fix(chat): keep in-flight answers alive across in-app view switches

Root cause (owner repro, verified in a real browser 2026-09-06): the
five navbar views (Chat, RAG, Sources, Tuning, History) were separate
HTML documents, so a navbar click was a REAL cross-document navigation
— the chat page unloaded, the in-flight SSE fetch was aborted, and the
phase-48 teardown (app/api/chat.py `finally`, "chat: turn cancelled")
stopped the model. Observed: send question -> click RAG mid-stream ->
click Chat -> the answer never finished: no `query_log` row, and on
return a dangling question with no brain record (the pre-token pagehide
partial persist skips because `acc` is empty).

Phase-48 LOCKED-DECISION REFINEMENT (owner-confirmed 2026-09-06,
flagged per AGENTS.md rule 3, not silently deviated): "real navigation
cancels the fetch" now means LEAVING THE APP — tab close,
external/other-document navigation, the Stop button. In-app navbar
switches are client-side view switches and no longer cancel.

Fix — Option A (SPA shell), chosen over B (Service Worker owns the
stream) and C (server-side turn registry + resume):
- frontend/index.html is the shell: ONE `<main id="main">` holds the
  five `<section class="view">` blocks; hidden views carry BOTH
  `hidden` and `inert` (WCAG — no focus/keyboard traversal). The
  shared header, the single `doc-modal-*` skeleton, and the
  `#app-version` footer each exist exactly once; the per-view copies
  from the four folded pages are dropped.
- New frontend/assets/router.js (vanilla module — no framework, no
  bundler, No-CDN rule intact): lazy-imports a view module on FIRST
  show only (mount-once, hide-forever — the chat view's in-flight SSE
  reader persists across switches; that persistence IS the fix);
  intercepts same-shell navbar links with preventDefault +
  history.pushState (never a document load); handles popstate; single
  writer of `.nav-link` active state (is-active + aria-current),
  document.title, and the per-view meta description (values carried
  over from the old pages' heads, brand-resolved at write time).
- Each folded page's JS becomes `export async function mount(root)` —
  root-scoped queries; `initSharedHeader()` dropped (the header boots
  once in the shell via the chat module; the admin flag comes from the
  same cached `fetchIsAdmin()` promise — zero extra requests).
- app/main.py: a small list-driven route factory serves the shell for
  /tuning.html, /sources.html, /git-sources.html, /history.html —
  registered AFTER the API routers and BEFORE the static catch-all
  (routes-first). The phase-33 caching middleware applies no-cache +
  `?v=` rewriting unchanged; app/core/caching.py needed NO change
  (the view paths did not change — pinned by the integration tests).
- The four old view .html files are DELETED (one source of truth);
  deep links to the old URLs keep working (the router picks the view
  from the pathname); `/?chat=<id>` is unaffected; the Containerfile
  bundles router.js (inlining the lazy view modules) and drops the
  folded page files.
- app/schemas.py: HistoryTurn.text cap 4000 -> 32000 — the shell
  keeps long saved answers in the chat, and the old cap (stricter than
  the 24_000-char total history budget) 422-rejected any second turn
  in such a chat (found by the phase-42 E2E suite on the shell).

Boundaries: login.html, shared.html, doc-edit.html, document.html
REMAIN separate documents (flow pages, not navbar tabs); a mid-stream
navigation to doc-edit/document.html still cancels per phase 48
(follow-up candidate, out of scope). The SSE API is unchanged. Real
departures still cancel the turn — phase 48 intact (pinned by
tests/e2e/test_stop_generation.py, unchanged, and by the new suite's
real-departure control).

Tests:
- Phase-20 suite REWRITTEN to the new semantics
  (tests/e2e/test_sources_midstream_bug.py): a navbar switch no longer
  cancels — the stream survives the switch and the FULL answer
  settles; the pagehide partial persist REMAINS for real departures
  (the partial's exact shape — first streamed chunk prefix, no done
  metadata — is still pinned there).
- NEW story suite tests/e2e/test_nav_switch_keeps_stream.py (mock
  LLM): the owner repro (send -> RAG mid-stream -> Chat: window
  sentinel survives = same document, FULL answer, exactly one brain
  turn in bor.chat.v1, exactly one settled query_log row, auto-saved
  row matches) + the same mid-stream switch against the other three
  views + the real-departure-still-cancels control + the no-switch
  baseline.
- tests/unit/test_frontend_router.py: source-level pins of the router
  invariants (click interceptor targets ONLY same-shell view paths,
  pushState-only switches, mount-once guard, hidden+inert pair,
  single-writer active state/title); shell-route integration tests
  (each folded path serves the shell with no-cache + `?v=` body; a
  non-view path still 404s); the file-reading unit pins re-pointed at
  the shell (the four view files are gone — the shell is the source
  of truth).

