Skip to content

test(F117): two fixtures follow the F117 message shape (unblocks fork gate Step 4) - #194

Merged
mindfn merged 1 commit into
develop_basefrom
fix/f117-fixture-message-shape
Sep 26, 2026
Merged

mindfn merged 1 commit into
develop_basefrom
fix/f117-fixture-message-shape

Conversation

@mindfn

@mindfn mindfn commented Sep 26, 2026

Copy link
Copy Markdown
Owner

Why

F117 changed the message shape on develop_base, but two test fixtures still used the old shape. As a result, every fork PR's pnpm gate fails at Step 4/6, even when the PR touches neither file. Found while gating #193 (comment 5843885399).

What

Fixture Old failure Fix
packages/web/src/stores/__tests__/chatStore-delivered-connector-source.test.ts TS2353: 'mentions' does not exist in type … in the full tsc --noEmit (tests included). The store's delivered-message type dropped mentions/userId (6392172, f8071e4), but the inline envelope still carried them. The inline envelope now comes from a helper, the same way the hold-ball notice next to it does. This keeps the real socket envelope, which still sends mentions/userId.
packages/api/test/personal-chrome-assistant-return-poller.test.ts Runtime error append requires one valid MessageFrom sender identity. MessageStore.append now requires from and rejects the catId projection (f426fb7). The file runs only in test:f247-chrome-host-spike, so public CI never exercises it. Both sources are now written as from: { kind: 'agent', catId: 'codex-sol' }.

Test-only change; production code is untouched.

Verification (at develop_base 7d57371 + this commit)

  • pnpm -r exec tsc --noEmit (gate Step 4) across all packages: exit 0. Before the fix, web failed.
  • node --import tsx --test test/personal-chrome-assistant-return-poller.test.ts: 6/6. Before the fix: 5/6.
  • vitest run chatStore-delivered-connector-source.test.ts: 3/3.
  • Rest of the f247 chrome-host line: 175/175 and 11/11. Nothing else in that line depended on the old shape.

[宪宪/claude-opus-5-5🐾]

🤖 Generated with Claude Code

Why: F117 changed the message shape on develop_base but two fixtures
still used the old one, so every fork PR's `pnpm gate` failed at Step 4
(reported from #193, comment 5843885399).

- web chatStore-delivered-connector-source: the inline delivered
  envelope carried `mentions`/`userId`, which the store's delivered-message
  type no longer declares (6392172, f8071e4), so the full tsc (tests
  included) rejected the literal. It now comes from a helper like the
  hold-ball notice beside it, which keeps the real socket envelope.
- api personal-chrome-assistant-return-poller: MessageStore.append now
  requires a MessageFrom sender and rejects the catId projection
  (f426fb7); the two sources are written as agent codex-sol. This file
  only runs in test:f247-chrome-host-spike, outside public CI.

Verified: `tsc --noEmit` across all packages exits 0; the poller file
6/6; the web file 3/3; the rest of the f247 line 175/175 + 11/11.

[宪宪/claude-opus-5-5🐾]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@mindfn
mindfn merged commit 52b80e1 into develop_base Sep 26, 2026
mindfn added a commit that referenced this pull request Sep 26, 2026
Why: record the Phase 3 status per the roadmap rule. The Queue-row owner
of the first design was replaced by the dispatch-state design Fable
ruled on (…000201); #196 implements it with carrier consumption
evidence, a waiting ref, read/unread settlement, the Queue Panel rows
and an honest recall refusal. Build and API full comparison clean; the
whole web suite is still running. Also notes #194/#195, the two F117
test leftovers that blocked every fork gate.

[宪宪/claude-opus-5-5🐾]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mindfn added a commit that referenced this pull request Sep 26, 2026
Why: roadmap rule — record each PR's status and the SHA 3001 runs. kimi
approved #196 at 84a40d1 (…000403) and accepted the rollback
trade-off; the web full-suite comparison is on the PR (5848460811). 3001
still runs 9aa0c01, so the next restart loads 2c, #194, #195 and
Phase 3 together.

[宪宪/claude-opus-5-5🐾]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mindfn added a commit that referenced this pull request Sep 27, 2026
…pter is unavailable (upstream zts212653#1531) (#193)

Cherry-pick of upstream zts212653#1531 (squash 1789afb; approved in review 5324110998 at f641375). It fixes zts212653#1532, the reply poller checking for ChatGPT Pro every second on Hosts where it isn't installed. While the adapter answers HOST_UNAVAILABLE, polling backs off exponentially up to one check per 60 s, timed on a monotonic clock, and only changes of state are logged.

Why now: Landy saw the once-a-second log line on this running instance and asked me to merge this fork PR myself (thread_mu8dg6h7l2x4ohsk …000470).

Evidence (on a tree byte-identical to this PR's head):
- pnpm gate --no-rebase steps 1-4 green, including step 4 (tsc with tests) after #194.
- Step 5 cut off by the 3 h wall limit with 0 failures (the machine keeps suspending).
- Upstream CI ran the full suite green on the identical change.
- Focused lane 21/21.

The fix takes effect after the operator restarts the running instance.

[宪宪/claude-opus-5-5🐾]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mindfn added a commit that referenced this pull request Sep 27, 2026
… 后 gate Step 1–4 通过

Why: zts212653#1531 这条线在 fork 侧收尾:Landy 让我自己合 fork-internal 的 #193(…000470);
记下 gate 证据(Step 5 被机器休眠拖过 3 小时墙钟上限截断,截断前 0 失败)和剩下的一步(operator 重启)。

[宪宪/claude-opus-5-5🐾]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mindfn added a commit that referenced this pull request Sep 27, 2026
Why: roadmap rule — record each PR's status and the SHA 3001 runs. kimi
approved #200 at d70dc35 (…000595) with no P1/P2; his one note (a direct
call without a Queue slot no longer has the 2x cap) only concerns tests
and scripts, since the Queue is production's single dispatch entry. 3001
still runs 9aa0c01, so the next restart loads 2c, #194, #195, Phase 3,
#197, #198 and Phase 4 together.

[宪宪/claude-opus-5-5🐾]

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant