Felipe Contreras wrote: > Denton Liu wrote: > > Hi Felipe, > > > > On Mon, Dec 28, 2020 at 01:23:01PM -0600, Felipe Contreras wrote: > > > Signed-off-by: Felipe Contreras <felipe.contreras@xxxxxxxxx> > > > --- > > > t/t9902-completion.sh | 7 +++++++ > > > 1 file changed, 7 insertions(+) > > > > > > diff --git a/t/t9902-completion.sh b/t/t9902-completion.sh > > > index a1c4f1f6d4..2e59fe4de0 100755 > > > --- a/t/t9902-completion.sh > > > +++ b/t/t9902-completion.sh > > > @@ -2380,4 +2380,11 @@ test_expect_success 'sourcing the completion script clears cached --options' ' > > > verbose test -z "$__gitcomp_builtin_notes_edit" > > > ' > > > > > > +test_expect_success '__git_complete' ' > > > + __git_complete foo __git_main && > > > + test "$(type -t __git_wrap__git_main)" == function && > > > > s/==/=/ for this and the other patch. > > Sure, I'll include that in the next version. > > But I expect people will make further comments on the whole method to > find out if a function exists. I just noticed there's already a check for that in git-completion.bash: if declare -f $completion_func >/dev/null 2>/dev/null I suppose that's prefered to: if test "$(type -t func)" = function -- Felipe Contreras
