Stabilize local dev merge

Align regression tests with the current Odysseus behavior after merging origin/dev into local main.

- keep phone/name-only contacts valid and cover null email without crashes

- pin explicit web-search false form submission in chat.js

- update Cookbook dependency/download completion tests for combined live + persisted output

- expose SGLang OS package repair hints from backend diagnosis

- treat MLX and MLX-community repos as servable on Apple Metal while keeping CUDA behavior unchanged

- keep desktop new-chat coverage on the shared preferred-model helper

- remove a hardcoded crop overlay portal z-index literal

- include the local agent-loop cleanup that removes the old manage_notes reminder repair shim

Verified with: docker run --rm -v /home/pewds/odysseus-cookbook-fresh:/app -w /app odysseus-cookbook-fresh-odysseus python3 -m pytest -q (4515 passed, 4 skipped).

tests/test_chat_route_tool_policy.py

@@ -254,6 +254,6 @@ def test_frontend_always_sends_explicit_allow_bash():
 def test_frontend_sends_explicit_allow_web_search_false_in_agent_mode():
     """chat.js must send allow_web_search=false when web toggle is off in agent mode."""
     source = _CHAT_JS.read_text(encoding="utf-8")
-    assert "allow_web_search', 'false'" in source, (
+    assert "fd.append('allow_web_search', el('web-toggle').checked ? 'true' : 'false')" in source, (
         "Frontend must send explicit allow_web_search=false in agent mode when toggle is off"
     )

tests/test_contacts_add_null_name.py

@@ -24,7 +24,11 @@ def _add_handler():
 def _stub_store(monkeypatch):
     created = []
     monkeypatch.setattr(cr, "_fetch_contacts", lambda *a, **k: [])
-    monkeypatch.setattr(cr, "_create_contact", lambda name, email: created.append((name, email)) or True)
+    monkeypatch.setattr(
+        cr,
+        "_create_contact",
+        lambda name, email="", address="", phones=None: created.append((name, email, address, phones or [])) or True,
+    )
     return created
 
 
@@ -33,10 +37,18 @@ def test_null_name_does_not_crash(_stub_store):
     result = asyncio.run(handler({"name": None, "email": "x@y.com"}, _admin="admin"))
     assert result["success"] is True
     # name fell back to the email local-part instead of crashing.
-    assert _stub_store == [("x", "x@y.com")]
+    assert _stub_store == [("x", "x@y.com", "", [])]
 
 
-def test_null_email_is_rejected_cleanly(_stub_store):
+def test_null_email_does_not_crash(_stub_store):
     handler = _add_handler()
     result = asyncio.run(handler({"name": "Bob", "email": None}, _admin="admin"))
-    assert result == {"success": False, "error": "Email required"}
+    assert result["success"] is True
+    assert _stub_store == [("Bob", "", "", [])]
+
+
+def test_phone_only_contact_is_allowed(_stub_store):
+    handler = _add_handler()
+    result = asyncio.run(handler({"name": "Bob", "email": None, "phone": "0805412 7841"}, _admin="admin"))
+    assert result["success"] is True
+    assert _stub_store == [("Bob", "", "", ["0805412 7841"])]

tests/test_cookbook_dependency_completion_regression.py

@@ -20,7 +20,9 @@ def test_background_status_poll_reconciles_into_local_tasks():
     source = _read("static/js/cookbookRunning.js")
 
     assert "const statusById = new Map(tasks.map(t => [t.session_id, t]));" in source
-    assert "const nextStatus = live.status === 'completed'" in source
+    assert "const completedByOutput = depDone || downloadDone;" in source
+    assert "const nextStatus = completedByOutput" in source
+    assert "live.status === 'completed'" in source
     assert "? 'done'" in source
     assert ": (live.status === 'error'" in source
     assert "? 'error'" in source
@@ -75,7 +77,8 @@ def test_background_poll_recovers_done_for_stopped_dependency_install():
     downgrading the card to crashed."""
     source = _read("static/js/cookbookRunning.js")
 
-    assert "const depDone = !!task.payload?._dep && _depInstallSucceeded(task.output);" in source
+    assert "const combinedOutput = `${task.output || ''}\\n${live.output_tail || ''}`;" in source
+    assert "const depDone = !!task.payload?._dep && _depInstallSucceeded(combinedOutput);" in source
     assert "(depDone || downloadDone) ? 'done' : (task.type === 'download' ? 'crashed' : 'stopped')" in source
 
