Repository navigation
claim-guard-bar: kill 5 surviving mutants and scale the honest-category limit #6679
Description
Follow-up to #6662 (merged 479990b), from the overnight improvement loop the owner asked for on 2026-10-06.
The round-2 reviewer's 28 new mutants left 5 survivors. None of them changes a verdict on the current set sizes, but each is a check that no test exercises:
- N16, N17: the round-up and round-down helpers are off by one at a remainder of 1 or d-1.
- N19: two EN table cells can be swapped without changing the sum or the zero count.
- N21: no passing test sits at exactly 200/100 for the minimum held-out size.
- N22: the control corpus silent-row figure (18) is not pinned.
The reviewer posted a diff that kills all 5, tested on a copy (21/21, vacuity 0, 28/28 killed): #6662 (comment). Landing it needs a reseal.
Design question. "At most 2 per honest category" is a flat count, so on a 5-row dev category it allows 40%. Decide whether the limit should be a percentage with a floor of 1 row.
Also: the PR body of #6662 still describes passes() and 15 tests.
Activity
test(spec): claim-guard-bar -- kill the five surviving review mutants (…
gHashTag commentedon Oct 6, 2026
Reopened after #6694 (merged as e20c1d4). The five survivors are killed. Two items are still open. (1) The design question: should 'at most 2 per honest category' become a percentage with a floor of 1 row? (2) The #6662 PR body still describes passes() and 15 tests. An optional third item: a 3-cell compensated table drift (+1/-2/+1 on evenly spaced axes) still survives the 1..14 weighted checksum. A base-8 packing fix was tested on a copy in #6694 (comment).
gHashTag commentedon Oct 7, 2026
Reopened 2026-10-07: the bulk close of issues without a
## Boundarysection was reverted by the owner. Nothing to do here; this comment replaces the earlier close notice.tri-claw commentedon Oct 8, 2026
The Queen cannot take this issue yet. A bee starts only on an issue whose body has every section below, and this one lacks:
- specs/x.t27); without it the Queen reserves nothing and no bee can take the issueFR-001: ... MUST ...exit 0, or a backticked commandShe tried to write them 2 times and could not (last: not a task: The issue references PR #6662 and a reviewer's diff in a comment but does not name any concrete file paths in this repository; the provided CANDIDATE PATHS list is empty, so no boundary can be establi). Add them and the next round takes it. The rule is
specs/policy/issue_shape.t27.