Junio C Hamano <gitster@xxxxxxxxx> writes: > Matthieu Moy <Matthieu.Moy@xxxxxxx> writes: > >> Junio C Hamano <gitster@xxxxxxxxx> writes: >> >>> Matthieu Moy <Matthieu.Moy@xxxxxxx> writes: >>> >>>> ... In this command, I obviously don't want to move the HEAD, >>>> so why the hell should git refuse "git reset --hard -- file1 >>>> file2"? >>> >>> (1) because allowing it begs the question "what would happen if >>> I replace that --hard with --soft", the answer to which has >>> to be "nothing happens". That answer is logically correct >>> but not useful from the end user's point of view; >> >> True, but that's already true of "git reset --hard", alone, isn't >>it? Unless I missed something, "git reset --soft" is a no-op if you >>don't provide arguments. > > I think I see where you are coming from. To you "git reset" (any of > three reset_type) is something special and different from "git reset > <any-commit>". If we take that approach, "git reset" and "git reset > HEAD" can behave differently and you can certainly argue "git reset > --soft" is a no-op. > > To me, it is just a shorthand of not giving an explicit commit > (and defaults to HEAD), and it is not particularly special. > Neither "git reset --hard not-HEAD" nor "git reset --soft > not-HEAD" is a no-op. > > That makes "git reset --soft not-HEAD -- paths" being a no-op > doubly confusing while "git reset --hard not-HEAD -- paths" does > what exactly "git checkout not-HEAD -- paths" does. Well, it is not "doubly confusing" in my book but "just as confusing". I would make the usage of everything consistent, and spew out a warning in the case of git reset --soft HEAD and git reset --soft anything files that says "`git-reset --soft HEAD' has no effect." or "`git-reset --soft' has no effect with a file list." I'd prefer a warning instead of an error since it is conceivable that the calls occur in the course of a script that does something useful even when $commit is HEAD. -- David Kastrup, Kriemhildstr. 15, 44793 Bochum - 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
