On Wed, Nov 14, 2018 at 11:12 AM Junio C Hamano <gitster@xxxxxxxxx> wrote: > > Nguyễn Thái Ngọc Duy <pclouds@xxxxxxxxx> writes: > > > One of the problems with "git checkout" is that it does so many > > different things and could confuse people specially when we fail to > > handle ambiguation correctly. > > You would have realized that this is way too noisy if you ran "make > test", which may have spewed something like this on the screen. Oh I realize it because it's part of my git build and I often use "git co <paths>". I'm just telling (or kidding?) myself that I'm just so used to the old behavior and may need some time to feel comfortable with the new one. > [19:09:19] t4120-apply-popt.sh ................................ ok 1624 ms ( 0.26 usr 0.21 sys + 5.31 cusr 3.51 csys = 9.29 CPU) > [19:09:20] t9164-git-svn-dcommit-concurrent.sh ................ skipped: Perl SVN libraries not found or unusable > [19:09:20] t1310-config-default.sh ............................ ok 177 ms ( 0.07 usr 0.01 sys + 0.89 cusr 0.66 csys = 1.63 CPU) > ===( 20175;154 1297/? 155/? 6/? 3/3 2/? 4/? 4/? 3/? 5... )===Checked out 1 path out of the index > Checked out 1 path out of the index > Checked out 1 path out of the index > Checked out 1 path out of the index > Checked out 1 path out of the index > [19:09:20] t1408-packed-refs.sh ............................... ok 310 ms ( 0.06 usr 0.00 sys + 0.69 cusr 0.52 csys = 1.27 CPU) > [19:09:20] t0025-crlf-renormalize.sh .......................... ok 246 ms ( 0.03 usr 0.01 sys + 0.34 cusr 0.22 csys = 0.60 CPU) > > I am very tempted to suggest to treat this as a training wheel and > enable only when checkout.showpathcount is set to true, or something > like that. Maybe we just drop it then. I'm not adding a training wheel. I'm trying to make this complex command safer somewhat. But maybe this is a wrong direction. I'll give the idea "switch-branch / restore-path alternative commands" a go some time. Then the new generation can just stick to those and old timers stay with "git checkout". -- Duy
