[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: repeated merges from a branch
From: |
david |
Subject: |
Re: repeated merges from a branch |
Date: |
Wed, 11 Dec 2002 10:37:16 -0600 (CST) |
> I have reached the point where I need to merge from a branch to main a second
> time with the same file. The documentation I have found dedicates just enough
> words to say use "cvs checkout -j once -j twice file" and it's the human's
> responsibility to guess the right value for "once". This strikes me as being
> a
> real nasty limitation. There is no meta data preserved to figure this out
> with?
> -CTH
>
That is correct. CVS adheres strictly to the RCS file format, and does
not extend it much. Therefore, there is no meta data preserved.
What most people do is use a tag for the last merge place. For
example, where I last worked we had a branch named RELEASE_6_0.
We also had a revision tag RELEASE_6_0_MERGED which marked the
last merged spot. Since it was SOP there to make changes on the
branch and merge to the trunk, we had a script which did so,
automatically moving the RELEASE_6_0_MERGED tag. (We also
had another script that moved the tag without merging, in the
relatively rare case of a change on 6.0 that should not go to
the head branch.)
So, while this is annoying, there is a standard CVS practice that
makes up for it. It is one of the issues addressed in the
Subversion project, whose goal is to be a better CVS.
--
Now building a CVS reference site at http://www.thornleyware.com
address@hidden