On Wed, Jun 15 2022, Junio C Hamano wrote: > Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes: > >> The existing "CURL" prerequisite is only used in one place, and we >> should probably name it "CURL_PROGRAM", > > True. It probably wants to become a lazy prerequisite defined only > in that program. Yes, I have that fix, but left it out for the "for v2.37.0" ... >> then rename "LIBCURL" to >> "CURL" as a follow-up, but for now (pre-v2.37.0) let's aim for the >> most minimal fix possible. > > OK. > >> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> >> --- >> t/lib-httpd.sh | 2 +- >> t/t5516-fetch-push.sh | 4 ++-- >> t/t5601-clone.sh | 4 ++-- >> t/test-lib.sh | 1 + >> 4 files changed, 6 insertions(+), 5 deletions(-) >> >> diff --git a/t/lib-httpd.sh b/t/lib-httpd.sh >> index 782891908d7..1f6b9b08d1d 100644 >> --- a/t/lib-httpd.sh >> +++ b/t/lib-httpd.sh >> @@ -29,7 +29,7 @@ >> # Copyright (c) 2008 Clemens Buchacher <drizzd@xxxxxx> >> # >> >> -if test -n "$NO_CURL" >> +if ! test_have_prereq LIBCURL >> then >> skip_all='skipping test, git built without http support' >> test_done > > This is not strictly needed in "the most minimal fix possible", is > it? Just checking. I figured if I was adding a $NO_CURL prereq I might as well change the only existing use of $NO_CURL to use it, which also assures the reader that our test suite is getting this from somewhere already. But I'll drop this in a re-roll.
