[AdaWorldAPI]

Implements D-DT64-7 §1a/§1b (plan on #1898) for the relational polyfill.

The problem

binding::prepare turned an algebra word into an opcode by reading the view's label and searching R2ILVocabulary for that name. The module doc said it directly: "the position a lens gives a word is free and only the name binds".

What changed

• New module polyfill. It defines POLYFILL_LENS = 0xFFFF_F000 (left side 0xFFFF = hardcoded, right side 0xF000 = the relational polyfill) and POLYFILL: 17 positions in EXECUTABLE order. Each position carries:
  • its opcode;
  • its arity;
  • its aliases: the R2IL name, a SQL name where one fits, and the Quack variant.
• binding::prepare accepts this lens.
  • Under it, word p is POLYFILL[p], and the view's label at p must be one of p's aliases.
  • A name that belongs to another position is refused with the new ExecRefusal::NotAnAlias. It never selects that other word.
  • The R2IL fold reading (0xC401) keeps its label binding unchanged.
• New dev-dependency on lance-graph-quack, used by the parity tests. Quack is OGAR-free and does not depend on this crate.

SQL via Quack

Quack has no SQL text parser. Its SQL surface is the typed Query { Filter, Agg }, so SQL registers through Quack's variant names. Methods registered by those names give exactly the same answers as Quack's own lower + mask-risc execute over the same lanes:

• 21 cases: Cmp::LtI32 combined with each of Agg::{Count, Any, All, SumI32, MinI32, MaxI32, CountDistinctOrderedU32}, over a partial, a full and an empty selection.
• One compound filter using Cmp::EqI32, Filter::negate and Filter::and.

The SQL keyword aliases (<, AND, COUNT, EXISTS, …) are recorded in the table, but no parser in the tree produces them yet.

Tests and disable runs

All tests are in tests/integration/polyfill.rs. Each guard was disabled once to confirm a test fails:

guard disabled	tests failing
alias check	1
positional path, falling back to label binding	4
lens not accepted	5
SUM and MIN opcodes swapped	3
duplicate alias	1
table arity check	1

The table-arity disable was green at first, because registration already ties the argument count to the declared arity. A test with a wrong runtime declare_algebra now makes that check load-bearing.

Also checked: the table matches R2ILVocabulary (name, arity, one result); R2IL, SQL and Quack views bind the same steps and the same plan id.

Local gates: cargo test -p lance-graph-r2il (all suites), clippy --all-targets -D warnings and fmt are clean; entries_index, citation_decay and append_only_gate pass.

Not covered

• Quack variants with no single executable word: Rows, BlendI32, ScatterOrU32, the Group* aggregates, Cmp::{Ne, Le, Gt, Ge}, the ternary matches, the strided compares and Range.
• Empty selections: SUM, MIN and MAX return the fold identity where SQL returns NULL.
• Still runtime-declared: the lens goes through declare_algebra and its opcodes are ogar-loco values. DT64 acting as the authority for 0xFFFF is a later item (D-DT64-7-4, O8).

Board entry: .claude/board/entries/2026-10-31-polyfill-positional-lens.md.

---

Generated by Claude Code

[coderabbitai]

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: 8ee4d29c-5671-4885-b80c-f6e98c9d3533

• 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

• X
• Mastodon
• Reddit
• LinkedIn

[chatgpt-codex-connector]

Codex Review Summary

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

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-31T21:00:00.371176Z	7d58148	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.