prizmkit 1.1.107 → 1.1.108

package/bundled/dev-pipeline/tests/test_generate_refactor_prompt.py

@@ -7,7 +7,6 @@ from continuation import append_continuation_handoff
 from generate_refactor_prompt import (
     build_replacements,
     generate_refactor_checkpoint,
-    load_log_size_section,
     merge_refactor_checkpoint_state,
 )
 
@@ -35,7 +34,7 @@ class TestRefactorContinuationHandoff:
 
         assert prompt.startswith("REFACTOR PROMPT")
         assert "This is a fresh headless AI CLI session for the same task." in prompt
-        assert "Do not attempt `/compact` as the recovery mechanism." in prompt
+        assert "Do not attempt `/compact` as the recovery mechanism." not in prompt
         assert ("Log " + "monitor") not in prompt
         assert ("log-" + "monitor") not in prompt
         assert (tmp_path / ".prizmkit/specs/R-007/continuation-summary.md").exists()
@@ -54,19 +53,102 @@ class TestRefactorCheckpointMerge:
         assert merged["steps"][1]["status"] == "pending"
 
 
+class TestExplicitProjectRoot:
+    def test_build_replacements_infers_project_root_from_refactor_list(self, tmp_path, monkeypatch):
+        monkeypatch.delenv("PRIZMKIT_PLATFORM", raising=False)
+        project = tmp_path / "project"
+        plan_path = project / ".prizmkit" / "plans" / "refactor-list.json"
+        plan_path.parent.mkdir(parents=True)
+        (project / ".prizmkit" / "config.json").write_text(
+            '{"platform":"claude","ai_cli":"claude"}', encoding="utf-8"
+        )
+        args = Namespace(
+            refactor_list=str(plan_path),
+            refactor_id="R-001",
+            session_id="session-1",
+            run_id="run-1",
+            retry_count="0",
+            resume_phase="null",
+            state_dir=str(project / ".prizmkit" / "state" / "refactor"),
+            output=str(project / "out.md"),
+            template=None,
+            mode=None,
+            critic=None,
+        )
+        refactor = {"id": "R-001", "title": "Restructure", "description": "d"}
+
+        replacements = build_replacements(
+            args, refactor, [refactor], {}, str(Path("dev-pipeline/scripts").resolve())
+        )
+
+        assert replacements["{{PROJECT_ROOT}}"] == str(project.resolve())
+        assert replacements["{{REVIEWER_SUBAGENT_PATH}}"].endswith(
+            ".claude/agents/prizm-dev-team-reviewer.md"
+        )
+        assert replacements["{{REVIEWER_SUBAGENT_PATH}}"].startswith(str(project.resolve()))
+        assert replacements["{{SESSION_STATUS_PATH}}"] == str(
+            project / ".prizmkit" / "state" / "refactor" / "R-001" /
+            "sessions" / "session-1" / "session-status.json"
+        )
+
+    def test_build_replacements_uses_cli_project_root_for_rendered_paths(self, tmp_path, monkeypatch):
+        monkeypatch.setenv("PRIZMKIT_PLATFORM", "claude")
+        execution_root = tmp_path / "worktree"
+        main_checkout = tmp_path / "main"
+        execution_root.mkdir()
+        main_checkout.mkdir()
+        args = Namespace(
+            refactor_list=str(main_checkout / ".prizmkit" / "plans" / "refactor-list.json"),
+            refactor_id="R-001",
+            session_id="session-1",
+            run_id="run-1",
+            retry_count="0",
+            resume_phase="null",
+            state_dir=str(main_checkout / ".prizmkit" / "state" / "refactor"),
+            output=str(main_checkout / ".prizmkit" / "state" / "refactor" / "R-001" / "sessions" / "session-1" / "bootstrap-prompt.md"),
+            template=None,
+            mode=None,
+            critic=None,
+            project_root=str(execution_root),
+        )
+        refactor = {"id": "R-001", "title": "Worktree Refactor", "description": "d"}
