Refactor WorkspaceManager.sanitize_repo to stash changes and raise errors on failure, add unit tests

gitea/workspace.py

@@ -67,6 +67,19 @@ class WorkspaceManager:
                 check=True, capture_output=True,
             )
             self._configure_repo_user(repo_path)
+
+            # Check for any uncommitted changes or untracked files
+            status_res = subprocess.run(
+                ["git", "-C", str(repo_path), "status", "--porcelain"],
+                check=True, capture_output=True, text=True
+            )
+            if status_res.stdout.strip():
+                logger.info(f"Uncommitted changes detected in {repo_path}. Stashing before sanitization.")
+                subprocess.run(
+                    ["git", "-C", str(repo_path), "stash", "push", "-u", "-m", "Auto-backup before agent sanitization"],
+                    check=True, capture_output=True
+                )
+
             subprocess.run(
                 ["git", "-C", str(repo_path), "reset", "--hard", "HEAD"],
                 check=True, capture_output=True,
@@ -96,7 +109,8 @@ class WorkspaceManager:
                     check=True, capture_output=True,
                 )
         except Exception as e:
-            logger.error(f"Error during sanitization: {e}")
+            logger.error(f"Error during sanitization: {e}", exc_info=True)
+            raise RuntimeError(f"Failed to sanitize repository {repo_full_name} at {repo_path}: {e}") from e
 
     def clone_repo(self, repo_full_name: str, clone_url: str | None = None) -> Path:
         repo_path: Path = self.get_repo_path(repo_full_name)

tests/test_workspace.py

@@ -1,4 +1,5 @@
 from pathlib import Path
+from typing import Any
 from unittest.mock import MagicMock, patch
 import pytest
 
@@ -95,3 +96,111 @@ def test_workspace_manager_fails_if_authenticated_user_has_no_login(
 
     with pytest.raises(RuntimeError, match="No authenticated user found."):
         workspace._configure_repo_user(Path("/tmp/mock-repo"))
+
+
+@patch("gitea.workspace.subprocess.run")
+@patch("gitea.workspace.GiteaClient")
+def test_workspace_manager_sanitize_repo_no_changes(
+    mock_client_class: MagicMock,
+    mock_run: MagicMock
+) -> None:
+    # Setup Gitea client mock
+    mock_client = MagicMock()
+    mock_client_class.return_value = mock_client
+    mock_user = MagicMock()
+    mock_user.full_name = "Agent Tester"
+    mock_user.login = "agent-test"
+    mock_user.email = "agent-test@example.com"
+    mock_client.get_authenticated_user.return_value = mock_user
+
+    # Mock subprocess.run for status check and others
+    def mock_run_side_effect(args: list[str], **kwargs: Any) -> MagicMock:
+        result = MagicMock()
+        result.returncode = 0
+        if "status" in args:
+            result.stdout = ""
+        else:
+            result.stdout = "some output"
+        return result
+
+    mock_run.side_effect = mock_run_side_effect
+
+    workspace = WorkspaceManager()
+    repo_path = Path("/tmp/mock-repo")
+
+    workspace.sanitize_repo("meeks/repo1", repo_path)
+
+    # Verify that stash was not called
+    stash_calls = [call for call in mock_run.call_args_list if "stash" in call[0][0]]
+    assert len(stash_calls) == 0
+
+    # Verify other expected git calls
+    reset_calls = [call for call in mock_run.call_args_list if "reset" in call[0][0]]
+    clean_calls = [call for call in mock_run.call_args_list if "clean" in call[0][0]]
+    assert len(reset_calls) > 0
+    assert len(clean_calls) > 0
+
+
+@patch("gitea.workspace.subprocess.run")
+@patch("gitea.workspace.GiteaClient")
+def test_workspace_manager_sanitize_repo_with_changes(
+    mock_client_class: MagicMock,
+    mock_run: MagicMock
+) -> None:
+    # Setup Gitea client mock
+    mock_client = MagicMock()
+    mock_client_class.return_value = mock_client
+    mock_user = MagicMock()
+    mock_user.full_name = "Agent Tester"
+    mock_user.login = "agent-test"
+    mock_user.email = "agent-test@example.com"
+    mock_client.get_authenticated_user.return_value = mock_user
+
+    # Mock subprocess.run to show modified files
+    def mock_run_side_effect(args: list[str], **kwargs: Any) -> MagicMock:
+        result = MagicMock()
+        result.returncode = 0
+        if "status" in args:
+            result.stdout = " M file.py\n?? untracked.py\n"
+        else:
+            result.stdout = ""
+        return result
+
+    mock_run.side_effect = mock_run_side_effect
+
+    workspace = WorkspaceManager()
+    repo_path = Path("/tmp/mock-repo")
+
+    workspace.sanitize_repo("meeks/repo1", repo_path)
+
+    # Verify stash push was called
+    stash_calls = [call for call in mock_run.call_args_list if "stash" in call[0][0]]
+    assert len(stash_calls) == 1
+    assert "push" in stash_calls[0][0][0]
+    assert "-u" in stash_calls[0][0][0]
+
+
+@patch("gitea.workspace.subprocess.run")
+@patch("gitea.workspace.GiteaClient")
+def test_workspace_manager_sanitize_repo_fails(
+    mock_client_class: MagicMock,
+    mock_run: MagicMock
+) -> None:
+    # Setup Gitea client mock
+    mock_client = MagicMock()
+    mock_client_class.return_value = mock_client
+    mock_user = MagicMock()
+    mock_user.full_name = "Agent Tester"
+    mock_user.login = "agent-test"
+    mock_user.email = "agent-test@example.com"
+    mock_client.get_authenticated_user.return_value = mock_user
+
+    # Mock remote set-url to fail
+    import subprocess
+    mock_run.side_effect = subprocess.CalledProcessError(1, "git remote set-url")
+
+    workspace = WorkspaceManager()
+    repo_path = Path("/tmp/mock-repo")
+
+    with pytest.raises(RuntimeError, match="Failed to sanitize repository"):
+        workspace.sanitize_repo("meeks/repo1", repo_path)