Skip to content

fix(queen): create the Queen schema before migrating; mend red server tests - #500

Merged
gHashTag merged 2 commits into
feat/queen-supervisorfrom
fix/red-server-tests
Sep 21, 2026
Merged

gHashTag merged 2 commits into
feat/queen-supervisorfrom
fix/red-server-tests

Conversation

@gHashTag

Copy link
Copy Markdown
Owner

Real bug

runPgMigrations failed on any fresh database: the pool pins search_path to queenSchema() and nothing created it (no schema has been selected to create in). Now CREATE SCHEMA IF NOT EXISTS first. Reproduced on PostgreSQL 16: old code 2 fail, new 3 pass.

Tests red on every PR (#495-#499)

  • queen-candidates: hard-coded /Users/playra/BrowserOS -> this checkout
  • queen-salvage-guards (2): need queend for a verdict -> it.if(queendPresent), as queen-adversarial-review does
  • queen-salvage unmerged: merge had no identity; runner hostname without a domain -> git refused before the index. Adds identity + asserts UU
  • navigation/windows hidden tests: accept 'Hidden windows are no longer supported' as a platform skip
  • dom XPath: poll up to 5 s instead of racing page load

Local: candidates 6/6, salvage 14/14, salvage-guards 13 pass + 2 skip, pglive 3/3 against PG16; tsc clean. Browser tests need CI.

🤖 Generated with Claude Code

…ests

The migrations failed on every fresh database: the pool pins search_path to
queenSchema() ('trios') and nothing created that schema, so a new
environment, a restore or CI got 'no schema has been selected to create in'
on every statement. runPgMigrations now runs CREATE SCHEMA IF NOT EXISTS
first; reproduced against PostgreSQL 16 (old code 2 fail, new code 3 pass),
and the live gate now inspects that schema instead of 'public'.

Tests that were red on every PR into this branch:
- queen-candidates: read the checkout it runs in, not /Users/playra/BrowserOS
- queen-salvage-guards: two tests need queend for a verdict, gated like
  queen-adversarial-review; the fake gets a measurement budget of 1
- queen-salvage: the merge meant to conflict now has an identity; a CI
  runner hostname with no domain made git refuse it before the index, and
  the test now asserts the conflict (UU) actually happened
- navigation/windows: newer BrowserOS says 'Hidden windows are no longer
  supported'; the skip helper accepts both wordings
- dom search_dom XPath: polls up to 5s instead of racing the page load

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@github-actions github-actions Bot added the fix label Sep 21, 2026
@github-actions

github-actions Bot commented Sep 21, 2026

Copy link
Copy Markdown

✅ Tests passed — 2377/2437

Suite Passed Failed Skipped
agent 87/87 0 0
build 9/9 0 0
cdp-protocol 5/5 0 0
eval 93/93 0 0
server-agent 272/272 0 0
server-api 1257/1316 0 59
server-browser 6/6 0 0
server-integration 10/11 0 1
server-lib 279/279 0 0
server-pglive 3/3 0 0
server-root 68/68 0 0
server-skills 31/31 0 0
server-tools 243/243 0 0
shared 14/14 0 0

View workflow run

A second search_dom test (nodeId) raced the page load in CI the run after
the XPath one was fixed. Both now go through searchUntil, which asks for up
to five seconds until the answer carries what the test looks for.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@gHashTag
gHashTag merged commit 669dd96 into feat/queen-supervisor Sep 21, 2026
15 of 16 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant