[cvsnt] Re: Mergepoint question

[Bo Berglund]

On Fri, 25 Jul 2003 08:07:24 +0100 (BST), news at derived-software.ltd.uk
(Phil Richards) wrote:

>The way I do it is:
>Flip the sandbox over to the mainline (reset sticky tags etc).
>cvs -q update -P -d -jHEAD -jTheBranchName
>
>(I have to apologise - normally I'm doing this through WinCvs
>so I *might* have got it slightly wrong :-))
>
>The principle is, however, that the delta you want to apply
>to HEAD is precisely the difference between HEAD and TheBranchName.
>
>This will handle file creation and deletion on the branch - we use
>this approach all the time in our development process and it appears
>to work without error.
>
>Oh, it is normally a good idea to do:
>
>cvs -q update
>
>immediately afterwards - this clears up any files that have been
>modified on the branch and are exactly the same as the one on the
>mainline.  This happens a lot with what we do: the last stage
>on the branch is to merge the mainline to the branch and rerun all
>the unit tests after having sorted out the merge conflists
>
>phil
>--

Just to clarify for my understanding I quote from Cederquist:

-jrevision 
With two `-j' options, merge changes from the revision specified with
the first `-j' option to the revision specified with the second `j'
option, into the working directory. 
With one `-j' option, merge changes from the ancestor revision to the
revision specified with the `-j' option, into the working directory.
The ancestor revision is the common ancestor of the revision which the
working directory is based on, and the revision specified in the `-j'
option.

Now what I want to make sure is that -jHEAD -jBranch will result in a
sandbox HEAD file that is exactly like the one on the Branch tip,
ready for commit. Is this how it works?

Note that I am not after *merging* the Branch in because I have tried
that and it invariably causes a lot of conflicts. Instead I want to
promote Branch to the new HEAD completely discarding what is in HEAD.


/Bo Berglund