Verification (this commit): full suite green — 1565 unit+integration
tests, app/ coverage 99% (>90% floor); ruff + pyright clean; the
phase's E2E suites green in isolation (house protocol, AGENTS.md rule
9). Owner repro verified in a real browser against the real LLM
(dev server :8010, headful Chromium): "tell me about everquest" ->
RAG mid-stream -> Chat — the answer completed with one brain bubble
and no error banner, `query_log` gained exactly one settled row
(deflected=True: the dev KB holds no EverQuest docs — the settle, not
the topic, is the proof), zero "chat: turn cancelled" lines for that
turn; the control (real navigation to /shared.html mid-stream) still
cancelled (no settled row, the cancel line logged, the partial
persisted on return). Screenshots: .agents/screenshots/76_manual_*.

Phase 76 (76_spa_nav_shell) complete — moved to
.agents/phases/complete/.
This commit is contained in:
2026-09-06 06:31:31 -04:00
parent 7e567bddf3
commit ffa919b8bf
78 changed files with 5548 additions and 3244 deletions
+64 -32
View File
@@ -2,7 +2,9 @@
The browser behavior is E2E-covered (tests/e2e/test_sync_upload_progress.py,
task 06); here we pin the source-level wiring in sources.js, styles.css,
and sources.html — the fmtSyncLabel contract (both kinds, file
and the shell's RAG view markup (index.html — sources.html / git-
sources.html folded in, phase 76 task 02) — the fmtSyncLabel contract
(both kinds, file
present/absent, counts only when total > 0), enterSyncRunningState
writing the full untruncated path to the button title + #sync-result,
the two-job tick decision tree (sync running > upload running > sync
@@ -29,9 +31,11 @@ from pathlib import Path
FRONTEND = Path(__file__).resolve().parents[2] / "frontend"
SOURCES_JS = FRONTEND / "assets" / "sources.js"
STYLES_CSS = FRONTEND / "assets" / "styles.css"
SOURCES_HTML = FRONTEND / "sources.html"
# Phase 76 (task 02): both HTML pages are folded into the ONE-document
# shell — the pinned comments now live in the RAG / Sources view
# sections of index.html.
SHELL_HTML = FRONTEND / "index.html"
GIT_SOURCES_JS = FRONTEND / "assets" / "git-sources.js"
GIT_SOURCES_HTML = FRONTEND / "git-sources.html"
def _js() -> str:
@@ -43,7 +47,7 @@ def _css() -> str:
def _html() -> str:
return SOURCES_HTML.read_text(encoding="utf-8")
return SHELL_HTML.read_text(encoding="utf-8")
def _gjs() -> str:
@@ -51,27 +55,38 @@ def _gjs() -> str:
def _ghtml() -> str:
return GIT_SOURCES_HTML.read_text(encoding="utf-8")
return SHELL_HTML.read_text(encoding="utf-8")
def _gfn(js: str, name: str) -> str:
"""The source of the first `function <name>` in git-sources.js
(up to the first line-leading closing brace — the house pin
pattern)."""
(brace balanced — since phase 76 task 02 the functions live inside
mount(root), so the closing brace is indented, not line-leading).
"""
fn = js.find(f"function {name}")
assert fn != -1, f"{name} must be defined in git-sources.js"
return js[fn : js.find("\n}", fn)]
open_idx = js.find("{", fn)
depth = 0
for i in range(open_idx, len(js)):
c = js[i]
if c == "{":
depth += 1
elif c == "}":
depth -= 1
if depth == 0:
return js[fn : i + 1]
raise AssertionError(f"unbalanced braces in {name}")
def _utick(js: str) -> str:
"""The upload poll tick inside startUploadPolling — from `const
tick = async () => {` to the next top-level function
(initUploadStatus), so the whole decision tree is in the slice."""
tick = async () => {` to the next function (initUploadStatus), so
the whole decision tree is in the slice."""
fn = js.find("function startUploadPolling")
assert fn != -1, "startUploadPolling must be defined in git-sources.js"
tick = js.find("const tick = async () => {", fn)
assert tick != -1, "the tick must live inside startUploadPolling"
end = js.find("\nasync function initUploadStatus", tick)
end = js.find("async function initUploadStatus", tick)
assert end != -1, "initUploadStatus must follow startUploadPolling"
return js[tick:end]
@@ -88,23 +103,34 @@ def _usubmit(js: str) -> str:
def _fn(js: str, name: str) -> str:
"""The source of the first `function <name>` in js (up to the first
line-leading closing brace — the house pin pattern from
tests/unit/test_sync_button.py)."""
"""The source of the first `function <name>` in js (brace balanced —
since phase 76 task 02 the functions live inside mount(root), so
