fix(sessions): invalidate session-list cache on CLI-visibility toggle + de-flake node test timeouts

ROOT FIX (product bug): POST /api/settings never invalidated the session-list
cache, so toggling show_cli_sessions/show_cron_sessions could serve a stale
sidebar for up to the cache TTL when the settings-file mtime stamp collided with
a cached entry. The handler now clears the cache directly on those toggles. This
is also the root cause of the recurring test_gateway_sync CI flake (session
absent from /api/sessions right after the visibility toggle).

DEFENSE-IN-DEPTH: cleanup_test_sessions fixture now resets show_cli_sessions to
default BEFORE each test too, not only in teardown.

FLAKE FIX (test infra): bump 31 node-subprocess correctness-harness timeouts from
10s -> 30s across 23 test files. Node spawn + V8 warmup under parallel CI shard
load can exceed 10s, causing TimeoutExpired flakes (e.g. test_1325 on #4445).
Only real node subprocess.run() sites changed — mock signatures, HTTP urlopen,
and deliberate timeout-behavior tests (test_api_timeout, timeout=0/1/2) untouched.

CHANGELOG.md

@@ -3,6 +3,12 @@
 
 ## [Unreleased]
 
+## [v0.51.502] — 2026-06-18 — Release RL (CLI-session visibility toggle invalidates the sidebar cache)
+
+### Fixed
+
+- **Toggling "show CLI/messaging sessions" now takes effect immediately instead of after a brief delay.** `POST /api/settings` did not invalidate the session-list cache, which is keyed partly on the visibility setting and otherwise only revalidated by the settings-file mtime stamp. When a visibility toggle wrote the settings file within the same mtime granularity as a cached entry (and resolved to the default-valued cache key), the sidebar could keep serving the stale session set for up to the cache TTL (~2.5s). The settings handler now clears the session-list cache directly whenever `show_cli_sessions` / `show_cron_sessions` change. This was also the root cause of an intermittent `test_gateway_sync` CI flake (a freshly-inserted CLI/gateway session occasionally absent from `/api/sessions` right after the visibility toggle).
+
 ## [v0.51.501] — 2026-06-18 — Release RK (Ctrl/Cmd+, opens Settings)
 
 ### Added

api/routes.py

@@ -9941,6 +9941,24 @@ def handle_post(handler, parsed) -> bool:
         saved = save_settings(body)
         saved.pop("password_hash", None)  # never expose hash to client
 
+        # Settings that change which sessions appear in the sidebar must
+        # invalidate the session-list cache directly. Relying on the cache's
+        # settings-file mtime stamp is fragile: a toggle that writes the
+        # settings file within the same mtime granularity as a cached entry (and
+        # produces the default-valued key, e.g. show_cli_sessions back to its
+        # True default) can leave a stale row set served for up to the cache TTL.
+        # This is the root cause of the intermittent gateway_sync test flake
+        # (a freshly-inserted CLI/gateway session occasionally absent from
+        # /api/sessions right after the visibility toggle). Invalidate explicitly.
+        if any(
+            k in body
+            for k in ("show_cli_sessions", "show_cron_sessions")
+        ):
+            try:
+                _clear_session_list_cache()
+            except Exception:
+                pass
+
         auth_enabled_after = is_auth_enabled()
         auth_just_enabled = bool(
             requested_password and auth_enabled_after and not auth_enabled_before

tests/conftest.py

@@ -1073,6 +1073,18 @@ def cleanup_test_sessions():
     was left in an unexpected state by a prior test in the same shard.
     """
     created: list[str] = []
+    # Defense-in-depth: reset the CLI-session visibility setting to its default
+    # BEFORE the test runs too, not only in teardown. Teardown-only reset relies
+    # on every sibling test being wrapped by this fixture AND on its teardown
+    # actually completing; a pre-test reset guarantees each test starts from a
+    # known visibility state regardless of what a prior test left behind, so a
+    # test that reads /api/sessions without first setting the toggle can't
+    # inherit a leaked value. Pairs with the POST /api/settings cache
+    # invalidation that fixes the gateway_sync row-absence flake at the source.
+    try:
+        _post(TEST_BASE, "/api/settings", {"show_cli_sessions": False})
+    except Exception:
+        pass
     yield created
 
