[cvsnt] Merging problems, getting strange conflicts....

[Tony Hoyle]

Bo Berglund wrote:
> Now, in the second (1-way) merge CVSNT should have noted that 1.8.2.10
> (tip of branch) already contained all of the head changes from the
> original branch point to the HEAD revision 1.11 since the mergepoints
> would tell it so.
> Consequently there is no need to get any data from HEAD, just copy over
> the contents of branch in the merge process.

Support for bidirectional merges varies between versions - it was 
removed a while back after a couple of people reported loss of data due 
to it, then after some discussion on the list put back as an option 
(just had a look and it's not in the 2.5.04 tree.. might be in the 
2.5.03 drops).  Evs has it, but that's a different mechanism.

> So why are the mergepoints not used?

The other case is if cvsnt detects *any* 'difficult' cases it'll fall 
back to branchpoint merges rather than risk losing stuff - these are 
mostly things like branches with no revisions (where the mergepoint is 
in the parent branch but finding it can be tricky).

Tony