fix: partial clone failure [DE-921] (#3970)

Signed-off-by: Mouad BANI <mouad-mb@outlook.com>

Author: mbani01

services/apps/git_integration/src/crowdgit/services/clone/clone_service.py

@@ -17,6 +17,7 @@
     get_default_branch,
     get_remote_default_branch,
     get_repo_name,
+    normalize_gerrit_remote,
     run_shell_command,
 )
 
@@ -370,6 +371,7 @@ async def clone_batches_generator(
         error_message = None
         total_execution_time = 0.0
         remote = repository.url.removesuffix(".git")
+        clone_remote = normalize_gerrit_remote(remote)
 
         batch_info = CloneBatchInfo(
             repo_path=temp_repo_path,
@@ -382,10 +384,10 @@ async def clone_batches_generator(
             batch_start_time = time.time()
 
             clone_with_batches = await self.determine_clone_strategy(
-                temp_repo_path, remote, repository.branch, repository.last_processed_commit
+                temp_repo_path, clone_remote, repository.branch, repository.last_processed_commit
             )
             if clone_with_batches:
-                batch_depth = await self._calculate_batch_depth(temp_repo_path, remote)
+                batch_depth = await self._calculate_batch_depth(temp_repo_path, clone_remote)
             await self._update_batch_info(
                 batch_info, temp_repo_path, repository.last_processed_commit, clone_with_batches
             )

services/apps/git_integration/src/crowdgit/services/utils.py

@@ -13,6 +13,20 @@
 from crowdgit.logger import logger
 
 
+def normalize_gerrit_remote(url: str) -> str:
+    """Rewrite Gerrit remote URLs to their Gitea equivalents for git clone/fetch operations.
+
+    Gerrit servers don't support the git shallow protocol extension (--deepen),
+    causing HTTP 500 errors during incremental fetching. Where a Gitea mirror
+    exists with full protocol support (e.g. opendev.org mirrors review.opendev.org),
+    we rewrite the host so that shallow cloning works.
+
+    Only used for git operations, not for URLs persisted to the database or
+    passed to other services.
+    """
+    return url.replace("https://review.opendev.org", "https://opendev.org")
+
+
 def safe_decode(data: bytes) -> str:
     """
     Safely decode bytes to string, handling various encodings that might be present in git output.