[cvsnt] Mergepoint numbering?

[Tony Hoyle]

John Kinson wrote:
> Tony, just to be clear, this is a bug: if I merge two branches to the 
> trunk, commit once, make further changes on both branches, then attempt 
> to merge a second time, then one of the branches may needlessly generate 
> conflicts as the mergepoint will be missing (the entire branch will be 
> merged again).

The client only has storage for a single mergepoint, and that data is 
generated by the server during the merge itself.

The 'mergepoint1' terminology came from an original idea of recording 
both -j options on a two-branch merge.  I only realized later that it's 
a bit pointless to do so.

> The workaround is obviously to only ever merge one branch at a time and 
> commit before merging the next branch, which is good practice for 
> traceability anyway.

This should probably be in the documentation somewhere.

Tony

-- 
Before you criticize someone, walk a mile in their shoes.
That way, if he gets angry, he's a mile away and barefoot.

Tony Hoyle <tmh at nodomain.org>  Key ID: 104D/4F4B6917 2003-09-13
Fingerprint: 063C AFB4 3026 F724 0AA2  02B8 E547 470E 4F4B 6917
FWD: 413300