lance-graph-dt64: the bind store for meaning and behaviour by AdaWorldAPI · Pull Request #1869 · AdaWorldAPI/lance-graph

AdaWorldAPI · 2026-10-30T16:07:37Z

Why a new PR

#1862 merged at head 4707683dc, before any lance-graph-dt64 commit was pushed. The crate and its follow-ups were left on the branch. This PR rebases them onto current main.

What

crates/lance-graph-dt64 is the DT bind store for meaning and behaviour (D-DRDT §1/§4, D-DT64-6 §1/§3). It is a workspace member that depends only on lance-graph-contract, holds bindings and their version log (never business data), and evaluates nothing.

Meaning (the classid is the lens)

view_class(classid) = ogar_codebook::classid_concept. Labels are resolved late through the ClassView.
field_view(&WideFieldMask) is the ERB/Redmine fieldview pattern, the same shape as OGAR ogar-render-askama::field_view minus the value: FieldRow { position, label, predicate }. action_ordinals() lists the actions.
field_meaning / mask_where / unspecified join each row to its property::PropertySpec (obligation, semantic type, marking). A risk mask is marking >= Pii. A field with no spec is UNKNOWN and never takes a default.

Behaviour (the classid switches over agnostic algebra)

Algebra methods form a classid-free catalogue (AlgebraEntry: id, sig_rev, polyfill, arity; immutable; D6).
A consumer classid's methods are MethodBindings:
- identity: MethodId + SigRev;
- ON / reads: WideFieldMask;
- WHEN: a K3 Guard (SetExpr + Logic);
- DO: ordered Effect { target, algebra: AlgebraRef, args };
- Population / Truth / value demand.
register refuses: unknown algebra, revision or arity mismatch, out-of-basis positions, duplicate MethodIds, and UndeclaredRead (a step reading what no earlier step wrote).
D5: a classid's behaviour is immutable; an override is register_derived plus a re-route. Only routes are versioned (routed_at).
resolve / resolve_at / admit give named refusals, never an empty answer. admit covers a dropped, added or mismatched guard and a stale SigRev.
footprint() / methods_touching(mask) let a meaning mask reach behaviour.

The XLOG session shaped much of this. Its per-Y count (group-count twice, then checked i64 Mul) and an ONNX Mul(int64) node reuse one catalogue entry, and only the two consumers carry classids.

Board: PR_ARC_INVENTORY / LATEST_STATE records for #1862; plan Amendment A.5 and its addenda.

Tests

cargo test -p lance-graph-dt64: 19 pass. clippy --all-targets -D warnings and fmt are clean.
Every guard was disable-checked: removed, its test red, restored. The runs are recorded in plan A.5.

Open, with the owners (EXO-50 27–29)

Who declares the algebra catalogue entries and their ids; whether tract's per-op classids become catalogue entries.
MethodSignature (dr-math) vs MethodRef identity.
Whether a frame records the catalogue's extent (D8).
OGAR CapabilityAuthority wiring (D6); persistence (EXO-55).

Generated by Claude Code

DT is the existing DTO binding store, addressed by ClassId (D-DRDT §1). This crate is that store as production code: a workspace member whose only dependency is the zero-dep contract. It holds bindings and their version log, never business data, and evaluates nothing. Closes the four D-DRDT §4 gaps: - MethodId + SigRev: identity independent of the source name; - Guard = set_algebra::SetExpr + Logic, the description Quack's WHERE evaluates; Guard::witness is its normal form (Kleene for K3); - ordered effects (field position + AlgebraRef, the D4 polyfill ref); - one lookup: resolve(view, classid, schema_rev), fields borrowed from the ClassView, methods from the store; resolve_at by DR position. D5: register is idempotent for an identical binding and refuses any other (BehaviourImmutable); an override is register_derived plus a re-route. Only routing is versioned: an append-only route log with routed_at(version, position). D-DRDT §5: admit(MethodRef) refuses a dropped, added or mismatched guard and a stale signature revision; resolve refuses another schema revision and a ClassView whose field basis moved. The first DR #4 falsifier passes: equal read/write sets, different effect order, both kept and distinct. 11 tests; clippy -D warnings and fmt clean. Board: STATUS_BOARD D-DT64-6-BS, plan Amendment A.5, entry 2026-10-30-d-dt64-6-bs. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

A classid is a codebook lens, never one per item (operator). So: - ClassSchema drops the declared view class; view_class(classid) is ogar_codebook::classid_concept, the codebook's own split. - Effect carries EffectOp::{Algebra(AlgebraRef), Call(MethodRef)}. A Call reuses another lens's entry by reference, is admitted when the caller registers (CallRefused wraps the reason), and callers_of(lens) lists who reuses it. Shaped by the XLOG session's critique: its per-Y count is three Calls into one algebra lens (group-count twice, checked i64 Mul), two classes bound in total. Route targets stay bare classids. 14 tests; clippy and fmt clean. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

