On Mon, Feb 11, 2013 at 06:17:27PM +0100, Michal Nazarewicz wrote: > > I am happy to put it off until it becomes a problem, but I wonder if the > > Git::credential() interface is sufficient to express what we would want. > > It only allows two return values: true for approve, false for reject. > > But we would want a tri-state: approve, reject, indeterminate. > > Being it tri-state is not a problem. The last can be easily represented > by undef. Yeah, I think undef makes sense there. > > Reading the Net::SMTP code, it doesn't look like the information is even > > available to us (it really just passes out success or failure), so I > > don't think we can even make it work now. But it may be better to > > prepare the public Git::credential interface for it now, so we do not > > have to deal with breaking compatibility later. > > I guess. I left it as is since git-send-email won't make use of the > indeterminate values, but I can add it in this patchset as well. Yeah, I am more worried about third-party uses outside of the Git tree, which we may then break if we change the meaning of "undef" later. Thanks. -Peff -- 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