Harden work-item failure recovery

tests/test_cli_e2e.py

@@ -43,6 +43,8 @@ if args[0] == "api":
         issue = int(endpoint.split("/issues/")[1].split("/")[0])
         blockers = state["dependencies"].get(str(issue), [])
         out([next(x for x in state["issues"] if x["number"] == number) for number in blockers])
+    elif endpoint.endswith("/pulls?state=open&limit=50"):
+        out([pr for pr in state["prs"].values() if not pr["merged"]])
     elif "/pulls/" in endpoint:
         number = int(endpoint.rsplit("/", 1)[1])
         out(state["prs"][str(number)])

tests/test_core.py

@@ -174,5 +174,54 @@ class StateTests(RepositoryFixture):
         self.assertFalse(workflow.state_path.exists())
 
 
+class RecoveryTests(RepositoryFixture):
+    def story(self) -> Story:
+        return Story(1, "US01-01", "Recover Workflow", 1, 1, frozenset(), (), {})
+
+    def test_failed_remote_claim_rolls_back_created_branch(self) -> None:
+        workflow = Workflow(self.repo, self.config)
+        workflow.gitea = Mock()
+        workflow.gitea.next_story.return_value = self.story()
+        workflow.gitea.claim.side_effect = WorkItemError("remote claim failed")
+        with self.assertRaisesRegex(WorkItemError, "remote claim failed"):
+            workflow.claim()
+        self.assertEqual(workflow.git.current_branch(), "main")
+        self.assertNotIn("us/US01-01", workflow.git.git("branch"))
+        self.assertFalse(workflow.state_path.exists())
+
+    def test_pr_creation_failure_resumes_after_push_without_recommit(self) -> None:
+        config = config_for("domverse/photoanalyzer", max_file_bytes=1000)
+        workflow = Workflow(self.repo, config)
+        story = self.story()
+        branch = workflow.git.create_story_branch(story)
+        workflow.save_state({
+            "issue": 1,
+            "story_id": story.story_id,
+            "title": story.title,
+            "branch": branch,
+            "status": "in-progress",
+        })
+        (self.repo / "feature.txt").write_text("implemented\n", encoding="utf-8")
+        workflow.gitea = Mock()
+        workflow.gitea.find_open_pr.return_value = None
+        workflow.gitea.create_pr.side_effect = [
+            WorkItemError("response lost"),
+            {"number": 9, "index": 9, "url": "https://example.test/pr/9"},
+        ]
+
+        with self.assertRaisesRegex(WorkItemError, "response lost"):
+            workflow.submit(("true",), confirmed=True)
+        pushed = workflow.load_state()
+        self.assertEqual(pushed["status"], "pushed")
+        commit = pushed["commit"]
+        self.assertEqual(workflow.git.git("status", "--porcelain"), "")
+
+        result = workflow.submit((), confirmed=True)
+        self.assertEqual(result["status"], "review")
+        self.assertEqual(result["commit"], commit)
+        self.assertEqual(workflow.gitea.create_pr.call_count, 2)
+        workflow.gitea.mark_review.assert_called_once()
+
+
 if __name__ == "__main__":
     unittest.main()

work_item/core.py

@@ -322,6 +322,13 @@ class Gitea:
         except json.JSONDecodeError as error:
             raise WorkItemError(f"Could not parse created pull request: {output}") from error
 
+    def find_open_pr(self, branch: str) -> dict[str, Any] | None:
+        pulls = self.api(f"{self.base}/pulls?state=open&limit=50")
+        for pull in pulls:
+            if (pull.get("head") or {}).get("ref") == branch:
+                return pull
+        return None
+
     def pr(self, number: int) -> dict[str, Any]:
         return self.api(f"{self.base}/pulls/{number}")
 
@@ -399,7 +406,12 @@ class Workflow:
         self.git.update_main()
         story = self.gitea.next_story()
         branch = self.git.create_story_branch(story)
+        try:
             self.gitea.claim(story)
+        except Exception:
+            self.git.git("switch", self.config.main_branch)
+            self.git.git("branch", "-D", branch)
+            raise
         state = {
             "issue": story.number,
             "story_id": story.story_id,
@@ -421,6 +433,15 @@ class Workflow:
         state = self.load_state()
         if self.git.current_branch() != state["branch"]:
             raise WorkItemError(f"Expected branch {state['branch']}, found {self.git.current_branch()}")
+        story = Story(
+            number=int(state["issue"]), story_id=state["story_id"], title=state["title"],
+            epic=int(state["story_id"][2:4]), sequence=int(state["story_id"][5:7]),
+            labels=frozenset(), assignees=(), raw={},
+        )
+        if state.get("status") == "pushed":
+            commands = tuple(state["tests"])
+            commit = state["commit"]
+        elif state.get("status") == "in-progress":
             commands = tuple(dict.fromkeys((*self.config.required_tests, *tests)))
             if not commands:
                 raise WorkItemError("At least one automated test command is required")
@@ -434,12 +455,14 @@ class Workflow:
                     "Review the changes and rerun with --yes to commit:\n"
                     f"{summary}\nFiles: {', '.join(path.as_posix() for path in paths)}"
                 )
-        story = Story(
-            number=int(state["issue"]), story_id=state["story_id"], title=state["title"],
-            epic=int(state["story_id"][2:4]), sequence=int(state["story_id"][5:7]),
-            labels=frozenset(), assignees=(), raw={},
-        )
             commit = self.git.stage_commit_push(story, state["branch"])
+            state.update({"status": "pushed", "commit": commit, "tests": list(commands)})
+            self.save_state(state)
+        else:
+            raise WorkItemError(f"Story cannot be submitted from status {state.get('status')!r}")
+
+        pr = self.gitea.find_open_pr(state["branch"])
+        if pr is None:
             pr = self.gitea.create_pr(story, state["branch"], commands, commit)
         pr_number = int(pr.get("index") or pr.get("number"))
         state.update({"status": "review", "pr": pr_number, "commit": commit, "tests": list(commands)})