+
+        replacements = build_replacements(
+            args,
+            refactor,
+            [refactor],
+            {},
+            str(main_checkout / "dev-pipeline" / "scripts"),
+        )
+
+        assert replacements["{{PROJECT_ROOT}}"] == str(execution_root.resolve())
+        assert replacements["{{SESSION_STATUS_PATH}}"] == str(
+            main_checkout / ".prizmkit" / "state" / "refactor" / "R-001" /
+            "sessions" / "session-1" / "session-status.json"
+        )
+        assert not replacements["{{SESSION_STATUS_PATH}}"].startswith(str(execution_root.resolve()))
+        assert replacements["{{DEV_SUBAGENT_PATH}}"].startswith(str(execution_root.resolve()))
+        assert "### Checkpoint Update Helper" in replacements["{{CHECKPOINT_SYSTEM}}"]
+        assert "python3 .prizmkit/dev-pipeline/scripts/update-checkpoint.py" in replacements["{{CHECKPOINT_SYSTEM}}"]
+        assert "pipeline progress mechanism" not in replacements["{{CHECKPOINT_SYSTEM}}"]
+
+
 class TestRefactorPromptMaxLogRemoval:
-    def test_refactor_log_size_section_is_checkpoint_guidance_without_max_log_env(self):
-        content = load_log_size_section(str(Path("dev-pipeline/scripts").resolve()))
-
-        assert "passive checkpoint guidance" in content
-        assert "Headless recovery must rely on durable checkpoints" in content
-        assert "fresh continuation session" in content
-        assert REMOVED_MAX_LOG_ENV not in content
-        assert REMOVED_MAX_LOG_HUMAN not in content
-        assert ("maximum-" + "log-size") not in content
-        assert ("log size " + "limit") not in content.lower()
-
-    def test_refactor_replacements_keep_session_log_without_max_log_placeholders(self):
+    def test_refactor_source_has_no_deleted_log_size_injection(self):
+        deleted_section = "log-size-" + "awareness.md"
+        loader = "load_" + "log_size_section"
+        placeholder = "{{LOG_" + "SIZE_AWARENESS}}"
+        deleted_title = "Session " + "Checkpoint Awareness"
+        source = Path("dev-pipeline/scripts/generate-refactor-prompt.py").read_text(encoding="utf-8")
+        template = Path("dev-pipeline/templates/refactor-bootstrap-prompt.md").read_text(encoding="utf-8")
+
+        assert (Path("dev-pipeline/templates/sections") / deleted_section).exists() is False
+        for content in (source, template):
+            for forbidden in (deleted_section, loader, placeholder, deleted_title):
+                assert forbidden not in content
+
+    def test_refactor_replacements_keep_session_log_without_max_log_placeholders(self, monkeypatch):
+        monkeypatch.setenv("PRIZMKIT_PLATFORM", "claude")
         args = Namespace(
             refactor_id="R-001",
             session_id="session-1",
@@ -90,3 +172,24 @@ class TestRefactorPromptMaxLogRemoval:
         )
         assert "{{" + REMOVED_MAX_LOG_ENV + "}}" not in replacements
         assert "{{" + REMOVED_MAX_LOG_HUMAN + "}}" not in replacements
+        assert "{{LOG_" + "SIZE_AWARENESS}}" not in replacements
+
+
+HEADLESS_FORBIDDEN_PROMPT_STRINGS = [
+    "### Interactive Claude Code Note",
+    "### Project Conventions",
+    "Scaffold / generated file awareness",
+    "### Resume Behavior",
+    "[NEEDS CLARIFICATION]",
+    "pause for interactive input",
+    "prizmkit-test-gate.py",
+    "/compact",
+    "Session " + "Checkpoint Awareness",
+    "{{LOG_" + "SIZE_AWARENESS}}",
+]
+
+
+def test_template_source_is_clean_for_headless_prompt_contract():
+    content = Path("dev-pipeline/templates/refactor-bootstrap-prompt.md").read_text(encoding="utf-8")
+    for forbidden in HEADLESS_FORBIDDEN_PROMPT_STRINGS:
+        assert forbidden not in content

