[haiku-development] Re: using a branch

  • From: Oliver Tappe <zooey@xxxxxxxxxxxxxxx>
  • To: haiku-development@xxxxxxxxxxxxx
  • Date: Sat, 15 Aug 2009 16:34:42 +0200

On 2009-08-15 at 14:39:49 [+0200], Ryan Leavengood <leavengood@xxxxxxxxx> wrote:
> On Sat, Aug 15, 2009 at 8:00 AM, Oliver Tappe<zooey@xxxxxxxxxxxxxxx> wrote:
> >
> > After you have created the branch, you either switch your local trunk 
> > sandbox
> > to the branch or make a separate checkout, depending whether or not you like
> > to be able to build the trunk, too. I personally always make a separate
> > checkout.
> 
> Given how long the process can take to checkout all our code (at least
> for those of us outside Europe), I might recommend copying the already
> existing trunk checkout to a new directory on your local machine and
> then doing svn switch if you want to keep the trunk version around.

Right.

> Also for those wondering the branch is fairly efficient on the server
> side since any unmodified code just points to the trunk version.
> 
> Also on a related note, is it acceptable or a good practice to only
> branch certain subdirectories? For example if one just wanted to
> refactor a small part of the code? AFAIK it will work in principle
> (after all trunk is just a directory to svn), I was just curious what
> others think of it.

I have been burnt by that line of thought recently, when I was pretty sure that 
I only needed to change a single configuration file for gcc. In a minimalistic 
approach, I branched just that one file. Of course, a couple of days later it 
occured to me that it would be better to adjust other files, too, requiring a 
second branch ...

So, I would leave it to subversion to optimize the branches and always do a 
full branch on trunk. If you switch your trunk, (or copy & switch, as you 
suggested above), it will be nearly as swift as having a small branch, but much 
more flexible.

cheers,
        Oliver

Other related posts: