phase: 120_failed_turn_retry
Build and Push Containers / build-and-push-app (push) Successful in 3m5s
Build and Push Containers / build-and-push-db (push) Failing after 2m13s

All verification complete. Final report:

**Phase 120 — Failed-turn retry: verification pass (all 3 tasks were done; final verification + 1 regression fix)**

**Verified:** `ChatMessage.failed`/`error` (≤500, `extra="forbid"` intact); `finalizeFailedTurn` funnel on the 3 failure paths (catch-else, stream-drop guard, zero-frame fallback) with `failed: true` + capped detail + `markLastRetryable`; `appendFailedNote` restore branch (Save-as-doc/Tune excluded); `showErrorBanner`/`retryLastTurn` byte-pinned untouched; only the three paths persist `failed: true` (grep + unit pin); no test asserts the old broken behavior.

**Defect found & fixed (rule 7):** a real navigate-away mid-turn let the browser's teardown fetch rejection (TypeError, not AbortError) leak into the failed funnel, persisting a phantom failed brain record — `test_sources_midstream_bug.py::test_no_orphan_brain_message_when_navigated_before_first_token` failed (2 `.msg` after reload) and violated the phase-20 navigate-away convention. Fixed: turn-scoped `leftThePage` flag (set unconditionally on `pagehide`, reset in `runTurn`) skips the funnel in the catch-else branch; pinned by new unit test `test_navigate_away_is_not_a_failed_turn`. No phase-overview/PLAN/todo/complete files touched; no commits made.

**Gates (exact):**
- `uv run pytest` → 2577 passed
- `uv run pytest --cov=app --cov-report=term-missing` → TOTAL 4271 stmts, 99% (>90%)
- `uv run pytest tests/e2e/test_failed_turn_retry.py -v --no-cov` → 4 passed (isolated)
- `uv run ruff check . && uv run pyright` → clean (0 errors)
- Regression E2E, isolated: `test_sources_midstream_bug.py` 6/6 (was 5/6); `test_llm_retry`/`test_tool_scaffolding_guardrails`/`test_stop_generation`/`test_navbar_refresh` 17/17

**Completion criteria:** (1) network error → banner + in-bubble Retry, re-ask without re-typing ✅ (E2E A); (2) refresh restores failed bubble + working Retry, no "new chat" ✅ (E2E C); (3) stopped/successful turns byte-identical ✅ (negative E2E, stop suite, byte-identity units); (4) pytest/coverage/lint/types ✅; (5) commit + phase move — left to the harness per pass rules.

**Notable:** deviation = the regression fix above (a navigation is not a failed turn; phase-20 partial-persist convention restored). Next pending phase: `121_git_source_tokens`.
This commit is contained in:
2026-09-24 18:50:36 -04:00
parent 0ff1f8c4d6
commit 3a0fc3db05
29 changed files with 2298 additions and 28 deletions
+34 -3
View File
@@ -60,6 +60,11 @@ Test → source mapping:
appears (role=alert, the "dropped the connection" copy), the last
retrying status is the highest attempt — "(4 of 4)…" — and the send
button re-enables (the banner path settles the state machine).
Phase 120 (locked A1): the zero-frame turn ALSO renders the
retryable failed bubble (the fixed honest line + the in-bubble
error note with the detail + the in-bubble Retry) and reveals the
banner Retry (its ``lastBrainWrap`` precondition finally holds on
the error path) — a network error is retryable.
"""
from __future__ import annotations
@@ -115,6 +120,21 @@ MOCK_ANSWER_MARKER = "Deterministic mock answer for E2E"
DEFLECT_PHRASE = r"haven't done anything like that"
ERROR_COPY = "The chat model dropped the connection — try again?"
def _js_const(name: str) -> str:
"""A frontend string constant, read from app.js (no JS/Python
drift — the E2E asserts against the SAME text the page renders)."""
js = (REPO / "frontend" / "assets" / "app.js").read_text(encoding="utf-8")
m = re.search(rf'const {name}\s*=\s*\n?\s*"([^"]*)"', js)
assert m, f"const {name} not found in app.js"
return m.group(1)
#: Phase 120: the fixed bubble text of a ZERO-frame failed turn (the
#: exhaustion case below) — read from app.js so this suite's assertion
#: tracks the page's constant.
FAILED_TURN_TEXT = _js_const("FAILED_TURN_TEXT")
# --------------------------------------------------------------------------
# DB seeding (TRUNCATE-then-seed, cf. test_agent_document_tools.py)
# --------------------------------------------------------------------------
@@ -501,7 +521,18 @@ def test_exhaustion_lands_on_the_error_banner(
assert not [f for f in frames if f.get("type") == "done"]
assert not [f for f in frames if f.get("type") == "delta"]
# No answer bubble was ever rendered (no frame ever streamed a
# token) — the user bubble is the only message in the DOM.
expect(page.locator("#messages > .msg.brain")).to_have_count(0)
# Phase 120 (locked A1): the zero-frame turn is no longer a bare
# question — it persists as a FAILED brain record (the phase-48
# ``stopped`` precedent) and renders the retryable error bubble:
# the fixed honest line (no frame ever streamed a token, so the
# bubble is NOT the empty-answer fallback), the in-bubble error
# note carrying the detail, the in-bubble Retry, and the banner
# Retry revealed too (its ``lastBrainWrap`` precondition finally
# holds on the error path — a network error is retryable).
brain = page.locator("#messages > .msg.brain")
expect(brain).to_have_count(1)
expect(brain.locator(".bubble")).to_have_text(FAILED_TURN_TEXT)
expect(brain.locator(".failed-note")).to_contain_text(ERROR_COPY)
expect(brain.locator(".retry-btn")).to_have_count(1)
expect(page.locator("#banner-retry")).to_be_visible()
expect(page.locator(".msg.user .bubble")).to_have_count(1)