package/bundled/dev-pipeline/tests/test_python_runner_parity.py

@@ -604,6 +604,7 @@ def _install_runner_session_fakes(monkeypatch, runners, *, statuses=("success",)
 
     prompt_calls = []
     launch_cwds = []
+    launch_configs = []
     updates = []
     cleanup_calls = []
     merge_calls = []
@@ -621,6 +622,7 @@ def _install_runner_session_fakes(monkeypatch, runners, *, statuses=("success",)
     )
     monkeypatch.setattr(runners, "detect_stream_json_support", lambda command, platform: SimpleNamespace(enabled=False))
     monkeypatch.setattr(runners, "start_item", lambda *args, **kwargs: SimpleNamespace(ok=True, text="", data={"new_status": "in_progress"}))
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", lambda *args, **kwargs: SimpleNamespace(attempted=False, committed=False, ok=True, result=None))
     monkeypatch.setattr(runners, "commit_pre_branch_bookkeeping", lambda *args, **kwargs: None)
     monkeypatch.setattr(runners, "commit_final_bookkeeping", lambda *args, **kwargs: None)
     monkeypatch.setattr(runners, "current_branch", lambda _root: "main")
@@ -652,6 +654,7 @@ def _install_runner_session_fakes(monkeypatch, runners, *, statuses=("success",)
 
         def run(self):
             launch_cwds.append(self.config.cwd)
+            launch_configs.append(self.config)
             return SimpleNamespace(progress_summary=SimpleNamespace(), fatal_error_code="")
 
     def fake_classify(*args, **kwargs):
@@ -690,6 +693,7 @@ def _install_runner_session_fakes(monkeypatch, runners, *, statuses=("success",)
     return SimpleNamespace(
         prompt_calls=prompt_calls,
         launch_cwds=launch_cwds,
+        launch_configs=launch_configs,
         updates=updates,
         cleanup_calls=cleanup_calls,
         merge_calls=merge_calls,
@@ -697,6 +701,36 @@ def _install_runner_session_fakes(monkeypatch, runners, *, statuses=("success",)
     )
 
 
+
+def test_runner_launch_config_preserves_artifact_paths_live_output_and_split_thresholds(monkeypatch, tmp_path):
+    from prizmkit_runtime import runners
+    from prizmkit_runtime.runner_models import family_for, parse_invocation
+
+    monkeypatch.setenv("PRIZMKIT_LIVE_OUTPUT", "0")
+    monkeypatch.setenv("HEARTBEAT_INTERVAL", "7")
+    monkeypatch.setenv("HEARTBEAT_STALE_THRESHOLD", "11")
+    monkeypatch.setenv("STALE_KILL_THRESHOLD", "13")
+    paths = _make_paths(tmp_path / "artifact-wiring")
+    family = family_for("feature", paths)
+    invocation = parse_invocation(family, "run", ())
+    invocation = invocation.__class__(**{**invocation.__dict__, "list_path": paths.feature_plan})
+    observed = _install_runner_session_fakes(monkeypatch, runners)
+    monkeypatch.setattr(runners, "run_git_plan", lambda *args, **kwargs: SimpleNamespace(ok=True, reason="", results=()))
+
+    status = runners._process_item(family, invocation, "F-001", paths, initial_metadata={})
+
+    config = observed.launch_configs[0]
+    expected_log = paths.feature_state_dir / "F-001" / "sessions" / "F-001-session-1" / "logs" / "session.log"
+    assert status == "completed"
+    assert config.live_output is False
+    assert config.live_progress_interval_seconds == 7
+    assert config.heartbeat_stale_threshold_seconds == 11
+    assert config.stale_kill_threshold_seconds == 13
+    assert config.log_path == expected_log
+    assert config.progress_path == expected_log.with_name("progress.json")
+    assert config.heartbeat_path == expected_log.with_name("heartbeat.json")
+
+
 @pytest.mark.parametrize(("kind", "item_id", "branch"), [("feature", "F-001", "dev/F-001-test"), ("bugfix", "B-001", "bugfix/B-001-test")])
 def test_feature_and_bugfix_explicit_worktree_launches_from_worktree_and_renders_worktree_root(monkeypatch, tmp_path, kind, item_id, branch):
     from prizmkit_runtime import runners
@@ -710,18 +744,26 @@ def test_feature_and_bugfix_explicit_worktree_launches_from_worktree_and_renders
     invocation = invocation.__class__(**{**invocation.__dict__, "list_path": paths.feature_plan if kind == "feature" else paths.bugfix_plan})
     observed = _install_runner_session_fakes(monkeypatch, runners)
     setup_calls = []
+    lifecycle_order = []
 