 
@@ -91,14 +94,14 @@ def test_background_poll_recovers_done_for_completed_download():
     normalized = " ".join(source.split())
     assert (
         "const downloadDone = task.type === 'download' "
-        "&& String(task.output || '').includes('DOWNLOAD_OK');"
+        "&& String(combinedOutput || '').includes('DOWNLOAD_OK');"
     ) in normalized
 
 
 def test_dependency_install_payload_keeps_env_path_for_refresh():
     source = _read("static/js/cookbook.js")
 
-    assert "env_path: _envState.envPath || ''" in source
+    assert "env_path: targetEnvPath || ''" in source
 
 
 def test_local_dependency_probe_refreshes_user_site_visibility():

tests/test_cookbook_diagnosis.py

@@ -29,7 +29,7 @@ def test_diagnose_sglang_native_dependency_errors():
     diagnosis = _diagnose_serve_output(output)
 
     assert diagnosis is not None
-    assert "SGLang native dependencies" in diagnosis["message"]
+    assert "SGLang native kernel/runtime" in diagnosis["message"]
     labels = [suggestion["label"] for suggestion in diagnosis["suggestions"]]
     assert any("libnuma-dev" in label for label in labels)
     assert any("python3.12-dev" in label for label in labels)

tests/test_cookbook_diagnosis_js.py

@@ -17,6 +17,6 @@ def test_sglang_native_dependency_diagnosis_is_exposed_to_browser():
 
     assert r"Python\.h" in source
     assert r"libnuma\.so\.1" in source
-    assert "SGLang native dependencies" in source
+    assert "SGLang native kernel/runtime" in source
     assert "libnuma-dev python3.12-dev build-essential" in source
     assert "sglang-kernel" in source

tests/test_cookbook_same_host_server_profiles_js.py

@@ -31,8 +31,8 @@ def test_selected_server_helpers_prefer_profile_key_before_host_fallback():
 
 def test_cookbook_submodules_resolve_visible_profile_selection():
     assert "_serverByVal?.(_ssv)" in DOWNLOAD
-    assert "_serverByVal?.(_envState.remoteServerKey || host)" in DOWNLOAD
-    assert "_serverByVal?.(_envState.remoteServerKey || _zh)" in DOWNLOAD
+    assert "_serverByVal?.(_envState.remoteServerKey || _envState.remoteHost || '')" in DOWNLOAD
+    assert "_serverByVal?.(zombieCandidate.remoteServerKey || zombieCandidate.payload?.remote_server_key || _zh)" in DOWNLOAD
     assert "_serverByVal(_envState.remoteServerKey || remoteHost)" in HWFIT
     assert "hk: _currentServerValue()" in HWFIT
     assert "sel.value = _currentServerValue();" in HWFIT

tests/test_hwfit_macos.py

@@ -43,13 +43,12 @@ def _fake_sysctl(brand="Apple M2 Pro", memsize_gb=32, wired_mb=None, display_jso
     return run
 
 
-def test_mlx_models_hidden_on_metal():
-    """MLX-quantized models can't be served by llama.cpp or Ollama (the only
-    Metal-capable engines Odysseus generates), so they must never be recommended
-    on Apple Silicon — even though the catalog tags them as Apple-only."""
+def test_mlx_models_visible_on_metal():
+    """Apple Silicon can serve MLX models through the MLX engine, so Cookbook
+    should surface MLX rows on Metal while still hiding them on CUDA."""
     results = rank_models(_metal_system(), limit=900)
     mlx = [m for m in results if str(m.get("quant", "")).startswith("mlx-")]
-    assert mlx == [], f"MLX models surfaced but cannot be served: {[m['name'] for m in mlx]}"
+    assert mlx, "MLX models should be available on Apple Silicon / Metal hardware"
 
 
 def _cuda_system():
@@ -65,17 +64,18 @@ def test_mlx_hidden_on_cuda_backend_unchanged():
     assert mlx == []
 
