On Thu, Oct 06, 2005 at 11:15:09AM +0900, Stephen J. Turnbull wrote: > >>>>> "Thomas" == Thomas Zander <[EMAIL PROTECTED]> writes: > Thomas> What we could do is the following; > > Thomas> V V V V V V [local] > Thomas> foo > Thomas> *********** [2005foo - the full name of the patch that holds \ > Thomas> the next part] > Thomas> bar > Thomas> ^ ^ ^ ^ ^ ^
One issue is that in general each section of the marking doesn't come from a single patch, it's the result of a sequence of patches. We could just mention the "last one", or perhaps even "any one of the patches that lead to that version". > That's the Interface That Users Demand[tm], but is it clear that we > can _implement_ that? I think that with the new conflictors, we should be able to do so, although it could be a bit of a pain. It might be possible to hack this into the current code, but it would be very difficult (for technical reasons regarding how mergers are implemented), and I think it'll be much better to spend the effort on conflictors. > [...] If the latter, it's quite possible that the conflict region is now > in a state that the user will never have seen before! In that case > "local" is a lie. Indeed, both versions presented may be in states that no user has ever seen before, but I still think we're not too hosed. [local] just means that "local changes would have led to this version." -- David Roundy http://www.darcs.net _______________________________________________ darcs-users mailing list [email protected] http://www.abridgegame.org/mailman/listinfo/darcs-users