     def fake_setup(runtime):
+        lifecycle_order.append("worktree_setup")
         setup_calls.append(runtime)
         runtime.worktree_path.mkdir(parents=True, exist_ok=True)
         return SimpleNamespace(ok=True, runtime=runtime, reason="")
 
+    def fake_preflight(*args, **kwargs):
+        lifecycle_order.append("preflight")
+        return SimpleNamespace(attempted=False, committed=False, ok=True, result=None)
+
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", fake_preflight)
     monkeypatch.setattr(runners, "ensure_linked_worktree", fake_setup, raising=False)
 
     status = runners._process_item(family, invocation, item_id, paths, initial_metadata={})
 
     expected_worktree = paths.state_dir / "worktrees" / branch.replace("/", "-")
     assert status == "completed"
+    assert lifecycle_order[:2] == ["preflight", "worktree_setup"]
     assert setup_calls[0].worktree_path == expected_worktree.resolve()
     assert observed.launch_cwds == [expected_worktree.resolve()]
     assert observed.prompt_calls[0]["project_root"] == paths.project_root
@@ -842,6 +884,7 @@ def test_explicit_worktree_materializes_ignored_support_assets_before_prompt_and
     )
     monkeypatch.setattr(runners, "detect_stream_json_support", lambda command, platform: SimpleNamespace(enabled=False))
     monkeypatch.setattr(runners, "start_item", lambda *args, **kwargs: SimpleNamespace(ok=True, text="", data={"new_status": "in_progress"}))
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", lambda *args, **kwargs: SimpleNamespace(attempted=False, committed=False, ok=True, result=None))
     monkeypatch.setattr(runners, "commit_pre_branch_bookkeeping", lambda *args, **kwargs: None)
     monkeypatch.setattr(runners, "commit_final_bookkeeping", lambda *args, **kwargs: None)
     monkeypatch.setattr(runners, "_session_id", lambda item_id: f"{item_id}-session")
@@ -921,6 +964,11 @@ def test_default_no_worktree_preserves_main_worktree_branch_launch(monkeypatch,
         lifecycle_order.append("start_item")
         return SimpleNamespace(ok=True, text="", data={"new_status": "in_progress"})
 
+    def fake_preflight(*args, **kwargs):
+        lifecycle_order.append("preflight")
+        return SimpleNamespace(attempted=False, committed=False, ok=True, result=None)
+
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", fake_preflight)
     monkeypatch.setattr(runners, "run_git_plan", fake_run_git_plan)
     monkeypatch.setattr(runners, "start_item", fake_start_item)
     monkeypatch.setattr(runners, "ensure_linked_worktree", lambda runtime: (_ for _ in ()).throw(AssertionError("worktree setup should not run")), raising=False)
@@ -932,7 +980,7 @@ def test_default_no_worktree_preserves_main_worktree_branch_launch(monkeypatch,
     assert observed.prompt_calls[0].get("execution_root") == paths.project_root
     assert "branch_create" in plan_names
     assert "branch_merge" in plan_names
-    assert lifecycle_order[:2] == ["branch_create", "start_item"]
+    assert lifecycle_order[:3] == ["preflight", "branch_create", "start_item"]
     assert branch_events == [("create", "dev/F-001-test"), ("merge", "dev/F-001-test")]
 
