Hi, On Thu, 27 Nov 2008, Junio C Hamano wrote: > Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes: > > > I have a strong suspicion that the narrow stuff will make the worktree > > mess pale in comparison. > > > > Note that I do not have time to review this myself (which is not > > helped at all by it being no longer a trivial single patch, but a full > > 10 patches!), but I really have a bad feeling about this. IMO it is > > substantially under-reviewed. > > Well, "a bad feeling" is not a convincing enough argument either, is it? > What kind of bad interaction are you fearing? I just remember the worktree stuff well enough. I had a bad gut feeling when it was proposed, and I had a bad impression of the (IMO way too intrusive) patch series implementing it. It was pretty buggy and affected Git in serious ways (in order to accomodate worktree, we broke operations in bare repositories at least once, for example). (So no, "bad feeling" is not convincing, but it is basically a primitive pattern matching in experiences that have not been fully analyzed, but that turned out to be bad enough.) I tried to fix it, but did not a very good job at it. In the meantime, I think I know why: there is no elegant way to implement this that is performant at the same time. (Just think of having git_dir be relative: this is a necessity for the performance, but ugly to implement in the presence of worktree where it may _need_ to be absolute). To me, the narrow patch series has all the looks of becoming the same type of nightmare: - it is intrusive, - it consists of a substantial number of patches (making bugs the opposite of shallow), - it is heavily under-reviewed, - it _needs_ a lot of changes to be accomodated, affecting common code paths, having all the potential to break existing workflows, and - there is as little interest in the feature from core Git developers as with worktree, literally guaranteeing that it will not, or only very slowly, and probably badly, get fixed if it breaks. And the worst part: I think that as with worktree, there has not been enough of kicking forth and back ideas how to design the beast, so I fully expect a subtle breakage that would require a redesign (which will be painful, with existing users of the feature). Maybe I am crying "wolf", but I _do_ want to caution against risking too much, too fast, with that feature. In other words, unless there is more interest in that feature, enough to generate a well-understood design before a good implementation, I'd rather see this patch series dropped. Ciao, Dscho -- 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