…nature The XLOG session found that a Call recorded where a result goes but not what it reads: the binding could not tell counting R and S apart from counting R twice. Effect now carries ordered args (Field position or Const). MethodBinding::inputs is part of the signature, and a Call with another arity is refused at registration (ArityMismatch in CallRefused). Field args are checked against the caller's basis. 15 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

Operator: the classid switches behaviour, labels and polyfill over otherwise agnostic addresses and algebra methods. The algebra lens classid and class-to-class Call were wrong and are removed. The store holds a classid-free algebra catalogue (AlgebraEntry: id, sig_rev, polyfill, inputs; immutable; D-DT64-6 D6). A consumer's effects bind an AlgebraRef with ordered operands, checked at registration (UnknownAlgebra, AlgebraRevisionMismatch, ArityMismatch). users_of lists the classids switching over an entry. XLOG's count and an ONNX Mul(int64) node reuse one entry; only the consumers carry classids. 15 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

Operator: classid x WideFieldMask ergonomics is the ERB Redmine fieldview pattern, as in OGAR ogar-render-askama::field_view. Resolved::field_view(mask) yields FieldRow { position, label, predicate } ascending, the position as layout address, labels borrowed late from the ClassView (OGAR FieldView minus the value; the store holds none). Resolved::action_ordinals() gives the methods as action addresses. 16 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

DT owns field semantics; the contract carries them per predicate as PropertySpec (obligation, codec route, semantic type, marking). field_meaning joins fieldview rows to their spec; mask_where derives WideFieldMasks from meaning (risk: marking >= Pii; obligation: Required); unspecified surfaces fields with no spec, which never take the spec default. 17 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

…ing masks register refuses UndeclaredRead: a step may read only declared reads or what an earlier step of the same method wrote, so a recipe's order is what it says. MethodBinding::footprint (ON, reads, writes) and Resolved::methods_touching(mask) let a risk or RBAC mask reach behaviour. 19 tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

#1862 merged at head 4707683, before the lance-graph-dt64 commits were pushed. PR_ARC_INVENTORY and LATEST_STATE record what landed and that the crate follows. Entries index regenerated after the rebase. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

coderabbitai · 2026-10-30T16:07:46Z

Important

🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration

Source mode: Auto
Repository: Unavailable
Branch: Unavailable
File: No configuration file from this repository applied
Commit: Unavailable
Layers: Organization UI
Review profile: CHILL
Plan: Advanced
Run ID: 4877c87c-cde6-4962-8c09-e6ee31dae226

Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands.}

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

chatgpt-codex-connector · 2026-10-30T16:09:59Z

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-30T16:10:32.300225Z	`1761d4c`	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

…-op classid question STATUS_BOARD: D-DT64-6-BS row is Shipped #1869; the D-TRACT-HARVEST-0 row still said "op signature -> classid", which #1870 corrected in the plans but not on the board. Plan D-DT64-6 Amendment A.5 addendum 7: one ONNX lens classid with op signatures as fields is the shape lance-graph-dt64 implements; the MethodKey vs MethodRef identity stays open. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

…1869 and #1871 PR_ARC_INVENTORY gains the #1869 and #1871 entries that were missing, LATEST_STATE records main after them, STATUS_BOARD moves the D-DT64-6-BS row, and a board entry records the dr module. Addendum 12 also records that #1871 merged without the thin-provisioned extent, which this branch restores. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

claude added 10 commits October 30, 2026 16:05

D-DT64-6 A.5: record the eight lance-graph-dt64 disable runs

c46645f

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

D-DT64-6 A.5: the fieldview is one masking pattern, not render-only

3745f47

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

STATUS_BOARD: D-DT64-6-BS is in PR #1869

1761d4c

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XAqKMPh2Seh2WLccCksAdY

AdaWorldAPI marked this pull request as ready for review October 30, 2026 16:09

AdaWorldAPI merged commit 9d119bb into main Oct 30, 2026
12 checks passed

AdaWorldAPI mentioned this pull request Oct 30, 2026

lance-graph-dt64: polyfill as one lens classid; thin-provisioned DR extent; board catch-up; fmt fix #1871

Merged

Navigation Menu

Repository navigation

lance-graph-dt64: the bind store for meaning and behaviour - #1869

lance-graph-dt64: the bind store for meaning and behaviour#1869
AdaWorldAPI merged 11 commits into
mainfrom
ccr-45a41fb4-0j3i0l

AdaWorldAPI commented Oct 30, 2026

Uh oh!

coderabbitai Bot commented Oct 30, 2026 •

edited

Loading

Uh oh!

Uh oh!

chatgpt-codex-connector Bot commented Oct 30, 2026 •

edited

Loading

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

AdaWorldAPI commented Oct 30, 2026

Why a new PR

What

Tests

Open, with the owners (EXO-50 27–29)

Uh oh!

coderabbitai Bot commented Oct 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot commented Oct 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codex Review Summary

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

coderabbitai Bot commented Oct 30, 2026 •

edited

Loading

chatgpt-codex-connector Bot commented Oct 30, 2026 •

edited

Loading