 
@@ -1024,6 +1072,7 @@ def test_no_worktree_recovers_completed_pending_branch_before_new_session(monkey
     monkeypatch.setattr(runners, "has_branch_completion_evidence", lambda *args, **kwargs: True)
     monkeypatch.setattr(runners, "commit_final_bookkeeping", lambda *args, **kwargs: None)
     monkeypatch.setattr(runners, "commit_task_branch_changes", lambda *args, **kwargs: None)
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", lambda *args, **kwargs: SimpleNamespace(attempted=False, committed=False, ok=True, result=None))
 
     def fake_update(family, invocation, item_id, session_status, session_id, project_root, **metadata):
         updates.append((session_status, session_id, metadata))
@@ -1076,6 +1125,7 @@ def test_worktree_recovers_completed_pending_branch_from_base_session_history(mo
     monkeypatch.setattr(runners, "generate_prompt", lambda *args, **kwargs: (_ for _ in ()).throw(AssertionError("should not launch prompt")))
     monkeypatch.setattr(runners, "commit_final_bookkeeping", lambda *args, **kwargs: None)
     monkeypatch.setattr(runners, "commit_task_branch_changes", lambda *args, **kwargs: None)
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", lambda *args, **kwargs: SimpleNamespace(attempted=False, committed=False, ok=True, result=None))
 
     def fake_update(family, invocation, item_id, session_status, session_id, project_root, **metadata):
         updates.append((session_status, session_id, project_root, metadata))
@@ -1187,6 +1237,33 @@ def test_no_worktree_runner_ignores_recovery_side_effect_branch(monkeypatch, tmp
     assert observed.updates[-1][2]["base_branch"] == "main"
 
 
+def test_no_worktree_preflight_commit_failure_aborts_before_branch_or_status(monkeypatch, tmp_path):
+    from prizmkit_runtime import runners
+    from prizmkit_runtime.runner_models import family_for, parse_invocation
+
+    monkeypatch.setenv("DEV_BRANCH", "dev/F-001-preflight-fail")
+    paths = _make_paths(tmp_path)
+    family = family_for("feature", paths)
+    invocation = parse_invocation(family, "run", ())
+    invocation = invocation.__class__(**{**invocation.__dict__, "list_path": paths.feature_plan})
+    events = []
+
+    monkeypatch.setattr(
+        runners,
+        "load_runtime_config",
+        lambda paths: SimpleNamespace(ai_client=SimpleNamespace(command="fake-ai", platform="codebuddy"), model="", effort=None),
+    )
+    monkeypatch.setattr(runners, "current_branch", lambda _root: "main")
+    monkeypatch.setattr(runners, "commit_preflight_ready_changes", lambda *args, **kwargs: SimpleNamespace(attempted=True, committed=False, result=SimpleNamespace(stderr="commit failed\n", stdout="")))
+    monkeypatch.setattr(runners, "run_git_plan", lambda *args, **kwargs: events.append("branch_create"))
+    monkeypatch.setattr(runners, "start_item", lambda *args, **kwargs: events.append("start_item"))
+
+    with pytest.raises(RuntimeError, match="Preflight dirty workspace preservation failed"):
+        runners._process_item(family, invocation, "F-001", paths, initial_metadata={})
+
+    assert events == []
+
+
 def test_foreground_runner_emits_structured_feature_progress(monkeypatch, tmp_path, capsys):
     from prizmkit_runtime import runners
     from prizmkit_runtime.runner_models import family_for, parse_invocation
@@ -1212,7 +1289,7 @@ def test_foreground_runner_emits_structured_feature_progress(monkeypatch, tmp_pa
     assert "────────────────────────────────────────────────────" in captured.err
     assert "Feature: F-001 — Readable output" in captured.err
     assert "Code retry: 1 / 3" in captured.err
-    assert "Infrastructure retry: 2 / 3" in captured.err
+    assert "Preflight: no user-visible changes to preserve before feature task F-001" in captured.err
     assert "Pipeline mode: lite (Tier 1 — Single Agent)" in captured.err
     assert "Agents: 1 (critic: disabled)" in captured.err
     assert "Spawning AI CLI session: F-001-session-1" in captured.err