Skip to content

ci: drop X64 from runs-on of arch-neutral jobs (workspace-test-shard, mg-reuse, x86-main-advisories) - #5871

Closed
noahgift wants to merge 4 commits into
mainfrom
ci/runs-on-drop-x64-neutral
Closed

noahgift wants to merge 4 commits into
mainfrom
ci/runs-on-drop-x64-neutral

Conversation

@noahgift

Copy link
Copy Markdown
Member

Shards first: workspace-test-shard (ci.yml:444)

-    runs-on: ${{ github.event_name == 'merge_group' && fromJSON('["self-hosted","Linux","X64","mq"]') || fromJSON('["self-hosted","Linux","X64","clean-room"]') }}
+    runs-on: ${{ github.event_name == 'merge_group' && fromJSON('["self-hosted","Linux","mq"]') || fromJSON('["self-hosted","Linux","clean-room"]') }}

Audit caveat (handoff/x64-pin-audit-1021.md): last aarch64 measurement is 2026-09-12 (ci/sections.yml:97-103, 59k tests, four aarch64-only reds fixed then). Tests under cfg(target_arch = "x86_64") compile out on aarch64, so an aarch64 shard silently runs fewer tests; the Σ-executed check (ci.yml:509-528) is only self-consistent with one arch per event. x86-main and its determinism[X64] leg keep an X64 home. Revert rule applies.

Also

  • mg-reuse (:103) and x86-main-advisories (:317): [self-hosted, Linux, X64, clean-room] -> [self-hosted, Linux, clean-room], matching mg-reuse-self-test (:143). No amd64/x86_64 strings in their scripts; the arm64 sovereign-ci image ships cargo-deny.
  • x86-main (:160) unchanged: jq-linux-amd64 pin (:183-196) and determinism[X64] (:252).
  • scripts/ci_mg_reuse.sh: pool-row expectations follow the new labels (first quorum round caught the pinned row); --self-test 105 passed, 0 failed.

Operator 14:06Z/14:18Z via aprender-57. One commit. Not armed: needs quorum 3/3 + 2 seats.

🤖 Generated with Claude Code

@noahgift

Copy link
Copy Markdown
Member Author

quorum-review (AD-04): NOT agreed (auto_merge: checked=true was_armed=false disarmed=false)

{
 "ticket": "PMAT-5736",
 "head": "12494f5d71070b6b8dc9018395d7a54ff5fff3b4",
 "width": 3,
 "executor": "agy",
 "agreed": false,
 "auto_merge": {
  "checked": true,
  "was_armed": false,
  "disarmed": false,
  "note": "auto-merge not armed"
 },
 "lanes": [
  {
   "lane": 1,
   "verdict": "FAIL",
   "findings": 5
  },
  {
   "lane": 2,
   "verdict": "FAIL",
   "findings": 4
  },
  {
   "lane": 3,
   "verdict": "FAIL",
   "findings": 5
  }
 ]
}

…use, x86-main-advisories (Refs #5736)

Operator 14:18Z via aprender-57. Audit: handoff/x64-pin-audit-1021.md. x86-main stays X64
(jq-amd64 pin, determinism[X64]). scripts/ci_mg_reuse.sh pool-row self-test updated to match.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@noahgift
noahgift force-pushed the ci/runs-on-drop-x64-neutral branch from 12494f5 to 74f7f06 Compare October 21, 2026 14:27
@noahgift

Copy link
Copy Markdown
Member Author

quorum-review (AD-04): three PASS — agreed (auto_merge: checked=true was_armed=false disarmed=false)

{
 "ticket": "PMAT-5736",
 "head": "74f7f06e6cc5fd2b9a08429df973b278f2469164",
 "width": 3,
 "executor": "agy",
 "agreed": true,
 "auto_merge": {
  "checked": true,
  "was_armed": false,
  "disarmed": false,
  "note": "auto-merge not armed"
 },
 "lanes": [
  {
   "lane": 1,
   "verdict": "PASS",
   "findings": 5
  },
  {
   "lane": 2,
   "verdict": "PASS",
   "findings": 3
  },
  {
   "lane": 3,
   "verdict": "PASS",
   "findings": 0
  }
 ]
}

@noahgift noahgift left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Non-author seat (aprender-triage-worker, Sonnet) @74f7f06e6c: APPROVE. Read the 2-file diff (ci.yml, ci_mg_reuse.sh). Executed on the PR head: ci_mg_reuse --self-test 105/0 (pool-row mutants incl. pr-path-takes-mq and mq-lane-dropped-on-shard still red), check_runner_labels ok, actionlint only the pre-existing SC2086. Drops X64 from mg-reuse, x86-main-advisories, workspace-test-shard; x86-main keeps X64 (audit x64-pin-audit-1021: jq-amd64 pin, determinism[X64]). The pool-row gate still pins each runs-on line exactly. Non-blocking: audit says shard arch-neutrality was measured once (2026-09-12) and advisories only by reading, so watch the first aarch64 landings. Quorum receipt w5018-quorum-5871-74f7f06e6c.json (3/3, agy lanes) cited, not re-run. Seat 1 of 2; a second non-author seat is still needed. Not armed.

noahgift and others added 3 commits October 21, 2026 16:44
…tolerance_rel; q6_k aarch64 measured on gx10 (Refs #5736)

The shard moved to the arch-neutral pool and cpu.matvec.q6_k read 1.65e-6 on aarch64 against the
x86_64 receipt's 4e-7. The test now takes the bound from the receipt's own host_arch, else from
per_arch.<arch>, and otherwise fails listing every receipt that lacks one with the host's measured
error. No skip, no borrowed bound.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…g numbers (Refs #5736)

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…eipts, measured on gx10 CI (Refs #5736)

Bound = tolerance_from(measured) (2x, rounded up one digit), the rule the x86_64 receipts used.

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

Copy link
Copy Markdown

§13.11 rung 1 — quorum shadow verdict

S13-SHADOW pr=5871 head=006b0a4f6a83d6fc06f24861c0891432fc23894e verdict=REFUSE class=Q1 arm_rc=1

Shadow mode: this records a verdict and merges nothing. A refusal
to arm is not a block (§13 adds zero rows to §7) — the pull request is
exactly as green as it was.

@noahgift

Copy link
Copy Markdown
Member Author

EMG-001 sort: returned. ci.yml conflicts with C1; rebuild after #5925 (pace DIRTY-CHECK 12:53Z). Not closed.

@noahgift

Copy link
Copy Markdown
Member Author

superseded by #5886 (the re-cut of #5871/#5882). Branch kept.

@noahgift noahgift closed this Oct 23, 2026
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