On Mon, Oct 04, 2010 at 08:42:08PM -0500, Jonathan Nieder wrote: > >From my point of view, I think it would be best to start with > the smallest usable piece, which is the raw format. It probably > makes the most sense to error out when -u and --detect-dir-renames are > used together. Then unified diff could be reenabled in a separate > patch series on top of this one. Well, there's probably not much to do - in the same order of work than adding the check to error out :) > Another nice feature might be to let the directory-move > detection feed back into file-move detection to make it more > accurate. Have you thought about this? Would it be feasible, > and if so, would it be useful? That was suggested by Junio at that time. It would surely be useful in some cases, but it's not clear to me how frequently it would. And I'd suspect that quite some design changes would be necessary: that would be a 2-way influence, in that a nearly-complete bulkmove could bump the score of the latest rename that would make it a full bulkmove. Probably not trivial to do right. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html