backend: create BackendError::InvalidHashLength

Strictly speaking, we could rely on e.g. `git2::Oid::from_str` to produce an error, but I figure that having an explicit error for a mismatching hash length might demystify some error condition in the future, since commit IDs and change IDs and potentially other backends' IDs may have different lengths, so this could flag a mismatch earlier/more obviously.

lib/src/backend.rs

@@ -173,6 +173,16 @@ content_hash! {
 
 #[derive(Debug, Error, PartialEq, Eq)]
 pub enum BackendError {
+    #[error(
+        "Invalid hash for object of type {object_type}: {hash} (expected {expected} bytes, got \
+         {actual} bytes)"
+    )]
+    InvalidHashLength {
+        expected: usize,
+        actual: usize,
+        object_type: &'static str,
+        hash: String,
+    },
     #[error("Object not found")]
     NotFound,
     #[error("Error: {0}")]

lib/src/git_backend.rs

@@ -180,7 +180,12 @@ impl Backend for GitBackend {
 
     fn read_file(&self, _path: &RepoPath, id: &FileId) -> BackendResult<Box<dyn Read>> {
         if id.as_bytes().len() != self.hash_length() {
-            return Err(BackendError::NotFound);
+            return Err(BackendError::InvalidHashLength {
+                expected: self.hash_length(),
+                actual: id.as_bytes().len(),
+                object_type: "file",
+                hash: id.hex(),
+            });
         }
         let locked_repo = self.repo.lock().unwrap();
         let blob = locked_repo
@@ -200,7 +205,12 @@ impl Backend for GitBackend {
 
     fn read_symlink(&self, _path: &RepoPath, id: &SymlinkId) -> Result<String, BackendError> {
         if id.as_bytes().len() != self.hash_length() {
-            return Err(BackendError::NotFound);
+            return Err(BackendError::InvalidHashLength {
+                expected: self.hash_length(),
+                actual: id.as_bytes().len(),
+                object_type: "symlink",
+                hash: id.hex(),
+            });
         }
         let locked_repo = self.repo.lock().unwrap();
         let blob = locked_repo
@@ -229,7 +239,12 @@ impl Backend for GitBackend {
             return Ok(Tree::default());
         }
         if id.as_bytes().len() != self.hash_length() {
-            return Err(BackendError::NotFound);
+            return Err(BackendError::InvalidHashLength {
+                expected: self.hash_length(),
+                actual: id.as_bytes().len(),
+                object_type: "tree",
+                hash: id.hex(),
+            });
         }
 
         let locked_repo = self.repo.lock().unwrap();
@@ -348,7 +363,12 @@ impl Backend for GitBackend {
 
     fn read_commit(&self, id: &CommitId) -> BackendResult<Commit> {
         if id.as_bytes().len() != self.hash_length() {
-            return Err(BackendError::NotFound);
+            return Err(BackendError::InvalidHashLength {
+                expected: self.hash_length(),
+                actual: id.as_bytes().len(),
+                object_type: "commit",
+                hash: id.hex(),
+            });
         }
 
         if *id == self.root_commit_id {

lib/src/revset.rs

@@ -86,6 +86,9 @@ fn resolve_full_commit_id(
     symbol: &str,
 ) -> Result<Option<Vec<CommitId>>, RevsetError> {
     if let Ok(binary_commit_id) = hex::decode(symbol) {
+        if repo.store().hash_length() != binary_commit_id.len() {
+            return Ok(None);
+        }
         let commit_id = CommitId::new(binary_commit_id);
         match repo.store().get_commit(&commit_id) {
             Ok(_) => Ok(Some(vec![commit_id])),