The crew reads two words written in one millisecond in the order they were written #65

Merged

grokbot merged 1 commit from grokbot/forge:fix/crew-parks-tie into main

2026-10-18 20:44:40 +00:00

grokbot commented

2026-10-18 20:20:07 +00:00

Owner

Which word on an issue is newest decided a park's strike, the planner's
list, an already-done note and the planner's split, and each compared
created_at alone. That is milliseconds: a park and a done note, a
person's word and a park, or an issue and the one split from it,
written in the same millisecond, tied, and the newer one read as not
newer. So a done issue stayed on the planner's list, a park after a
done note did not undo it, a person's word hid the parks after it, and
the planner was refused a split into an issue it had just opened. On a
fast runner that is chance: crew-parks failed this way on main
(the planner re-planning…, parked twice and then found done).

Each comparison now orders by (created_at, id). id() is monotonic, so
within a millisecond the id says which came first, the same order every
listing here already sorts by. handedOverAt becomes handedOver, the
newest park's time and id, so a split is checked against both. An issue
opened before, in the same millisecond, still cannot stand in.

Validation

pipeline/validate on 6c30ec09: baseline in 672s — bun 1.3.14, CI-shaped env (scrubbed vars, runner PATH), exact check.yml shard + forge.json commands.

step	result	time
`bun install --frozen-lockfile`	ok (exit 0)	0s
`forge.json check.fast`	BASELINE (exit 1)	190s
`check.yml gate (1) pull_request`	BASELINE (exit 1)	15s
`check.yml gate (2) pull_request`	ok (exit 0)	272s
`check.yml gate (3) pull_request`	ok (exit 0)	195s

⚠️ node v20.19.2 here, v22 on CI (tsc runs under node)
⚠️ check.yml gate (1) pull_request stopped early at a BASELINE failure (fail-fast): its later files were not run in this step

test output (failures, totals)

▶ bun install --frozen-lockfile
  ok  exit 0  0s
▶ forge.json check.fast
(fail) egress, reported: an undeclared public host goes through and is counted; a private one never does > an undeclared public host is let through and counted [1.16ms] [shard 10]
Ran 3755 tests across 371 files in 150.9s; 1 failed.
classify: reran test/contain.test.ts test/crew-early.test.ts test/usage-report.test.ts  alone on origin/main dea83e0f and on this branch
  BASELINE  exit 1  190s
▶ check.yml gate (1) pull_request
(fail) egress, reported: an undeclared public host goes through and is counted; a private one never does > an undeclared public host is let through and counted [2.63ms] [shard 3]
Ran 113 tests across 3 files in 14.6s; 1 failed.
classify: reran test/contain.test.ts  alone on origin/main dea83e0f and on this branch
  BASELINE  exit 1  15s
▶ check.yml gate (2) pull_request
Ran 1406 tests across 136 files in 271.6s (136 shards).
  ok  exit 0  272s
▶ check.yml gate (3) pull_request
Ran 1441 tests across 136 files in 194.5s (136 shards).
  ok  exit 0  195s

Which word on an issue is newest decided a park's strike, the planner's list, an already-done note and the planner's split, and each compared `created_at` alone. That is milliseconds: a park and a done note, a person's word and a park, or an issue and the one split from it, written in the same millisecond, tied, and the newer one read as not newer. So a done issue stayed on the planner's list, a park after a done note did not undo it, a person's word hid the parks after it, and the planner was refused a split into an issue it had just opened. On a fast runner that is chance: crew-parks failed this way on main (`the planner re-planning…`, `parked twice and then found done`). Each comparison now orders by (created_at, id). `id()` is monotonic, so within a millisecond the id says which came first, the same order every listing here already sorts by. `handedOverAt` becomes `handedOver`, the newest park's time and id, so a split is checked against both. An issue opened before, in the same millisecond, still cannot stand in.  ### Validation `pipeline/validate` on `6c30ec09`: **baseline** in 672s — bun 1.3.14, CI-shaped env (scrubbed vars, runner PATH), exact check.yml shard + forge.json commands. | step | result | time | |---|---|---| | `bun install --frozen-lockfile` | ok (exit 0) | 0s | | `forge.json check.fast` | BASELINE (exit 1) | 190s | | `check.yml gate (1) pull_request` | BASELINE (exit 1) | 15s | | `check.yml gate (2) pull_request` | ok (exit 0) | 272s | | `check.yml gate (3) pull_request` | ok (exit 0) | 195s | - ⚠️ node v20.19.2 here, v22 on CI (tsc runs under node) - ⚠️ check.yml gate (1) pull_request stopped early at a BASELINE failure (fail-fast): its later files were not run in this step <details><summary>test output (failures, totals)</summary> ``` ▶ bun install --frozen-lockfile ok exit 0 0s ▶ forge.json check.fast (fail) egress, reported: an undeclared public host goes through and is counted; a private one never does > an undeclared public host is let through and counted [1.16ms] [shard 10] Ran 3755 tests across 371 files in 150.9s; 1 failed. classify: reran test/contain.test.ts test/crew-early.test.ts test/usage-report.test.ts alone on origin/main dea83e0f and on this branch BASELINE exit 1 190s ▶ check.yml gate (1) pull_request (fail) egress, reported: an undeclared public host goes through and is counted; a private one never does > an undeclared public host is let through and counted [2.63ms] [shard 3] Ran 113 tests across 3 files in 14.6s; 1 failed. classify: reran test/contain.test.ts alone on origin/main dea83e0f and on this branch BASELINE exit 1 15s ▶ check.yml gate (2) pull_request Ran 1406 tests across 136 files in 271.6s (136 shards). ok exit 0 272s ▶ check.yml gate (3) pull_request Ran 1441 tests across 136 files in 194.5s (136 shards). ok exit 0 195s ``` </details>

grokbot added 1 commit

2026-10-18 20:20:07 +00:00

The crew reads two words written in one millisecond in the order they were written

check / gate (3) (pull_request) Successful in 4m3s

Details

check / gate (1) (pull_request) Successful in 11m9s

Details

check / gate (2) (pull_request) Successful in 18m9s

Details

check / check (pull_request) Successful in 1s

Details

preview / preview (pull_request) Has been skipped

Details

6c30ec0918

Which word on an issue is newest decided a park's strike, the planner's
list, an already-done note and the planner's split, and each compared
`created_at` alone. That is milliseconds: a park and a done note, a
person's word and a park, or an issue and the one split from it,
written in the same millisecond, tied, and the newer one read as not
newer. So a done issue stayed on the planner's list, a park after a
done note did not undo it, a person's word hid the parks after it, and
the planner was refused a split into an issue it had just opened. On a
fast runner that is chance: crew-parks failed this way on main
(`the planner re-planning…`, `parked twice and then found done`).

Each comparison now orders by (created_at, id). `id()` is monotonic, so
within a millisecond the id says which came first, the same order every
listing here already sorts by. `handedOverAt` becomes `handedOver`, the
newest park's time and id, so a split is checked against both. An issue
opened before, in the same millisecond, still cannot stand in.

grokbot merged commit 5881c081a7 into main

2026-10-18 20:44:40 +00:00

grokbot referenced this pull request from a commit

2026-10-18 20:44:42 +00:00

Merge pull request 'The crew reads two words written in one millisecond in the order they were written' (#65) from grokbot/forge:fix/crew-parks-tie into main