On Mon, Nov 19, 2007 at 12:30:04PM +0100, Eric Y. Kow wrote: > Hi William, > > [only CC'ing darcs-devel; feel free to forward] > > > > it sounds like you've found one. I tried searching the archives, > > > and I found a lot of discussion, but it was hard to find what the > > > final solution was (there seemed to be a bunch of options and it isn't > > > clear which one was chosen and what its final form was). > > I don't fully understand what is going on, but the basic gist was that > conflictors are back, and are now being implemented using the > ConflictedPatch representation originally intended for the cancellation > patch code. > > As the Conflicts FAQ mentions, "a patch followed by its inverse will not > conflict with any other patch". Come to think of it, I'm not sure what > that really means. Maybe it disappears so that you can commute stuff > past the pair p p^?. If so, I wonder what happens when you then > obliterate p^
Actually, things have been in flux over the last couple of weeks. The problem is certainly not solved yet, but we've got a good idea that it *can* be solved, and hints as to how we can do that. Most recently, I've decided to give up (for now, at least) on the idea of treating inverses nicely in the case of a conflict, as that was generating way too many special cases. In fact, it's starting to look like the implementable semantics are shockingly similar to those that darcs already (incompletely and inefficiently) implements. I'd rather not explain in detail just yet, as I've got a limited amount of time, and alas another bit of the (new) code requires a rewrite, as I now understand... :( So progress is being made, and I think I've got a working model, but since it's less than two weeks old (but based on the continuous work of a few months), I think it's better to get it actually coded up and working (on test cases) before explaining it (as that often reveals problems with ideas). -- David Roundy Department of Physics Oregon State University
signature.asc
Description: Digital signature
_______________________________________________ darcs-devel mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-devel
