Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
On 2003-09-05, Glen Starrett <grstarrett at cox.net> wrote:
> What is "bidirectional merging", and how is it used? It sounds
> interesting, but I'm not sure how I would use it.
The old code very nicely handled repeated merging from one place
(branch) to another (mainline).
The new code extends this to allow a merge operation to be performed
in the opposite direction after having done this.
I "needed" it for the following scenario (flying fish branches that
live for significant periods of time):
On branch (b1):
... do some work on fred
cvs update -jHEAD fred
... do some more work on fred
cvs update -jHEAD fred
... do some more work on fred
cvs update -jHEAD fred
Deliver branch code to mainline:
I used to have to do:
cvs update -jHEAD -jb1 fred
Now I will be able to:
cvs update -jb1 fred
(As a side effect, this will also record the delivery of the
branch as a mergepoint on the mainline, something the 2 -j
form can't.)
And, if I wanted to, I could continue development on b1, merge from
HEAD again (or not as the case may be), and then deliver the branch
to the mainline (again). Not that I do that very often.
With the previous version, using "cvs update -jb1 fred" could result
in merge conflicts because it was reapplying changes to the mainline.
It's not perfect - you tend to find commits for a lot of null-deltas
coming in. But I'm working on that :-)
phil
--
change name to "phil" for email