On Tue, Jul 11, 2006 at 06:40:54PM -0700, Zack Weinberg wrote:
> I was thinking about using commit date as a further heuristic, i.e.
> when we have two LCAs neither of which is an ancestor of the other,
> merge the newest one first; furthermore, when we have three or more
> heads with the same LCA, merge the newest two first.
I don't have any really clear picture of whether this heuristic would
help or hurt. My general bias is towards simple and predictable
systems that involve minimal cognitive clutter (you can just _hear_ my
math background seeping through), so this doesn't excite me too much
but, you know, hey, if you can make it work :-).
> However, it
> seems like a huge pain to get from a revision_id to its commit date,
> and in fact I'm not sure the date cert is guaranteed to exist.
These are both true. (In addition to the mechanisms mentioned elsewhere
on the thread, a simple interrupted netsync can leave one with a revision
but not all of its certs.)