Merge algorithms

[Stephen J. Turnbull]

Eli Zaretskii writes:

 > Well, granted, I understand the technical reasons for this
 > behavior.  I just don't see why merge punts so easily in this case.

Probably because you will still lose unless the workflow is very
hierarchical.  If you ever merge back from trunk into a
long-running-branch, then when you do merge into trunk, a naive "merge
source log entries float to top of merged ChangeLog" algorithm will
certainly result in interleaving the branch log entries with the trunk
entries.  Now what?

I think the proposed plug-in will probably work well for 95% of Emacs
contributors 100% of the time, and for the other 5%, 95% of the time.
It's probably worth doing on that basis.  But projects with many
people who are used to working in a more distributed fashion might not
get such good results.