Merge pull request #2167 from GitoxideLabs/copilot/fix-3952f55e-8faf-4737-886f-09e74cab4ca8

Make `gix::Repository::has_object()` consider empty blob ids to be always present

Author: Byron

gix/src/repository/object.rs

@@ -49,6 +49,14 @@ impl crate::Repository {
                 repo: self,
             });
         }
+        if id == ObjectId::empty_blob(self.object_hash()) {
+            return Ok(Object {
+                id,
+                kind: gix_object::Kind::Blob,
+                data: Vec::new(),
+                repo: self,
+            });
+        }
         let mut buf = self.free_buf();
         let kind = self.objects.find(&id, &mut buf)?.kind;
         Ok(Object::from_data(id, kind, buf, self))
@@ -86,6 +94,8 @@ impl crate::Repository {
     /// Obtain information about an object without fully decoding it, or fail if the object doesn't exist.
     ///
     /// Note that despite being cheaper than [`Self::find_object()`], there is still some effort traversing delta-chains.
+    /// Also note that for empty trees and blobs, it will always report it to exist in loose objects, even if they don't
+    /// exist or if they exist in a pack.
     #[doc(alias = "read_header", alias = "git2")]
     pub fn find_header(&self, id: impl Into<ObjectId>) -> Result<gix_odb::find::Header, object::find::existing::Error> {
         let id = id.into();
@@ -95,6 +105,12 @@ impl crate::Repository {
                 size: 0,
             });
         }
+        if id == ObjectId::empty_blob(self.object_hash()) {
+            return Ok(gix_odb::find::Header::Loose {
+                kind: gix_object::Kind::Blob,
+                size: 0,
+            });
+        }
         self.objects.header(id)
     }
 
@@ -109,7 +125,7 @@ impl crate::Repository {
     #[doc(alias = "exists", alias = "git2")]
     pub fn has_object(&self, id: impl AsRef<gix_hash::oid>) -> bool {
         let id = id.as_ref();
-        if id.to_owned().is_empty_tree() {
+        if id.to_owned().is_empty_tree() || id.to_owned().is_empty_blob() {
             true
         } else {
             self.objects.exists(id)
@@ -130,6 +146,12 @@ impl crate::Repository {
                 size: 0,
             }));
         }
+        if id == ObjectId::empty_blob(self.object_hash()) {
+            return Ok(Some(gix_odb::find::Header::Loose {
+                kind: gix_object::Kind::Blob,
+                size: 0,
+            }));
+        }
         self.objects.try_header(&id).map_err(Into::into)
     }
 
@@ -144,6 +166,14 @@ impl crate::Repository {
                 repo: self,
             }));
         }
+        if id == ObjectId::empty_blob(self.object_hash()) {
+            return Ok(Some(Object {
+                id,
+                kind: gix_object::Kind::Blob,
+                data: Vec::new(),
+                repo: self,
+            }));
+        }
 
         let mut buf = self.free_buf();
         match self.objects.try_find(&id, &mut buf)? {

gix/tests/gix/repository/object.rs

@@ -1,3 +1,5 @@
+use gix_odb::Header;
+use gix_pack::Find;
 use gix_testtools::tempfile;
 
 use crate::util::named_subrepo_opts;
@@ -507,6 +509,76 @@ mod find {
         );
         Ok(())
     }
+
+    #[test]
+    fn empty_blob_can_always_be_found() -> crate::Result {
+        let repo = basic_repo()?;
+        let empty_blob = gix::hash::ObjectId::empty_blob(repo.object_hash());
+        assert_eq!(repo.find_object(empty_blob)?.into_blob().data.len(), 0);
+        assert!(repo.has_object(empty_blob));
+        assert_eq!(
+            repo.find_header(empty_blob)?,
+            gix_odb::find::Header::Loose {
+                kind: gix_object::Kind::Blob,
+                size: 0,
+            },
+            "empty blob is considered a loose object"
+        );
+        assert_eq!(
+            repo.try_find_object(empty_blob)?
+                .expect("present")
+                .into_blob()
+                .data
+                .len(),
+            0
+        );
+        assert_eq!(
+            repo.try_find_header(empty_blob)?,
+            Some(gix_odb::find::Header::Loose {
+                kind: gix_object::Kind::Blob,
+                size: 0,
+            }),
+            "empty blob is considered a loose object"
+        );
+        Ok(())
+    }
+}
+
+#[test]
+fn empty_objects_are_always_present_but_not_in_plumbing() -> crate::Result {
+    let repo = empty_bare_in_memory_repo()?;
+    let empty_blob_id = gix::hash::ObjectId::empty_blob(repo.object_hash());
+
+    assert!(
+        repo.has_object(empty_blob_id),
+        "empty object is always present even if it's not"
+    );
+    assert!(!repo.objects.contains(&empty_blob_id));
+
+    let header = repo.find_header(empty_blob_id)?;
+    assert_eq!(header.kind(), gix_object::Kind::Blob);
+    assert_eq!(header.size(), 0);
+    assert_eq!(repo.objects.try_header(&empty_blob_id)?, None);
+
+    let header = repo.try_find_header(empty_blob_id)?.expect("should find header");
+    assert_eq!(header.kind(), gix_object::Kind::Blob);
+    assert_eq!(header.size(), 0);
+
+    let obj = repo.find_object(empty_blob_id)?;
+    assert_eq!(obj.kind, gix_object::Kind::Blob);
+    assert_eq!(obj.data.len(), 0);
+
+    let mut buf = Vec::new();
+    assert_eq!(repo.objects.try_find(&empty_blob_id, &mut buf)?, None);
+
+    let obj = repo.try_find_object(empty_blob_id)?.expect("should find object");
+    assert_eq!(obj.kind, gix_object::Kind::Blob);
+    assert_eq!(obj.data.len(), 0);
+
+    let blob = obj.try_into_blob()?;
+    assert_eq!(blob.data.len(), 0);
+
+    Ok(())
 }
 
 mod tag {