Repository navigation
specs/queen/dispatch.t27: 39 of 134 mutants survive on master (land after #6934) #7430
Description
Problem
tri mutate spec --file specs/queen/dispatch.t27 on master 05e633d03 (Railway lab, master release build, --jobs 8 --timeout 60 --zig-threads 6): 95 of 134 killed, 39 survived, 0 hung, 0 unviable. The spec has 6 tests. It was left out of #7077's pass because open PR #6934 edits it, and it has had no whole-file pass since.
Survivors by function (line numbers on 05e633d03, spec sha256 prefix 271df7d7c08c):
| Function | Survivors | Lines, mutant kind |
|---|---|---|
next_state |
24 | 181 drop-guard + flip-cmp; 186, 187, 191-197, 201, 211 drop-guard; 190, 205 swap-logic; 206, 207 drop-guard + flip-cmp; 214 3 flip-cmp + 2 swap-logic |
claim_of |
8 | 224 drop-guard + swap-logic; 229 drop-guard; 231 drop-guard + flip-cmp; 236 swap-logic; 237 flip-cmp + swap-logic |
chosen_of |
3 | 256, 259, 260 drop-guard |
column_of |
1 | 250 drop-guard |
counts_against_issue |
1 | 283 ret-default |
stall_reapable |
1 | 300 flip-cmp |
followup_rounds |
1 | 304 flip-cmp |
24 drop-guard survivors in next_state mean most of its transitions are never asserted: removing the guard that selects a transition changes no test result.
Expected
- Every survivor is classified, one line each, as a dead line (removed), a test gap (an assert added; the mutant applied by hand first and seen to die, then seen to survive on the parent), or an equivalent mutant (the reason written down), as specs/queen: whole-file mutation pass over review_valve, priority, task_analysis, views (and lotus, brain_summaries) #7077 did for the other Queen specs.
- Asserts only where a test gap is found; no change to what a function returns.
Acceptance criteria
t27c parse,typecheck,gen-rust,gen-verilog,gen-cexit 0 on the spec.t27c test-report specs/queen/dispatch.t27printsvacuous passes 0 of N.- A second
tri mutate spec --file specs/queen/dispatch.t27leaves only survivors this issue's PR names as equivalent.
Order
Do not start before #6934 is merged or closed: it edits this file, and a second branch on the same lines makes a conflict, not progress.
Boundary
specs/queen/dispatch.t27
.trinity/seals/queen_QueenDispatch.json
User Scenarios
-
Given the
specs/queen/dispatch.t27file with 39 surviving mutants on master commit05e633d03, When the classification work is complete, Then every survivor is categorized as a dead line (removed), a test gap (assert added and verified), or an equivalent mutant (reason documented), and a subsequenttri mutate spec --file specs/queen/dispatch.t27reports only the survivors explicitly named as equivalent in the PR. -
Given the spec passes all compilation and generation steps, When
t27c test-report specs/queen/dispatch.t27is run, Then it printsvacuous passes 0 of Nindicating no vacuous test passes remain.
Requirements
- FR-001: All 39 surviving mutants MUST be classified, one line each, as dead line, test gap, or equivalent mutant with written justification.
- FR-002: Test asserts MUST be added only where a test gap is identified; function return behavior MUST NOT be changed.
- FR-003:
t27c parse specs/queen/dispatch.t27MUST exit 0. - FR-004:
t27c typecheck specs/queen/dispatch.t27MUST exit 0. - FR-005:
t27c gen-rust specs/queen/dispatch.t27MUST exit 0. - FR-006:
t27c gen-verilog specs/queen/dispatch.t27MUST exit 0. - FR-007:
t27c gen-c specs/queen/dispatch.t27MUST exit 0. - FR-008:
t27c test-report specs/queen/dispatch.t27MUST printvacuous passes 0 of N. - FR-009: A subsequent
tri mutate spec --file specs/queen/dispatch.t27MUST leave only survivors that the PR explicitly names as equivalent mutants. - FR-010: Work MUST NOT start before PR Queen: log why a bee run ended with no diff (56% of queen-* branches in 7 days, cause not recorded) #6934 is merged or closed to avoid conflicts on the same file.
Activity
github-actions commentedon Oct 17, 2026
Re-filed as #9886.
An attempt claimed this issue on 2026-10-07 and left an empty branch. A dispatch row that has spent its retry ceiling is released once and then keeps its issue for good, which is the right rule for a third identical failure and the wrong outcome for an issue nobody ever attempted twice. The new issue carries the same work, the same boundary, and today's brief.
Closing this one so the swarm can take the work again.