[cvsnt] Re: Branch merging - this seems wrong...

[Tony Hoyle]

Andreas Krey wrote:
> if I merge either direction, I always have A3 as the baseline revision and
> A4 and B3 as the two others. Merging the changes from A3 to A4 into B3
> produces the same result as merging the changes from A3 to B3 into A4:
> merging is a symmetric operation.

Merging is *not* symmetric.

Merging from A to B *does not introduce the changes in B back to A*

Your explanation is broken - there is no such thing as 'the changes from 
A3 to B3' as they are on different branches.  There is no direct delta 
without going via the branchpoint.

Your choices are:

Merge the changes from B2 to B3 into A4 as you propose, ignoring all 
changes in B prior to B2.

Or

Merge all the changes to B0 to B3 into A4 making A4 a merged copy of B, 
keeping all changes.

I can't see how your version makes any sense at all in the general case 
(although of course you're free to specify that with -j options).

Tony