     for sid in created:

tests/test_1325_user_fenced_code.py

@@ -46,7 +46,7 @@ def _run_user_render(text_input):
     try:
         result = subprocess.run(
             ['node', tf.name, json.dumps(text_input)],
-            capture_output=True, text=True, timeout=10
+            capture_output=True, text=True, timeout=30
         )
         if result.returncode != 0:
             raise RuntimeError(f"node error: {result.stderr}")

tests/test_1707_workspace_filename_click.py

@@ -278,7 +278,7 @@ setTimeout_(() => {
     )
     r = subprocess.run(
         [NODE, "-e", js],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if r.returncode != 0:
         raise RuntimeError(f"node failed: {r.stderr}")

tests/test_chat_start_provider_fallback.py

@@ -126,7 +126,7 @@ def _run_helper(driver_path, payload):
         [NODE, driver_path, str(UI_JS_PATH), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")
@@ -139,7 +139,7 @@ def _run_model_state_helper(driver_path, payload):
         [NODE, driver_path, str(UI_JS_PATH), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")

tests/test_issue1154_fenced_code_leak.py

@@ -64,7 +64,7 @@ def _render(driver_path, markdown: str) -> str:
     import subprocess
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH)],
-        input=markdown, capture_output=True, text=True, timeout=10,
+        input=markdown, capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_issue1188_fuzzy_match.py

@@ -77,7 +77,7 @@ def _find(driver_path, model_id: str, options: list[str], preferred: str | None
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH),
          json.dumps({"modelId": model_id, "options": options, "preferredProvider": preferred})],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_issue1446_glued_heading_lift.py

@@ -233,7 +233,7 @@ def _render(driver_path: str, markdown: str) -> str:
         input=markdown,
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_issue1618_yaml_json_diff_newline_preserve.py

@@ -169,7 +169,7 @@ def _render(driver_path, markdown: str) -> str:
         input=markdown,
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_issue1771_session_model_switch_sync.py

@@ -164,7 +164,7 @@ def _run_sync(driver_path, *, session_model, initial_value="@expensive:gpt-5.5",
         [NODE, driver_path, str(UI_JS_PATH), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")

tests/test_issue2881_workspace_chat_links.py

@@ -63,7 +63,7 @@ def _render(driver_path: str, markdown: str) -> str:
         input=markdown,
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(result.stderr)

tests/test_issue3360_multi_slash_model_collision.py

@@ -114,7 +114,7 @@ def _find(driver_path, model_id, options, preferred=None):
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH),
          json.dumps({"modelId": model_id, "options": options, "preferredProvider": preferred})],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")
@@ -124,7 +124,7 @@ def _find(driver_path, model_id, options, preferred=None):
 def _norm_keys(driver_path, ids):
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH), json.dumps(ids)],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_issue3368_model_prefix_shadowing.py

@@ -164,7 +164,7 @@ def _resolve(driver, query, groups, sel_options):
     r = subprocess.run(
         [NODE, driver, str(COMMANDS_JS_PATH),
          json.dumps({"query": query, "groups": groups, "selOptions": sel_options})],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if r.returncode != 0:
         raise RuntimeError(f"node driver failed: {r.stderr}")
@@ -175,7 +175,7 @@ def _find(driver, model_id, options, preferred=None):
     r = subprocess.run(
         [NODE, driver, str(UI_JS_PATH),
          json.dumps({"modelId": model_id, "options": options, "preferredProvider": preferred})],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if r.returncode != 0:
         raise RuntimeError(f"node driver failed: {r.stderr}")
@@ -185,7 +185,7 @@ def _find(driver, model_id, options, preferred=None):
 def _best(driver, query, options):
     r = subprocess.run(
         [NODE, driver, str(COMMANDS_JS_PATH), json.dumps({"query": query, "options": options})],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if r.returncode != 0:
         raise RuntimeError(f"node driver failed: {r.stderr}")

tests/test_issue3429_uri_scheme_model_ids.py

@@ -92,7 +92,7 @@ def norm_driver(tmp_path_factory):
 def _labels(driver_path, ids):
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH), json.dumps(ids)],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")
@@ -102,7 +102,7 @@ def _labels(driver_path, ids):
 def _norm_keys(driver_path, ids):
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH), json.dumps(ids)],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_issue3460_cron_session_unread.py