the closing brace is indented, not line-leading; the house pin
pattern from tests/unit/test_sync_button.py)."""
fn = js.find(f"function {name}")
assert fn != -1, f"{name} must be defined in sources.js"
return js[fn : js.find("\n}", fn)]
open_idx = js.find("{", fn)
depth = 0
for i in range(open_idx, len(js)):
c = js[i]
if c == "{":
depth += 1
elif c == "}":
depth -= 1
if depth == 0:
return js[fn : i + 1]
raise AssertionError(f"unbalanced braces in {name}")
def _tick(js: str) -> str:
"""The poll tick inside startSyncPolling — from `const tick = async
() => {` to the next top-level function (startSync), so the whole
two-job decision tree is in the slice."""
() => {` to the next function (startSync), so the whole two-job
decision tree is in the slice."""
fn = js.find("function startSyncPolling")
assert fn != -1, "startSyncPolling must be defined in sources.js"
tick = js.find("const tick = async () => {", fn)
assert tick != -1, "the tick must live inside startSyncPolling"
end = js.find("\nasync function startSync", tick)
end = js.find("async function startSync", tick)
assert end != -1, "startSync must follow startSyncPolling"
return js[tick:end]
@@ -309,8 +335,9 @@ def test_reattach_adopts_a_running_upload_only() -> None:
assert '"upload", upload.current_file, upload.files_done, upload.files_total' in branch
assert 'emitSyncStatus({ state: "running" })' in branch
assert "startSyncPolling()" in branch
# the idle settle is the fall-through (the last statement)
assert body.rstrip().endswith("applySyncIdle(status);")
# the idle settle is the fall-through (the last statement — the
# brace-balanced body ends with the closing brace)
assert body.rstrip().removesuffix("}").rstrip().endswith("applySyncIdle(status);")
# ---------- the section header + the page comment ----------
@@ -332,10 +359,11 @@ def test_section_header_documents_the_two_job_contract() -> None:
def test_sources_html_comment_documents_the_live_announcer() -> None:
"""The #sync-result comment in sources.html documents the phase-64
dual role: the live file label (both kinds) while either job runs,
untruncated for the aria-live announcer, and empty after an upload
settles (A3 — the upload's counts live on the Sources page)."""
"""The #sync-result comment in the shell's RAG view (formerly
sources.html) documents the phase-64 dual role: the live file
label (both kinds) while either job runs, untruncated for the
aria-live announcer, and empty after an upload settles (A3 — the
upload's counts live on the Sources page)."""
html = _html()
idx = html.find('id="sync-result"')
assert idx != -1
@@ -634,23 +662,27 @@ def test_boot_reattach_branches() -> None:
assert "status.error" in fail
assert "uploadError.hidden = false" in fail
assert 'status.state === "idle"' not in body, "idle does nothing — no branch"
# The boot IIFE: after the list loads, the re-attach runs (admin
# branch only — the anonymous path returns before it).
# The mount's tail (phase 76 task 02 — the boot IIFE is gone):
# after the list loads, the re-attach runs (admin branch only — the
# anonymous path returns before it), and it is the LAST statement
# of mount(root).
i_boot = js.rfind("await loadSources();")
tail = js[i_boot:i_boot + 400]
assert "await initUploadStatus();" in tail
assert "})();" in tail
assert "})();" not in js, "the top-level boot IIFE is gone (mount owns boot)"
assert tail.rstrip().removesuffix("}").rstrip().endswith("await initUploadStatus();")
# ---------- the page comment ----------
def test_git_sources_html_comment_documents_the_202_contract() -> None:
"""The #archive-upload-form comment in git-sources.html documents
the phase-64 202 contract (the phase-49 synchronous paragraph
marked superseded): the 202 = "safely on disk" + the JS-created
toast (no markup), the live "Processing…" label via the status
poll, and the 409 re-attach without an error banner."""
"""The #archive-upload-form comment in the shell's Sources view
(formerly git-sources.html) documents the phase-64 202 contract
(the phase-49 synchronous paragraph marked superseded): the 202 =
"safely on disk" + the JS-created toast (no markup), the live
"Processing…" label via the status poll, and the 409 re-attach
without an error banner."""
html = _ghtml()
idx = html.find('id="archive-upload-form"')
assert idx != -1