Repository navigation
Conversation
…LAB nightly schedules (PMAT-5847) Refs #5847 PMAT-Ticket: PMAT-5847 PMAT-Batch: BATCH-aprender-build-speed Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…a so a REBASE-replayed PR is judged whole Refs #5859 PMAT-Ticket: PMAT-5859 PMAT-Batch: BATCH-aprender-build-speed Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…IGKILLed Refs #5868 PMAT-Ticket: PMAT-5868 PMAT-Batch: BATCH-aprender-build-speed Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…+3 ticks Refs #5798 PMAT-Ticket: PMAT-5798 PMAT-Batch: BATCH-aprender-build-speed Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…40->120 / 180->120; shard permissions block added; LAB audits created x86-main and workspace-test-shard each gain a failure() && merge_group step that runs `gh run cancel`. merge_group timeouts drop to 120 min (x86-main was 240, the shard 180). Both jobs also set FAT_TIMEOUT_SCALE=1 under merge_group: fat_driver doubles every section timeout by default, which made the 110-min nextest step (ci/sections.yml:520) effectively 220 min, above the 120-min job wall; at scale 1 the step budget (110) fires before the wall (120). Evidence, 14 completed merge_group runs of ci.yml read 2026-10-22: workspace-test-shard n=27 jobs p50 17 / p90 30 / max 43 min; x86-main n=14 p50 29 / p90 38 / max 42 min, so 120 is ~2.8x the worst observed. pull_request/push keep 240/180. workspace-test-shard gets a NEW permissions block (contents: read, checks: read, actions: write) so the cancel is authorised and the check-suite reader keeps working. Adds ci/step-budgets.tsv and the LAB tools scripts/lab/ci_failfast_audit.sh and scripts/ci/step_budget_check.sh, which no PR or release job runs (C324 r5). The cancel is a request: whether it beats the always() gate is unmeasured (see check_ci_unwedge.sh:6-24, #3229) and pending quorum. Refs #5394 PMAT-Ticket: PMAT-5394 PMAT-Batch: BATCH-aprender-build-speed Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…eck that does not exist scripts/affected_set.sh cited scripts/check_affected_readers.sh, which is not in the tree. The comment now says what is true (a crate missing from READERS is skipped silently). A 12-row --self-test on a synthetic graph covers the empty diff, closure + readers, a leaf, a CORE crate, non-.rs under crates/, build.rs, an unplaced path, docs, ALL winning over a placed path and both --filter forms. Dropping the build.rs case turns its row RED. Refs #5974 PMAT-Ticket: PMAT-5974 PMAT-Batch: BATCH-aprender-silicon-coverage Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
noahgift
commented
Oct 22, 2026
noahgift
left a comment
noahgift
left a comment
Member
Author
There was a problem hiding this comment.
Seat 1 (aprender-w5404, non-author): APPROVE @aae18cfeeca04bb8cd7ff25d2a9132e6ae90b5bd. EMG-7 GREEN (7 commits, 0 merges, base ba95f48). Seat-2 blocker fixed: FAT_TIMEOUT_SCALE=1 on both merge_group jobs (fat_driver.py:1305 reads it; <1 refused) makes the 110-min nextest step 110 < the 120-min wall; PR runs keep default scale and 240/180. affected_set self-test is a nightly LAB step, not a gate. Fail-fast cancel is merge_group only. Cited p90 30/38 min (n=27/14) is w5032's figure; I did not re-measure it. Timeout change is a two-way door (revert rule).
This was referenced Oct 22, 2026
noahgift
added
the
parked
Member
Author
|
Parked by R4 triage (conflicting). Ticket: -. The ticket stays open and links this PR; reopen or re-cut when rebased. Head aae18cf. |
Member
Author
|
Corrected ticket line. Ticket: #6367. It is open and links this PR; any earlier "Ticket:" line on this PR is superseded. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
work/5950-v17: affected_set, ladder, LAB workflow and merge-group fixes (7 commits)
Linear on main, no merge commits, one PMAT-Ticket per commit (work/5950-v17 @aae18cfeec). Supersedes #6007 (v16: MG timeout 120 < inner 220 min), #6006 (v15) and #5991 (v13) and the v12 branch. v16: the affected_set --self-test left x86-main and is now a guards-nightly LAB step with a lab_ticket opener and manifest row (wired check: 3 unwired, ratchet PASS); merge-group timeouts 100->120, above the shard's inner 110-min step. Built by w5032. Not armed; seats pending.
v17 fix (w5032): both MG jobs set FAT_TIMEOUT_SCALE=1 under merge_group, so the nextest step is 110 min, under the 120 wall. Evidence, 14 MG runs: shard n=27 p90 30 / max 43 min; x86-main n=14 p90 38 / max 42 min. fat_driver self-test 60/60.
ont-delta: none no new type, shape or reason; fold of existing tickets
🤖 Generated with Claude Code