 
-def test_only_gguf_models_recommended_on_metal():
-    """llama.cpp and Ollama (the only Metal engines) need GGUF. Safetensors-only
-    repos — incl. vLLM-only AWQ/GPTQ/FP8 — can't be served on Metal, so every
-    model recommended on Apple Silicon must ship a servable GGUF."""
+def test_only_gguf_or_mlx_models_recommended_on_metal():
+    """Metal recommendations must be servable locally: either GGUF for
+    llama.cpp/Ollama or MLX for the MLX engine."""
     catalog = {m["name"]: m for m in get_models()}
     unservable = [
         r["name"] for r in rank_models(_metal_system(), limit=900)
-        if not (catalog.get(r["name"], {}).get("is_gguf")
+        if not (str(r.get("quant", "")).startswith("mlx-")
+                or str(r.get("name", "")).startswith(("mlx-community/", "lmstudio-community/"))
+                or catalog.get(r["name"], {}).get("is_gguf")
                 or catalog.get(r["name"], {}).get("gguf_sources"))
     ]
-    assert unservable == [], f"{len(unservable)} non-GGUF models on Metal, e.g. {unservable[:3]}"
+    assert unservable == [], f"{len(unservable)} non-servable models on Metal, e.g. {unservable[:3]}"
 
 
 def test_qwen_catalog_entries_point_at_verified_gguf_repos():

tests/test_hwfit_manual_backend.py

@@ -67,11 +67,9 @@ def test_manual_ram_mode_wipes_gpu_and_unified_flag():
     assert "unified_memory" not in s
 
 
-def test_simulated_metal_box_only_recommends_gguf():
+def test_simulated_metal_box_only_recommends_gguf_or_mlx():
     """End-to-end: a simulated Metal box must rank exactly like a real Mac —
-    only models shipping a servable GGUF (llama.cpp/Ollama) survive. Before
-    'metal' was accepted, this box ranked as CUDA and surfaced safetensors-only
-    repos the Mac can't serve."""
+    only locally servable GGUF or MLX models survive."""
     system = _apply_manual_hardware(
         {"backend": "cuda", "available_ram_gb": 32.0, "total_ram_gb": 64.0},
         manual_mode="gpu", manual_vram_gb="48", manual_backend="metal",
@@ -79,7 +77,9 @@ def test_simulated_metal_box_only_recommends_gguf():
     catalog = {m["name"]: m for m in get_models()}
     unservable = [
         r["name"] for r in rank_models(system, limit=900)
-        if not (catalog.get(r["name"], {}).get("is_gguf")
+        if not (str(r.get("quant", "")).startswith("mlx-")
+                or str(r.get("name", "")).startswith(("mlx-community/", "lmstudio-community/"))
+                or catalog.get(r["name"], {}).get("is_gguf")
                 or catalog.get(r["name"], {}).get("gguf_sources"))
     ]
-    assert unservable == [], f"{len(unservable)} non-GGUF models on simulated Metal, e.g. {unservable[:3]}"
+    assert unservable == [], f"{len(unservable)} non-servable models on simulated Metal, e.g. {unservable[:3]}"

tests/test_new_chat_model_preference.py

@@ -27,6 +27,11 @@ def test_new_chat_prefers_pending_and_current_model_before_default():
 def test_desktop_new_chat_actions_use_shared_preference_helper():
     source = APP_JS.read_text(encoding="utf-8")
 
+    shared_handler = _slice(
+        source,
+        "async function _handleNewChatAction",
+        "// New session button on icon rail",
+    )
     rail_handler = _slice(
         source,
         "// New session button on icon rail",
@@ -38,7 +43,8 @@ def test_desktop_new_chat_actions_use_shared_preference_helper():
         "const sidebarNewChatBtn = el('sidebar-new-chat-btn');",
     )
 
-    assert "if (await _createDirectChatFromPreferredModel()) return;" in rail_handler
-    assert "if (await _createDirectChatFromPreferredModel()) return;" in brand_handler
+    assert "if (preferModel && await _createDirectChatFromPreferredModel()) return;" in shared_handler
+    assert "await _handleNewChatAction();" in rail_handler
+    assert "await _handleNewChatAction();" in brand_handler
     assert "const dc = await _refreshDefaultChat();" not in rail_handler
     assert "const dc = await _refreshDefaultChat();" not in brand_handler