Repository navigation
types classification: write the verdicts for Entry, Inputs, SkillMeta (Refs #2774) - #9145
Merged
Merged
Conversation
… (Refs #2774, Refs #9137) The classified half of the ledger heal #9137 started: the ratchet ledger reached 89 names, but three of them had no written verdict, so the 'Every conflicted type name has a written verdict' step stayed red on master. The definitions accumulated across the 10-10/10-11 port waves (gen_hash_table #7069, the openxc7-synth d_* ports, the browseros skills ports, mvp_v3 #8807); both ledgers were last blessed at 3b69b91, before some of those landed. Verdicts, from reading all six bodies: - Entry: DISTINCT -- a consensus-chain step (mvp_v3) vs a hash-table bucket cell (port); zero shared fields, no import edge. - Inputs: DISTINCT (4 defs) -- the Verilog port-list convention: each openxc7-synth module's own wire bundle, three of them literally 'clk : bool' alone; blink_working.t27 already uses BlinkInputs. - SkillMeta: DRIFT -- one concept in two halves of one port; catalog's is the three-field projection of loader's loaded record. Also drops the stale AttentionOutput row from the DRIFT table in docs/TYPE_CONFLICTS.md: the name is in neither the ratchet ledger nor the tree's conflicted set, and its row is what kept the md count (51) above the json's DRIFT count (50). After this: 89 rows, 51 DRIFT / 38 DISTINCT, json verdicts and both md headers agree. tri types classified: OK (0 unjudged, 0 stale); tri types ratchet: CLEAN (89/89), both on a fresh origin/master checkout (4751426). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
gHashTag
enabled auto-merge (squash)
October 11, 2026 16:58
Contributor
gHashTag
added a commit
that referenced
this pull request
Oct 12, 2026
…9290) specs/claw/recipes.t27: 19 rows, one per verified recipe, parallel arrays pinned by tests (same discipline as the gates inventory). GATES are workflow stems joined to specs/claw/gates.t27 NAMES, never display names. Classes: C_AUTO 13 (loop fixes, PR, auto-merge), C_REVIEW 3 (prepared patch, person merges), C_OWNER 3 (never executed; report names the keystroke). corpus-ratchet carries 5 rows, and the types pair (bless + verdicts) is pinned to land in ONE PR -- the #9145 lesson. Unknown kinds fail closed to fallback-manual. Folds the claw-af-seed rows (specs/claw/autofix.t27, never PR'd) into this one table so a single SSOT remains; display names re-keyed to real stems. Ledger: str_eq 2 recorded in the same commit -- a deliberate local helper, same shape as gates.t27's (specs cannot import). The two remaining dupe errors (slices_equal, starts_with) are inherited master reds owned by the night loop's next train item; NOT blessed here. AF5 #9287 (owner 2026-10-12): t27o transpiles foreign code into .t27 with no AI in the loop; spec-first-port moves REVIEW -> AUTO on the first verified transpile. Direction recorded in the header, table stays truthful for today. Lab-verified: typecheck OK, 11/11 tests, 0 FAIL, 0 vacuous; dupe scan scoped. Closes #9282. Part of #9271 and #9281. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
gHashTag
added a commit
that referenced
this pull request
Oct 12, 2026
…0 names, 21 read and judged (11 DRIFT, 10 DISTINCT) (#9293) Port wave (#9258 trios-train-cpu, #9259 trios-crypto, #9188 hir-cost-models) grew the conflicted-type set by 21; observed stable at exactly 21 across two master tips (21c6171, f235d90; no port landed after 16:20:49Z). Every NEW name was read at both definition sites, evidence quoting declarations and import edges, per the #9145 standard -- then blessed in the same PR as the verdicts (a bless alone leaves the classified gate red). - docs/reports/type_conflicts.json: ledger 89 -> 110 - docs/reports/type_conflicts_classified.json: +21 verdicts (11 DRIFT, 10 DISTINCT), census 62/48, ASCII-escaped - docs/TYPE_CONFLICTS.md: +21 table rows, section counts 51->62 / 38->48 Verified on the branch: tri types ratchet RATCHET: CLEAN; tri types classified OK (every conflicted name has a written verdict, no stale verdicts). Refs #9281 Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The classified half of the heal #9137 started. The ratchet ledger reached 89 names, but the 'Every conflicted type name has a written verdict' step stayed red on master: Entry, Inputs and SkillMeta had no rows in
docs/reports/type_conflicts_classified.json.Root cause is older than #8823: the definitions accumulated across the 10-10/10-11 port waves (gen_hash_table #7069, the openxc7-synth d_* ports, the browseros skills ports, mvp_v3 #8807), and both ledgers were last blessed at 3b69b91 (#8535), before some of those landed. #9137 restored the ratchet ledger; this writes the three verdicts the classified gate demands, from reading all six bodies:
ev/op/job/who/arg/amt, indexed by the JO_* opcode block) vs gen_hash_table.t27:29 hash-table bucket cell (key/value/used). Zero shared fields, no import edge.clk : boolalone; blink_working.t27:26 already spells the per-module reading withBlinkInputs.Also drops the stale
AttentionOutputrow from the DRIFT table indocs/TYPE_CONFLICTS.md-- the name is in neither the ratchet ledger nor the tree's conflicted set, and that row is what kept the md count (51) above the json DRIFT count (50). After this PR: 89 rows, 51 DRIFT / 38 DISTINCT, and the json verdict block plus both md headers agree.Verified on a fresh origin/master checkout (4751426):
tri types classified-> OK, 0 unjudged, 0 stale;tri types ratchet-> CLEAN (89/89).Refs #2774, Refs #9137
🤖 Generated with Claude Code