranch>

However, I think all this sort of "problems" we are having could be just fixed with the pushkey idea (http://mercurial.selenic.com/wiki/PushkeyConcept).

On the remote, an extension can be enabled to understand certain commands, meaning that we could drop a named branch or even an anonymous branch. A fast idea that comes to mind is performing an "hg clone -r ..." with all the wanted branches, leaving those that aren't of interest. Your old clone on the remote would become deleted, and the new one renamed to have the same name as the old repository.

This would be capable (being run on the remote) of removing all kind of branches (named and unnamed), having a very similar git behavior. We have to take into account the "genie is out of the bottle" problem, but in theory this is going to be done in personal repositories before they get merged upstream, so I believe this is not a problem.

Comments ?