Repository navigation
Spec Guards and Seal Coverage red on master: 8 seals stale after #10014 edited four fpga/board specs #10052
Description
Problem
Spec Guards and Seal Coverage are red on master (60d332d, d03afb0) and therefore red on every open PR,
including PRs that touch nothing near these specs (for example #10048, which adds one spec file).
python3 tools/check_seal_currency.py on master reports STALE generated-code hash: 8:
| Seal | Spec |
|---|---|
| BoardMinimalXC7A100T.json | specs/boards/xc7a100t_minimal.t27 |
| boards_BoardMinimalXC7A100T.json | specs/boards/xc7a100t_minimal.t27 |
| boards_BoardWukongV1.json | specs/boards/wukong_v1.t27 |
| Placement.json | specs/fpga/placement.t27 |
| fpga_Placement.json | specs/fpga/placement.t27 |
| Stdlib.json | specs/fpga/stdlib.t27 |
| fpga_FpgaStdlib.json | specs/fpga/stdlib.t27 |
| fpga_Stdlib.json | specs/fpga/stdlib.t27 |
Cause: #10014 (2f39502, 2026-10-17) edited all four specs and did not reseal them. The seals were last written
on 2026-10-02 to 2026-10-10.
Repair
Use the repair the check itself names: t27c seal <spec> --save && tri seals sync-twins for each of the four specs.
Run it on the Railway lab, not on a Mac. Do not use --update-baseline. The check says baselining gen-drift is
the wrong repair, because it would record the drift as accepted debt instead of recording what the compiler now
produces.
The PR holds only tool-written .trinity/seals/*.json files.
Done when
- The PR with
Closesfor this issue is open and holds only tool-written seal files. python3 tools/check_seal_currency.pyon the lab reportsSTALE generated-code hash: 0.- Spec Guards and Seal Coverage pass on that PR.
Separate: cli-tri is also red on master (not fixed here)
cli/tri/src/gates.rs:9012, in the unit check the_motivating_workflow_records_its_own_reason, runs
include_str!("../../../.github/workflows/brain-seal-refresh.yml"). #9117 (9a092df) deleted that workflow, so
cli-tri no longer compiles its checks. The repair is an edit to hand-written Rust, which needs the owner's
owner-approved-foreign label. It is only reported here; this issue's PR does not touch gates.rs.
Options for the owner: delete that one check (its motivating workflow is gone), or point it at an inline
fixture string that carries the tri:cause-removed marker.
Activity
github-actions commentedon Oct 17, 2026
Closed by the issue-shape gate (specs/policy/issue_shape.t27, #7498): an issue must have the template's sections. Edit the body; an edit that passes reopens it.
- specs/x.t27); without it the Queen reserves nothing and no bee can take the issueFR-001: ... MUST ...seals: reseal the four fpga/board specs #10014 edited (Closes #10052) (…
gHashTag commentedon Oct 17, 2026
Two checks fail on this PR. Neither comes from its diff, which holds only 8 seal JSON files:
cli/tri/src/gates.rs:9012usesinclude_str!on.github/workflows/brain-seal-refresh.yml, which ci: one queen-publish job opens the derived-record PRs the labs' verdicts allow (Closes #9092) #9117 deleted. Fixing it needsowner-approved-foreign.specs/tri/t27b/ts_emit.t27, and 18 of its function bodies (put_ind,to_front,sym_of,tnode,put_bytes,resolve, ...) are copies of the bodies inspecs/tri/t27b/verilog_emit.t27. They need to be shared through one module, or blessed deliberately withtools/dupe_scan.py --bless. That is a decision about t27b: ts -- consts, scalar fns and tests as TypeScript, decided in .t27 plans (Closes #9994) #10047's design, so it is left out of this seal-only PR.Once this PR merges, Spec Guards and Seal Coverage should go green on master.