Re: [boost] CVS development branches?

From: Raja R Harinath (harinath_at_[hidden])
Date: 2001-03-10 22:23:40

Beman Dawes <bdawes_at_[hidden]> writes:
> At 07:15 AM 3/8/2001 -0500, John Maddock wrote:
>  >>So far so good.  But doing that will leave the initially committed 
> version
>  >>of each file as part of the main trunk.  So the questions are,
>  >
>  >No if your working copy is the branch, the a "cvs add" will add the file 
> to
>  >the branch, it won't be part of the main trunk and will be located in the
>  >attic of the actual repository.
> 
> I figured that out, finally, by mistake.  The sequence of operations that 
> works is tag (AKA branch), update (on the branch), add, commit.
> 
> Doing an update when you know nothing has been added yet is 
> counterintuitive   to me.  I did it by mistake originally, and then went 
> back and analyzed why it worked.

If you're not too worried about hard disk space, checking out a new
directory for the branch is a better idea. 

  $ cd /path/to/checkout/of/head
  $ cvs tag -b -r foo-branch 

  $ cd /somewhere/else
  $ cvs -d :pserver:foo_at_cvs.repo.sitory:/cvsroot/boost co -r foo-branch boost

This way, you don't get confusing behaviour with permanent tags
&c. (unless you KNOW how CVS works).

- Hari
-- 
Raja R Harinath ------------------------------ harinath_at_[hidden]
"When all else fails, read the instructions."      -- Cahn's Axiom
"Our policy is, when in doubt, do the right thing."   -- Roy L Ash