2008/5/10 Ian Lynagh <[EMAIL PROTECTED]>:
> I've been trying to understand how conflictors work, prove that they do,
> and prove that they do quickly. So far I've mostly done the first of
> these :-)
What kind of semantics for conflictors you propose here?
Say you had conflicting branches A and B:
r1: A
r2: B
and r1 pulls B, resulting r1: A conflictor [ø A B]. Now if r2
generates rollback of B (~B), and r1 pulls it, too, do we see
r1: A conflictor [A A :B] conflictor [{A, B} {} B:~B]
like we do now or something like
r1: A conflictor [A A :B] rotcilfnoc [A A :B]
as I would naïvely expect? Do you expect some difficulties with that
kind of approach (besides completely ruining current conflict
resolution semantics)?
Or something more complicated, so that if we pulled A to r2, we would see
r2: B ~B A
as end result?
--
Pekka.Pessi mail at nokia.com
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users
