Skip to content

docs(analog): decision (c) for the two load constants, "no output changes" corrected (Closes #12069) - #12942

Merged
gHashTag merged 1 commit into
masterfrom
lane/analog-12069-decision-c
Oct 31, 2026
Merged

docs(analog): decision (c) for the two load constants, "no output changes" corrected (Closes #12069)#12942
gHashTag merged 1 commit into
masterfrom
lane/analog-12069-decision-c

Conversation

@gHashTag

@gHashTag gHashTag commented Oct 31, 2026

Copy link
Copy Markdown
Collaborator

What

Applies the default that #12069 states in its Requirements for the case that nobody chose by 2026-10-27: option (c), keep SL_DC_LOAD_R and SL_TG_IN_LOAD of specs/analog/sar_logic.t27 as documented survivors. The issue had no comment or edit after 2026-10-24 when this was read on 2026-10-31.

Docs only. No spec, seal or generated file is touched, and no non-t27 code is added.

Changes

  • docs/analog/seal-refusal-2026-10-25.md: new section 8 "Decision (c), 2026-10-31" records the decision, what analog: four single-line constants in sar_kick_net and sar_logic that no test pins #12067 met and did not meet, and the follow-up option (a) (make the load a parameter), not started.
  • Honesty correction in three places (section 6 of the note, the SC-002 row of docs/analog/v0.1-release-notes.md, the row of docs/analog/loop-report-2026-10-26.md): the phrase "change no output of the integer-ps model" was too strong. The probes of section 7 show that three of the four plants leave both probe sums identical, while SL_TG_IN_LOAD 429 moves one probe sum by 55 ps, below every test bound. The new wording says "no test gives a signal".

Not claimed

No speed or accuracy figure. The two constants are fitted values; the note only says that the test suite cannot tell a step of 1 from the fitted value.

Closes #12069
Refs #12067
Refs #10486

🤖 Generated with Claude Code

@gHashTag @claude
docs(analog): decision (c) for the two load constants, "no output cha…
875e364 
…nges" corrected (Closes #12069)

Applies the default that #12069 states for the case that nobody chose by
2026-10-27: both load constants of sar_logic stay as documented survivors,
no non-t27 code, no spec or seal touched.

- seal-refusal note: new section 8 records the decision, the rescoped
  success criterion of #12067 and the (a) follow-up option.
- The phrase "change no output of the integer-ps model" overclaimed:
  SL_TG_IN_LOAD 429 moves one probe sum by 55 ps, below every test bound.
  Replaced by "no test gives a signal" in the note, the SC-002 row of the
  v0.1 release notes and the loop report row.

Refs #12067
Refs #10486

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

github-actions Bot commented Oct 31, 2026

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-10-31 02:15:34 UTC

Summary

Status Count
Total Open PRs 28
PRs with Failing Checks 27
PRs with All Checks Green 1
READY 0
FAILING 27
PENDING 0
NO CHECKS YET 0

These columns do not partition: 0 + 27 + 0 + 0 = 27, and there are 28 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).

@gHashTag

gHashTag commented Oct 31, 2026

Copy link
Copy Markdown
Collaborator Author

Merge basis (read at 09:20 +07 on 2026-10-31, head 875e36412e2b85cf6df0ed6284a9890a8c0e76d6):

  • The four required checks of ruleset t27-master-protection pass: validate, check-linked-issue, parse-ratchet, own-language. 23 checks pass, 4 are skipped, 1 is pending (Corpus ratchet; it reads specs, and this diff holds three .md files only).
  • Four checks are red; I read each log and none involves a file of this diff (git diff --name-only origin/master shows only docs/analog/loop-report-2026-10-26.md, docs/analog/seal-refusal-2026-10-25.md, docs/analog/v0.1-release-notes.md):
    • gate-topology, step "Gate inventory must match the workflows directory": t27os and t27os-release are in specs/claw/gates.t27 but their workflow files are gone, and stale-arm-repair (added on master by crons: the stale-arm repair workflow live -- paid by the check-conflicts.sh deletion (Refs #12606, #12813) #12846) is not in the spec. Master's drift.
    • cli-tri, step "No census moved without saying so": "workflow files read" 71 -> 70, the same two removed workflow files. Master's drift.
    • coverage, step "Every seal still describes its spec": gen-drift on specs of other lanes (specs/ml/transformer/*, specs/tri/trees/*, specs/tri/utils/bytes.t27 and others); no analog spec is named. Master's drift.
    • Verilog widths, step "Scan the corpus": specs/t27os/pty_server.t27 no longer carries an absurd width and must be dropped from the ledger. Not touched here.

No spec, seal or generated file is changed, so none of the four can move with this diff. If Corpus ratchet ends red after the merge, I will name it in a follow-up comment.

🤖 Addressed by Claude Code

@gHashTag
gHashTag merged commit 763a892 into master Oct 31, 2026
27 of 32 checks passed
This was referenced Oct 31, 2026
Closed
Closed
Closed
Merged
Closed
Merged
gHashTag added a commit that referenced this pull request Oct 31, 2026
@gHashTag @claude
docs(analog): loop report update after the epic closed (Closes #12946) (
3267743 
#12947)

Appends section 9 to docs/analog/loop-report-2026-10-26.md, 25 lines added,
0 deleted: #12116 closed by its default on 2026-10-27, decision (c) of
#12069 applied by PR #12942 with #12067 closed with its deviation, epic
#10486 closed by PR #12838 with the cooperate and speed-up specs. The
superseded lines of sections 1, 4 and 8.3 are named and left as written.
No spec, seal or generated file, no new claim.

Refs #10486

Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
@gHashTag gHashTag mentioned this pull request Oct 31, 2026
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

analog: make the two sar_logic load constants visible to a test (follow-up of #12067)

1 participant

@gHashTag