Web lists-archives.com

[PATCH 1/2] help_unknown_ref(): duplicate collected refnames

When "git merge" sees an unknown refname, we iterate through the refs to
try to suggest some possible alternates. We do so with for_each_ref(),
and in the callback we add some of the refnames we get to a
string_list that is declared with NODUP, directly adding a pointer into
the refname string our callback received.

But the for_each_ref() machinery does not promise that the refname
string will remain valid, and as a result we may print garbage memory.

The code in question dates back to its inception in e56181060e (help:
add help_unknown_ref(), 2013-05-04). But back then, the refname strings
generally did remain stable, at least immediately after the
for_each_ref() call. Later, in d1cf15516f (packed_ref_iterator_begin():
iterate using `mmapped_ref_iterator`, 2017-09-25), we started
consistently re-using a separate buffer for packed refs.

The fix is simple: duplicate the strings we intend to collect. We
already call string_list_clear(), so the memory is correctly freed.

Signed-off-by: Jeff King <peff@xxxxxxxx>
 help.c           |  2 +-
 t/t7600-merge.sh | 14 ++++++++++++++
 2 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/help.c b/help.c
index a9e451f2ee..d3b3f64e3c 100644
--- a/help.c
+++ b/help.c
@@ -766,7 +766,7 @@ static int append_similar_ref(const char *refname, const struct object_id *oid,
 static struct string_list guess_refs(const char *ref)
 	struct similar_ref_cb ref_cb;
-	struct string_list similar_refs = STRING_LIST_INIT_NODUP;
+	struct string_list similar_refs = STRING_LIST_INIT_DUP;
 	ref_cb.base_ref = ref;
 	ref_cb.similar_refs = &similar_refs;
diff --git a/t/t7600-merge.sh b/t/t7600-merge.sh
index 7f9c68cbe7..7551ae3488 100755
--- a/t/t7600-merge.sh
+++ b/t/t7600-merge.sh
@@ -867,4 +867,18 @@ test_expect_success EXECKEEPSPID 'killed merge can be completed with --continue'
 	verify_parents $c0 $c1
+test_expect_success 'merge suggests matching remote refname' '
+	git commit --allow-empty -m not-local &&
+	git update-ref refs/remotes/origin/not-local HEAD &&
+	git reset --hard HEAD^ &&
+	# This is white-box testing hackery; we happen to know
+	# that reading packed refs is more picky about the memory
+	# ownership of strings we pass to for_each_ref() callbacks.
+	git pack-refs --all --prune &&
+	test_must_fail git merge not-local 2>stderr &&
+	grep origin/not-local stderr