Skip to content

seals: re-save 128 seal files for 70 specs (Refs #12972) - #12985

Merged
gHashTag merged 2 commits into
masterfrom
claude/ecstatic-mclean-c5cf93
Oct 31, 2026
Merged

gHashTag merged 2 commits into
masterfrom
claude/ecstatic-mclean-c5cf93

Conversation

@gHashTag

@gHashTag gHashTag commented Oct 31, 2026 •

Copy link
Copy Markdown
Collaborator

Refs #12972

What

Re-saves 128 seal files (70 specs) that the coverage check ("Every seal still describes its spec") lists as stale or gen-drifted on master 4cfd2f2 (Seal Coverage run 41172959376: 135 seals, 126 gen-drift, 9 stale).

  • Cause of the 126 gen-drift seals: the gen-rust slice-field capture change (afea7d4) moved gen_hash_rust of specs that were not touched. Re-sealing says the new output is the wanted one.
  • Cause of the 9 stale seals: specs edited without a re-seal.
  • Method: t27c seal <spec> --save from a compiler whose bootstrap/ and cli/tri sources equal master (no diff against origin/master). Twin seal files naming the same spec were updated by the same command. No absolute home path appears in the diff or in .trinity/seals.
  • Second commit: drops 23 lines from tools/seal_baseline.txt (the baselined gen-drift seals that now hold: the 29 gen-drift entries minus config-migrate, runtime-execute, server-router and their twins, which spec-guards still reports). Deletions only. No spec, compiler source or workflow file is edited.

Not in this PR (7 of the 135 seals)

Seal Spec Why left out
parser.json, parser_parser.json compiler/parser/parser.t27 gen-drift; needs a master-built t27c, none is on disk
t27os_T27osGlassWindow.json, t27os_T27osKeys.json, t27os_T27osPtyLogic.json, t27os_T27osQueenDash.json specs/t27os/{glass_window,keys,pty_logic,queen_dash}.t27 stale; same reason
fpga_fpga_wasm_board.json specs/fpga/wasm_board_ax7203.t27 stale; the spec does not generate (emit-bitexact, out of scope); not written with gen_hash=none

The compiler binary used for the 128 was built in a sibling worktree and has since been cleaned; the Mac has 4 GiB free and load 24, so no build was run here (heavy builds belong on the lab or a T27 OS pane). So coverage is expected to go from 135 listed seals to these 7, not to zero. Issue #12972 stays open for them.

Heads-up: PR 12967 edits gates.t27, which makes claw_ClawGates and claw_ClawGatesPins stale again when it merges.

🤖 Generated with Claude Code

…e ones (Refs #12972)

The gen-rust emitter change afea7d4 moved gen_hash_rust of specs that were
not touched, and a handful of specs were edited without a re-seal, so the
coverage check lists 135 seals that no longer describe their spec.
This re-saves 128 of them with t27c seal --save from a compiler whose
bootstrap sources equal master. The remaining seals (compiler/parser/parser.t27,
specs/fpga/wasm_board_ax7203.t27 and four specs/t27os specs) need a freshly
built compiler and follow in a later commit.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-31 07:37:27 UTC

Summary

Status Count
Total Open PRs 34
PRs with Failing Checks 33
PRs with All Checks Green 1
READY 0
FAILING 33
PENDING 0
NO CHECKS YET 0

These columns do not partition: 0 + 33 + 0 + 0 = 33, and there are 34 open PRs. A PR is being counted twice or not at all.

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=e75b498e5e79 != 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).

…12972)

The coverage run prints "NOTE 23 baselined seal(s) now hold. Drop their
lines so the gate holds them". They are the 29 gen-drift entries of
tools/seal_baseline.txt minus the six that spec-guards still reports as
drifted (config-migrate, runtime-execute, server-router and their twins).
Deletions only; no seal, spec or compiler file is touched.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-31 08:05:50 UTC

Summary

Status Count
Total Open PRs 36
PRs with Failing Checks 36
PRs with All Checks Green 0
READY 0
FAILING 36
PENDING 0
NO CHECKS YET 0

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=e75b498e5e79 != 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).

@gHashTag
gHashTag merged commit 8255399 into master Oct 31, 2026
27 of 34 checks passed
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.

1 participant