[AdaWorldAPI]

Lab probe in the workspace-excluded drf-focus-probe crate (examples/global_union.rs). Its question: can 64 independently mapped 64×64 registers, each a SealedMap of u16 global ordinals plus a 4,096-bit local selection, be reduced to one global 256×256 mask (1,024 × u64, 8 KiB) using the existing algebra? No production code changed. Only a dev-dependency on lance-graph-mask-risc was added to the lab crate.

Decision

• The global mask exists as a demanded result. It can be built two ways, and both equal an independent set oracle on every fixture:
  • ndarray::simd::mask_scatter_or_u32, called 64 times; it does not zero its output, so the calls union;
  • one mask-risc Terminal::ScatterOrU32 program over the 262,144-row population. One program is required because the executor zeroes Out::Mask at the start of each program.
• Further Boolean algebra over G is a contract blocker through the current mask-risc terminal. ScatterOrU32's survival condition (ir.rs:329-337) forbids consuming the scattered mask, and it stays unchanged.
• One form is legal today: G AND h. Build h with Ternlog + Keep over global planes, Gather it through each map under the local mask, then scatter. G itself is never read, and parity is exact.
• Not expressible through the current terminal: G OR h, NOT G, G XOR h, and AND-ing two scattered masks. That is a statement about mask-risc, not about the operation. ndarray's scatter-or preserves its destination, so seeding the demanded output with h could yield G OR h as one terminal result. That is a hypothesis for a later contract review, not a capability.
• The demonstrated consumer needs none of these. G is the demanded output, and its zero bits already exclude the other 61,440 addresses without a complement.

Measured (ns per run, median of 31 batches; reported, not gated)

fixture	selected slots → members	ndarray ×64	mask-risc, one program
clustered 64×64	4,096 → 4,096 (64 nonzero words)	9–12 µs	15–19 µs
scattered	4,096 → 4,096 (1,008 nonzero words)	11 µs	18–19 µs
chaotic 50 %	130,990 → 56,927	157–168 µs	198–227 µs
full	262,144 → 65,536	267–411 µs	376–488 µs

Preparation is charged separately and costs more than the scatter at small selections: binding the 64 maps takes 58–108 µs, and widening u16 → u32 (a 1 MiB derived lane) takes 35–45 µs.

Controls (all asserted)

• Identity maps, collisions, same entity with different local coordinates, and same slot number with different entities.
• Stale seal, out-of-universe ordinal, version mismatch and scope mismatch: each is refused before any union.
• Disable runs fire: XOR-ing the index lane breaks scatter parity; dropping the gather gate breaks G AND h parity.
• G is membership only, never a count, a Bag, provenance or a K3 result.

Gaps named, not built

• No u16-index scatter or gather in ndarray.
• No multi-register union witness; the check that all 64 maps share one source and one scope lives only in the probe.
• The mask-risc path needs the 64 local masks laid out contiguously.

Further Boolean operators over G wait for a concrete consumer to justify them.

Details: .claude/board/entries/2026-10-26-dr256-global-union-lab.md. Unrelated to Lance I/O: P1 (#1789) failed, and G says nothing about bytes read.

🤖 Generated with Claude Code

https://claude.ai/code/session_01X6y3drwKSE2zSgoexheLFX

[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: d0999b20-be41-4b29-a927-0fe30ba6046f

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

Warning

Billing warning: we have not been able to collect payment for this subscription for more than 72 hours. Please update the payment method or pay any pending invoices in Billing to avoid service interruption.

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

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

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f7f51c1fad

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".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[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 2026-10-26T17:49:44.710992Z	f7f51c1	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.