On Fri, Apr 12, 2019 at 11:32:25AM +0900, Junio C Hamano wrote: > Jeff King <peff@xxxxxxxx> writes: > > > On Thu, Apr 11, 2019 at 03:20:52PM -0300, Bárbara de Castro Fernandes wrote: > > > >> This new proposed --amend option, although semantically different, > >> would have a very similar functionality to the already existing -f > >> option. So should we, perhaps, change -f's behavior to treat the tag > >> as a new one, treating the old one as if it never existed (as I think > >> Junio was saying)? By this I mean the command should fail if the user > >> doesn't give a SHA-1 and the previous message wouldn't be preloaded. > >> --amend, on the other hand, would give the user an opportunity to > >> revise the tag by opening, by default, the editor with the > >> pre-existing message unless given the '--no-edit' option, and if not > >> given a SHA-1 it would keep on using the previous one. > > > > Yes, that's what I'd expect it to do (so yes, it's also different from > > "-f" in that it defaults to the existing tag destination instead of > > HEAD). > > Do you mean you'd expect "--amend" to do that, which is different > from what "-f" does, so they should not be conflated into one? > > If so, I think that makes sense and changing the behaviour of "-f" > is too confusing. Yes, sorry to be unclear. The "it" in my sentence was "--amend". -Peff
