t7415: don't bother creating commit for symlink test
authorJeff King <peff@peff.net>
Mon, 11 Jun 2018 08:35:40 +0000 (04:35 -0400)
committerJunio C Hamano <gitster@pobox.com>
Mon, 11 Jun 2018 17:56:04 +0000 (10:56 -0700)
Early versions of the fsck .gitmodules detection code
actually required a tree to be at the root of a commit for
it to be checked for .gitmodules. What we ended up with in
159e7b080b (fsck: detect gitmodules files, 2018-05-02),
though, finds a .gitmodules file in _any_ tree (see that
commit for more discussion).

As a result, there's no need to create a commit in our
tests. Let's drop it in the name of simplicity. And since
that was the only thing referencing $tree, we can pull our
tree creation out of a command substitution.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t7415-submodule-names.sh

index a770d92..541bd81 100755 (executable)
@@ -135,13 +135,10 @@ test_expect_success 'fsck detects symlinked .gitmodules file' '
                tricky="[foo]bar=true" &&
                content=$(git hash-object -w ../.gitmodules) &&
                target=$(printf "$tricky" | git hash-object -w --stdin) &&
-               tree=$(
-                       {
-                               printf "100644 blob $content\t$tricky\n" &&
-                               printf "120000 blob $target\t.gitmodules\n"
-                       } | git mktree
-               ) &&
-               commit=$(git commit-tree $tree) &&
+               {
+                       printf "100644 blob $content\t$tricky\n" &&
+                       printf "120000 blob $target\t.gitmodules\n"
+               } | git mktree &&
 
                # Check not only that we fail, but that it is due to the
                # symlink detector; this grep string comes from the config