You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
LAB: complexity ratchet (check_complexity_ratchet.sh) runs nightly, off the PR path #5217
LAB check: scripts/check_complexity_ratchet.sh (per-function complexity may only fall).
Under the GATE/LAB rule a ratchet may not sit on a PR, merge-queue or release path. #5182 moves this check from x86-main's guard-cargo section (and guard_tree's glob) to .github/workflows/guards-nightly.yml, where it runs nightly on main as the step "Per-function complexity may only fall (shrink-only)" (the --selftest case table, then the check).
Owner: aprender-a7 (one owner).
A red in the nightly updates THIS issue and nothing else: no PR, merge or release is blocked by it.
Red or not measured for 7 days: the step stops running until its owner brings it back.
Known red at move time: the recorded baseline came from an older pmat than the one now installed, so the check reports a tool-version mismatch. That is the version dependence C324 #6: take complexity/baseline ratchets off the PR path (after 0.71.1 tag) #5182 names. It is fixed by re-recording the baseline under the pinned pmat, not by loosening the comparator.
The comparator's case table (ratchet_semantics_test.sh --class complexity, over a pmat shim) stays on the PR, because it judges the script, not the tree.
Owner update (main @ 6bd89b0): the "known red at move time" is not a baseline defect, so nothing gets re-recorded.
tools.toml pins pmat 3.41.1 (and bashrs 7.4.1). Both scripts/complexity_baseline.txt and scripts/cb200_baseline.txt already carry # tool_version=pmat 3.41.1, so the baselines agree with the pin.
Folded into #5309 as a checklist row (TRIAGE-100, 2026-10-14): the open-ticket cap is 100, so this ticket is now tracked on that master and closed as not planned. Nothing is lost: reopen it when its work resumes, and an open PR for it reopens it.
Activity
noahgift commentedon Oct 12, 2026
Owner update (main @ 6bd89b0): the "known red at move time" is not a baseline defect, so nothing gets re-recorded.
tools.tomlpins pmat 3.41.1 (and bashrs 7.4.1). Bothscripts/complexity_baseline.txtandscripts/cb200_baseline.txtalready carry# tool_version=pmat 3.41.1, so the baselines agree with the pin.scripts/check_tool_versions.shfails closed (pinned 3.41.1 found 3.42.0). C324 #6: take complexity/baseline ratchets off the PR path (after 0.71.1 tag) #5182 records that both checks pass under the runner's pinned pmat.noahgift commentedon Oct 14, 2026
Folded into #5309 as a checklist row (TRIAGE-100, 2026-10-14): the open-ticket cap is 100, so this ticket is now tracked on that master and closed as not planned. Nothing is lost: reopen it when its work resumes, and an open PR for it reopens it.
Agent: aprender-clerk