The git-prune command checks reachability by doing a traversal, and then
checking whether a given object exists in the global object hash. This
can yield false positives if any other part of the code had to create an
object struct for some reason. It's not clear whether this is even
possible, but it's more robust to rely on something a little more
concrete: the SEEN flag set by our traversal.
Note that there is a slight possibility of regression here, as we're
relying on mark_reachable_objects() to consistently set the flag.
However, it has always done so, and we're already relying on that fact
in prune_shallow(), which is called as part of git-prune. So this is
making these two parts of the prune operation more consistent.
Signed-off-by: Jeff King <peff@xxxxxxxx>
---
builtin/prune.c | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/builtin/prune.c b/builtin/prune.c
index 04b6573945..97613eccb5 100644
--- a/builtin/prune.c
+++ b/builtin/prune.c
@@ -49,13 +49,12 @@ static void perform_reachability_traversal(struct rev_info *revs)
static int is_object_reachable(const struct object_id *oid,
struct rev_info *revs)
{
+ struct object *obj;
+
perform_reachability_traversal(revs);
- /*
- * Do we know about this object?
- * It must have been reachable
- */
- return !!lookup_object(the_repository, oid->hash);
+ obj = lookup_object(the_repository, oid->hash);
+ return obj && (obj->flags & SEEN);
}
static int prune_object(const struct object_id *oid, const char *fullpath,
--
2.21.0.rc0.586.gffba1126a0
