I think branch would be the right approach in this case. I could not find a way to grant per branch permissions. Doug, is there a way to do that?
If I don't see any objections, I will create a branch on Monday. Olga -----Original Message----- From: Antonio Magnaghi [mailto:[EMAIL PROTECTED] Sent: Friday, November 30, 2007 9:07 AM To: [email protected] Subject: RE: guideline on branching Olga, I was wondering then for the specific case of the implementation of the Pig abstraction layer if we'll branch and, if so, one of the committers would need to set that up. Antonio -----Original Message----- From: Olga Natkovich [mailto:[EMAIL PROTECTED] Sent: Tuesday, November 27, 2007 8:32 AM To: [email protected] Subject: RE: guideline on branching Because for major changes it can break existing functionality and you don't want to keep things in your local copy till everything works because you don't get any backup for it. Olga -----Original Message----- From: Nigel Daley [mailto:[EMAIL PROTECTED] Sent: Monday, November 26, 2007 9:10 PM To: [email protected] Subject: Re: guideline on branching Why not just create patches occasionally from your local workspace and attach them to the Jira issue for others to review? On Nov 26, 2007, at 5:11 PM, Olga Natkovich wrote: > I think creating branches is a better option for reasons listed by > Antonio. If nobody objects to this, we could add this info to our > development process. > > Olga > > -----Original Message----- > From: Antonio Magnaghi [mailto:[EMAIL PROTECTED] > Sent: Monday, November 26, 2007 4:46 PM > To: [email protected] > Subject: guideline on branching > > I have a general question about how to handle a common situation. > > > > I would like to start some work to implement features as described in > spec docs on the wiki, namely the abstraction layer portion. This > will, most likely, require multiple check-ins. > > > > On one side, I could take a snap shot of the code base and work > locally on my dev machine. I don't think this is ideal for several > reasons: > this > would prevent others interested in the work to give feedback as I > progress, additionally I would need to back up my code periodically... > > > > One other possibility is to just make a branch, work on the branch and > then, when the changes have been approved, merge the branch with the > head. However, branching requires some coordination: to decide what > feature(s) really deserve a branch, avoid excessive branch > proliferation, coordination to merge the branch... > > > > In general, do we have some guideline to follow in this regard? > > > > Thanks, > > -a. >
