On Tue, Apr 28, 2009 at 11:29:43PM -0400, Jeff King wrote:
> Note the repeated use of "hopefully". :) Maybe the earlier message is
> too hidden to rely on. We might be able to get by with checking "errno"
> for ENOTDIR after trying to lock the ref and using a different message,
> but I don't know how portable that will be.
Hmm, that actually doesn't work. errno is properly EACCESS in your
example, but the D/F problem doesn't actually set errno, since it is git
itself, and not a failed syscall, that determines that "foo/bar" is not
available because "foo" exists (and git must do it, because "foo" may be
a packed ref).
So I think we would need to simulate the errno setting, like the patch
below. That should generate the hint only when it would actually be
useful.
---
diff --git a/builtin-fetch.c b/builtin-fetch.c
index 0bb290b..ad00bd2 100644
--- a/builtin-fetch.c
+++ b/builtin-fetch.c
@@ -181,9 +181,9 @@ static int s_update_ref(const char *action,
lock = lock_any_ref_for_update(ref->name,
check_old ? ref->old_sha1 : NULL, 0);
if (!lock)
- return 2;
+ return errno == ENOTDIR ? 2 : 1;
if (write_ref_sha1(lock, ref->new_sha1, msg) < 0)
- return 2;
+ return errno == ENOTDIR ? 2 : 1;
return 0;
}
diff --git a/refs.c b/refs.c
index e65a3b4..79795d0 100644
--- a/refs.c
+++ b/refs.c
@@ -893,8 +893,10 @@ static struct ref_lock *lock_ref_sha1_basic(const char *ref, const unsigned char
* name is a proper prefix of our refname.
*/
if (missing &&
- !is_refname_available(ref, NULL, get_packed_refs(), 0))
+ !is_refname_available(ref, NULL, get_packed_refs(), 0)) {
+ last_errno = ENOTDIR;
goto error_return;
+ }
lock->lk = xcalloc(1, sizeof(struct lock_file));
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html