Ævar Arnfjörð Bjarmason [off-list ref] writes:
+ write_script .git/hooks/applypatch-msg <<-\EOF &&
+ true
+ EOF
+ write_script .git/hooks/unknown-hook <<-\EOF &&
+ true
+ EOF
git bugreport -s hooks &&
+
grep applypatch-msg git-bugreport-hooks.txt &&
- ! grep prepare-commit-msg git-bugreport-hooks.txt
+ ! grep unknown-hook git-bugreport-hooks.txt &&
+ ! grep -F .sample git-bugreport-hooks.txt
'
The above is not wrong per-se, but because we control which hooks
are and are not visible to "git bugreport" we run in this test, I
wonder if we should do
git bugreport -s tmp &&
sed -ne '/^\[Enabled Hooks\]$/,/^$/p' git-bugreport-tmp.txt |
sort >actual &&
sort >expect <<-\EOF &&
[Enabled Hooks]
applypatch-msg
EOF
test_cmp expect actual
instead. It is not like checking the "git help -a" output for list
of commands, which may grow or shrink as the system evolves.