Skip to content

ci(seals): re-seal 6 specs stale after wave-1 merges; drop 9 held baseline lines (Closes #11660) - #11661

Merged
gHashTag merged 1 commit into
masterfrom
fix/reseal-wave1
Oct 22, 2026
Merged

gHashTag merged 1 commit into
masterfrom
fix/reseal-wave1

Conversation

@gHashTag

Copy link
Copy Markdown
Collaborator

Fixes both Spec Guards and Seal Coverage that went red on master after the wave-1 merges.

Locally re-verified on the exact master tree: STALE generated-code hash = 0; Seal Coverage OK with no NOTE.

Closes #11660

Co-authored-by: Claude claude@anthropic.com

…eline lines (Closes #11660)

Wave-1 merges changed six specs after their seals were last written, so
four gen hashes per spec described output the compiler no longer
produced (#6726 -> graph_dfs, #6927 -> claim-guard-bar; devices,
lanes_view, session_store, verdict_notes from the same wave). Re-seal
with the master t27c (t27c seal <spec> --save) to record current
output rather than baselining the drift.

claim-guard-bar additionally carried a deliberately failing placeholder
test (assert(1 > 2)) left behind by #6927 exit-code testing, which
refused every re-seal. Removed; the spec's 21 real tests all pass
(t27c test: failed=0).

Also drop the 9 baseline lines the coverage gate reports as now holding
(TernaryShift, TriHtml, TriTerminal, TriXml, encoding_TriHtml,
encoding_TriXml, isa_TernaryShift, tools_GftGeneralizeDemo,
utils_TriTerminal) so the gate holds them instead of excusing them.
Local gates: STALE=0, coverage OK with no NOTE.

Closes #11660

Co-authored-by: Claude <claude@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-22 15:39:26 UTC

Summary

Status Count
Total Open PRs 50
PRs with Failing Checks 48
PRs with All Checks Green 2
READY 1
FAILING 48
PENDING 0
NO CHECKS YET 0

These columns do not partition: 1 + 48 + 0 + 0 = 49, and there are 50 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=5aec047d42ec != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@tri-claw

tri-claw Bot commented Oct 22, 2026

Copy link
Copy Markdown

Summary by t27-bees

Checked by the t27 compiler

t27c parse, parse-complete and typecheck on each changed .t27 file at d538111. Nothing was generated, built or run. A pass means the file is well formed, not that it is right.

  • specs/automation/claim-guard-bar.t27: parses and typechecks.

Read by the model

This PR re-seals 6 specs that went stale after wave-1 merges, removes a deliberately failing placeholder test from claim-guard-bar.t27, and drops 9 held baseline entries for specs with known test failures. All gen_hashes and spec_hashes updated, indicating upstream changes propagated through the dependency graph.

Worth a look

  • .trinity/seals/TriGraphDfs.json:10-11 and .trinity/seals/graph_TriGraphDfs.json:27-28 — Both seals now report tests.blocked: "does not compile: spec.zig:5:7: error: duplicate struct member name 'std'". The generated Zig code for graph_dfs.t27 fails to compile, so Zig tests cannot run. If Zig codegen is required for this spec, this is a regression.

  • specs/automation/claim-guard-bar.t27:472-474 (removed lines) — The deliberately failing test assert(1 > 2) was removed. Its comment says it existed "to check exit code behavior." If the test harness or CI relies on at least one failing test to verify non-zero exit codes, that signal is gone.

  • tools/seal_baseline.txt — Nine tests-fail entries were dropped (TernaryShift, TriHtml, TriTerminal, TriXml, encoding_TriHtml, encoding_TriXml, isa_TernaryShift, tools_GftGeneralizeDemo, utils_TriTerminal). If any of these specs still fail tests on master, Seal Coverage will now flag them instead of treating them as known-bad baselines. The author reports local pass, but the baseline removal shifts the gate's tolerance.

  • .trinity/seals/automation_automation::claim_guard_bar.json:20 — The reseal shows "vacuous": 1. One of the 21 passing tests is vacuous (covers no new paths). Worth confirming this isn't the removed failing test being misclassified.

  • .trinity/seals/graph_TriGraphDfs.json:4-15 — New closure array lists specs/std.t27 with "spec_hash": "missing". This dependency appears unresolved in the seal metadata, which may indicate a tooling gap or a genuine missing spec.


Read: 9 files, +75 -59, the whole diff. Head d538111. Model: openai-compatible/nvidia/nemotron-3-ultra-550b-a55b. Commands: @t27-bees review | summary | help | pause | resume. Free for every repository.

@gHashTag
gHashTag merged commit bb736ac into master Oct 22, 2026
33 of 35 checks passed
@gHashTag
gHashTag deleted the fix/reseal-wave1 branch October 26, 2026 03:47
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

ci(seals): 6 specs stale after wave-1 merges (graph_dfs, claim-guard-bar, devices, lanes_view, session_store, verdict_notes) + 9 baseline lines now hold

2 participants