Thread (13 messages) 13 messages, 3 authors, 2025-09-20
STALE279d

[PATCH 1/6] stash: tell setup_revisions() to free our allocated strings

From: Jeff King <hidden>
Date: 2025-09-19 22:40:29
Subsystem: the rest · Maintainer: Linus Torvalds

Possibly related (same subject, not in this thread)

In "git stash show", we do a first pass of parsing our command line
options by splitting them into revision args and stash args. These are
stored in strvecs, and we pass the revision args to setup_revisions().

But setup_revisions() may modify the argv we pass it, causing us to leak
some of the entries. In particular, if it sees a "--" string, that will
be dropped from argv. This is the same as other cases addressed by
f92dbdbc6a (revisions API: don't leak memory on argv elements that need
free()-ing, 2022-08-02), and we should fix it the same way: by passing
the free_removed_argv_elements option to setup_revisions().

I've added a test here which fails when built with SANITIZE=leak because
it calls "git stash show --". This by itself is not a very
interesting invocation, because there is nothing after the "--", and
thus the "--" is not really doing anything.

But I think the current parsing in show_stash() is a little
questionable. It splits the arguments into revision options and stash
options solely based on the presence of a leading dash, with no regard
to "--" at all. So:

  git stash show -- foo

will take "foo" as a stash option before we even pass anything to
setup_revisions(). And something like:

  git stash show -- 1

will show stash@{1}. But I would expect anything after the "--" to be a
pathspec. So in this example it would show only the part of the diff
that touched "foo". And something like:

  git stash show -p 1 -- foo

would treat "1" as a stash and "foo" as a pathspec.

That may be something we want to fix, but I want to focus here on the
leak-fixing without changing behavior. So this test is a little odd, but
does what we want without locking us in to any particular behavior (we
only care that "--" by itself does not change the output nor leak).

Signed-off-by: Jeff King <redacted>
---
I wonder if anybody actually cares that "git stash show -- foo" will
treat "foo" as a stash. If not, then it would probably be a fairly easy
#leftoverbits project to teach it to stop there and retain everything
after as a rev argument (which would then treat it like a pathspec).

 builtin/stash.c  | 3 ++-
 t/t3903-stash.sh | 6 ++++++
 2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/builtin/stash.c b/builtin/stash.c
index f5ddee5c7f..e5ab3c4cf5 100644
--- a/builtin/stash.c
+++ b/builtin/stash.c
@@ -957,6 +957,7 @@ static void diff_include_untracked(const struct stash_info *info, struct diff_op
 static int show_stash(int argc, const char **argv, const char *prefix,
 		      struct repository *repo UNUSED)
 {
+	struct setup_revision_opt opt = { .free_removed_argv_elements = 1 };
 	int i;
 	int ret = -1;
 	struct stash_info info = STASH_INFO_INIT;
@@ -1015,7 +1016,7 @@ static int show_stash(int argc, const char **argv, const char *prefix,
 		}
 	}
 
-	argc = setup_revisions(revision_args.nr, revision_args.v, &rev, NULL);
+	argc = setup_revisions(revision_args.nr, revision_args.v, &rev, &opt);
 	if (argc > 1)
 		goto usage;
 	if (!rev.diffopt.output_format) {
diff --git a/t/t3903-stash.sh b/t/t3903-stash.sh
index 0bb4648e36..1c9e589bbe 100755
--- a/t/t3903-stash.sh
+++ b/t/t3903-stash.sh
@@ -1741,4 +1741,10 @@ test_expect_success 'submodules does not affect the branch recorded in stash mes
 	)
 '
 
+test_expect_success 'stash show handles --' '
+	git stash show >expect &&
+	git stash show -- >actual &&
+	test_cmp expect actual
+'
+
 test_done
-- 
2.51.0.568.g6b54b97edf
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help