Ramsay Jones <ramsay@xxxxxxxxxxxxxxxxxxxx> writes: > Signed-off-by: Ramsay Jones <ramsay@xxxxxxxxxxxxxxxxxxxx> > --- > > Hi Lénaïc, > > As I said, I had a test failure from your previous series (Linux Mint) > which was solved with this patch. > > ATB, > Ramsay Jones > > t/t7900-maintenance.sh | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) This test comes from outside Lénaïc's series. I guess I am seeing a botched semantic merge conflict resolution of mine where the two topics met. The need for this fix made me a bit worried that the behaviour of resolve_scheduler(SCHEDULER_AUTO) forces us to always name the scheduler or risk a similar breakage, but in real life, nobody needs to "force" use of, say, launchctl on a platform where launchctl is not available so it is OK. > diff --git a/t/t7900-maintenance.sh b/t/t7900-maintenance.sh > index 8955aea9c8..36a4218745 100755 > --- a/t/t7900-maintenance.sh > +++ b/t/t7900-maintenance.sh > @@ -609,11 +609,11 @@ test_expect_success 'start and stop macOS maintenance' ' > > test_expect_success 'use launchctl list to prevent extra work' ' > # ensure we are registered > - GIT_TEST_MAINT_SCHEDULER=launchctl:./print-args git maintenance start && > + GIT_TEST_MAINT_SCHEDULER=launchctl:./print-args git maintenance start --scheduler=launchctl && > > # do it again on a fresh args file > rm -f args && > - GIT_TEST_MAINT_SCHEDULER=launchctl:./print-args git maintenance start && > + GIT_TEST_MAINT_SCHEDULER=launchctl:./print-args git maintenance start --scheduler=launchctl && > > ls "$HOME/Library/LaunchAgents" >actual && > cat >expect <<-\EOF &&