@@ -51,7 +51,7 @@ def _run_node(script: str) -> dict:
         check=True,
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     return json.loads(result.stdout)
 

tests/test_issue3691_model_picker_show_all.py

@@ -1394,7 +1394,7 @@ def _run_dropdown_driver(driver_path: str, payload: dict | None = None) -> dict:
         [NODE, driver_path, str(REPO / "static" / "ui.js"), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")
@@ -1562,7 +1562,7 @@ def test_runtime_inplace_expand_then_search_clear_preserves_expanded_group(_driv
         [NODE, _driver_paths["inplace"], str(REPO / "static" / "ui.js"), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node inplace driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")
@@ -1621,7 +1621,7 @@ def test_runtime_inplace_endpoint_error_group_renders_open(_driver_paths):
         [NODE, _driver_paths["endpoint_error"], str(REPO / "static" / "ui.js"), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node endpoint_error driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")
@@ -1663,7 +1663,7 @@ def test_runtime_inplace_expand_with_preexisting_options_reveals_them(_driver_pa
         [NODE, _driver_paths["preexisting"], str(REPO / "static" / "ui.js"), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node preexisting driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")

tests/test_issue3959_model_suffix_dedup.py

@@ -138,7 +138,7 @@ def test_backend_frontend_parity_complex_aggregator_proxy():
         tmp = f.name
     try:
         result = subprocess.run(
-            ["node", tmp], capture_output=True, text=True, timeout=10
+            ["node", tmp], capture_output=True, text=True, timeout=30
         )
         assert result.returncode == 0, f"JS execution failed: {result.stderr}"
         js_pairs = json.loads(result.stdout.strip())

tests/test_issue500_message_list_virtualization.py

@@ -26,7 +26,7 @@ def _run_node(source: str) -> str:
             cwd=str(REPO_ROOT),
             capture_output=True,
             text=True,
-            timeout=10,
+            timeout=30,
         )
     finally:
         script_path.unlink(missing_ok=True)

tests/test_issue500_session_list_virtualization.py

@@ -26,7 +26,7 @@ def _run_node(source: str) -> str:
             cwd=str(REPO_ROOT),
             capture_output=True,
             text=True,
-            timeout=10,
+            timeout=30,
         )
     finally:
         script_path.unlink(missing_ok=True)

tests/test_model_default_boot_precedence.py

@@ -307,7 +307,7 @@ def _run_populate_driver(
         [NODE, driver_path, str(REPO / "static" / "ui.js"), json.dumps(payload)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed:\nSTDOUT={result.stdout}\nSTDERR={result.stderr}")

tests/test_reasoning_chip_js_behaviour.py

@@ -108,7 +108,7 @@ def _apply(driver_path, value):
     import json as _json
     result = subprocess.run(
         [NODE, driver_path, str(UI_JS_PATH), _json.dumps(value)],
-        capture_output=True, text=True, timeout=10,
+        capture_output=True, text=True, timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_renderer_js_behaviour.py

@@ -80,7 +80,7 @@ def _render(driver_path, markdown: str) -> str:
         input=markdown,
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(f"node driver failed: {result.stderr}")

tests/test_session_copy_link_action.py

@@ -53,7 +53,7 @@ def _run_session_search_helper(cases):
         [NODE, "-e", driver, json.dumps(cases)],
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     assert result.returncode == 0, result.stderr
     return json.loads(result.stdout)

tests/test_session_lineage_collapse.py

@@ -24,7 +24,7 @@ def _run_node(source: str) -> str:
         capture_output=True,
         encoding="utf-8",
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(result.stderr)

tests/test_tool_card_preview_summary.py

@@ -62,7 +62,7 @@ def _preview(driver_path: str, tc: dict, display_snippet: str = "") -> str:
         input=json.dumps({"tc": tc, "displaySnippet": display_snippet}),
         capture_output=True,
         text=True,
-        timeout=10,
+        timeout=30,
     )
     if result.returncode != 0:
         raise RuntimeError(result.stderr)