Skip to content

spec(crons): MERGE-STEP lands the night loop's own CLEAN night/ PRs, one per tick (Closes #12111) - #12112

Merged
github-actions[bot] merged 2 commits into
masterfrom
night/129-merge-step-own
Oct 24, 2026
Merged

github-actions[bot] merged 2 commits into
masterfrom
night/129-merge-step-own

Conversation

@gHashTag

Copy link
Copy Markdown
Collaborator

Closes #12111

The tick-128 retrospective: MERGE-STEP read only armed PRs, so #12106 (a night/ PR, lab-green on both legs) sat about 25 minutes CLEAN and unarmed, and tick 127 found nothing armed and stopped.

merge_step_verdict in specs/crons/t27os-night-loop.t27:

Lab: run 40134593892 on 86928cb, challenge sha256 d781018521ff..., receipt runs/86928cb1cc0d84dc1f30f272eed96983665e026e.d7810185.json in Trinity-S-AI-DNA/t27os:

  • t27c tests=42|pass=42|FAIL=0|vacuous=0|blocked=0
  • t27b tests=42|pass=42|FAIL=0|vacuous=0|asserts=109
  • seal=true: night-loop resealed (second commit, seal file only).

see Trinity-S-AI-DNA/t27os#3

🤖 Generated with Claude Code

gHashTag and others added 2 commits October 25, 2026 04:30
…ith a lab receipt, one per tick (Closes #12111)

The tick-128 retrospective: #12106 sat ~25 min CLEAN and unarmed because
MERGE-STEP read only armed PRs. merge_step_verdict adds the loop's own
non-draft night/ PR with a lab receipt for its head; another actor's
unarmed lane and every draft stay NOT_OURS (negative control).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…12111)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-24 21:36:51 UTC

Summary

Status Count
Total Open PRs 47
PRs with Failing Checks 40
PRs with All Checks Green 7
READY 4
FAILING 40
PENDING 0
NO CHECKS YET 0

These columns do not partition: 4 + 40 + 0 + 0 = 44, and there are 47 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=55c0f1b1b32d != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@tri-claw

tri-claw Bot commented Oct 24, 2026

Copy link
Copy Markdown

Summary by t27-bees

Checked by the t27 compiler

t27c parse, parse-complete and typecheck on each changed .t27 file at f8ad25c. Nothing was generated, built or run. A pass means the file is well formed, not that it is right.

  • specs/crons/t27os-night-loop.t27: parses and typechecks.

Read by the model

The change adds a merge_step_verdict function to the night-loop cron spec that decides which PRs to merge each tick. Previously only armed PRs were considered; now the loop's own non-draft night/ PRs that are CLEAN and have a lab receipt for their head sha are also eligible—one per tick, lowest number first. This fixes #12111 where a green night PR sat unmerged for 25 minutes because it wasn't armed.

Worth a look

  • specs/crons/t27os-night-loop.t27:263 — merge_step_verdict returns MS_WAIT for a non-armed night PR without a receipt, but also returns MS_WAIT for a non-armed night PR that is clean == false (e.g., UNSTABLE) even if receipt == true (line 272: if (clean == false) { if (armed) return MS_LEAVE_GITHUB; return MS_WAIT; }). The test at line 297 expects MS_WAIT for this case, but the receipt is effectively ignored when the PR isn't clean—this may be intentional (a receipt on an unstable head isn't sufficient), but the logic could be read as "receipt overrides clean" which it doesn't.
  • specs/crons/t27os-night-loop.t27:260 — The function takes both clean and dirty as separate booleans. If both are false (e.g., mergeStateStatus is UNSTABLE or BLOCKED), the non-armed path returns MS_WAIT (line 272–275) while the armed path returns MS_LEAVE_GITHUB. This asymmetry means an armed UNSTABLE PR is left to GitHub but an unarmed night UNSTABLE PR with a receipt waits indefinitely—could cause a night PR to stall if its status flaps between CLEAN and UNSTABLE.
  • specs/crons/t27os-night-loop.t27:268 — lower_ready and merged gate the final MS_MERGE. If lower_ready is true but the lower-numbered PR ultimately gets MS_SKIP_DIRTY or MS_LEAVE_GITHUB (not merged), this tick wastes its merge slot because merged doesn't increment and the higher PR sees lower_ready == true next tick. The spec assumes the lower PR will merge, but the verdicts don't guarantee it.
  • specs/crons/t27os-night-loop.t27:283 — Test the_12106_case_is_merged_by_the_tick passes clean: true, dirty: false, receipt: true. There's no test for a night PR that has a receipt but clean: false (the UNSTABLE case at line 297 expects MS_WAIT), so the behavior for "receipt on unstable head" is only exercised by one assertion and not cross-checked against the armed path.
  • specs/crons/t27os-night-loop.t27:291 — Test negative_control_another_actors_clean_unarmed_lane_is_not_ours uses receipt: true for the lane PR. The function returns MS_NOT_OURS before checking receipt (line 262: if (armed == false and night_branch == false) return MS_NOT_OURS), so the receipt is irrelevant there—but the test includes it, which could mislead a reader into thinking receipts matter for non-night branches.

Read: 2 files, +79 -10, the whole diff. Head f8ad25c. Model: openai-compatible/nvidia/nemotron-3-ultra-550b-a55b. Commands: @t27-bees review | summary | help | pause | resume. Free for every repository.

@github-actions
github-actions Bot merged commit 41c2814 into master Oct 24, 2026
37 checks passed
@gHashTag
gHashTag deleted the night/129-merge-step-own branch October 26, 2026 03:43
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.

spec(crons): MERGE-STEP also lands the night loop's own CLEAN night/ PRs

1 participant