[AdaWorldAPI]

Codegraph::impact returns an empty set for a public symbol whose callers live in crates the store does not hold (#2000). This adds Codegraph::impact_report, which keeps impact's masks exactly and adds a typed list of what the answer could not see. It then answers three questions about report-rs PR #33 from a Charon store, graded blind against the manual EXO-185 findings. Board: EXO-229, EXO-230.

Changes

• crates/lance-graph-codegraph/src/impact_report.rs (new, additive): Scope, Targets, SymbolState, Gap, ImpactReport, Codegraph::impact_report. impact is unchanged.
  • Gaps: external consumers unobserved, targets not extracted or undeclared, excluded regions (opaque modules, features off), unresolved dispatch, incomplete harvest, stale or unrecorded files, and .rs files under a read directory that files.tsv does not list.
  • The external-consumer gap fires when the symbol, or anything that reaches it, is public. When the consumer set is declared closed, every stored crate's coverage is checked instead, including a crate whose call into the symbol sits in code it did not extract.
  • Missing coverage is never written into Graded.unknown.
  • A missing symbol in a stored crate carries its crate's target, region, harvest and file gaps, because each is a place it may still be defined.
  • closed_in_scope() is local; Coverage::no_known_gap() alone is never treated as closure.
• examples/cg_query.rs: a report subcommand with --targets, --exclude, --consumers-closed.
• tests/impact_coverage.rs: 14 tests.
• .claude/probes/2026-11-11-codegraph-recon-2/: report, questions, held-out answer key, both arms' answers, grades, raw store output.
• .codegraph/ regenerated; board entry.

Findings (details in REPORT.md)

• report-rs store (81705d9, lib, default features, crate::app opaque because the full crate ran out of memory): the CRATE row says COMPLETE, yet 14 of 59 src/ files were never read (iam.rs, main.rs, project_plan/*, …). SourceFilesUnread reports them from files.tsv.
• feat(bgz17): RaBitQ encoding, PCDVQ weighted distance, sigma-band palette D1: rabitq_compat.rs — RaBitQEncoding struct with: - OrthogonalMatrix (Hadamard rotation) - encode(): normalize → rotate → sign-quantize → binary + corrections - distance_rabitq(): unbiased L2² estimate with correction factors - distance_palette(): O(1) matrix lookup - distance_corrected(): generative decompression with LFD correction D6: Base17::l1_weighted() — PCDVQ direction-weighted L1 distance. dim 0 (sign) weighted 20×, dims 1-6 (exponent) 3×, dims 7-16 (mantissa) 1×. DistanceMatrix::build_pcdvq_weighted() uses weighted distances. D7: Palette::from_sigma_bands() — training-free codebook from empirical distribution. Equal-percentile bands = GQ-style guaranteed uniform utilization. No empty clusters by construction. 114 tests pass (was 108), 6 new tests for the added features. https://claude.ai/code/session_01CdqyUTUfjKZuk8YGJzv6LB #33 finding recovered: workbook_delta::WorkbookSource::poll has no library caller and no path to Ledger::ingest; Ledger::pull → source_fake::pull is dynamic and reported UNKNOWN, not as an edge. Each absence carries its scope.
• Blind grading (12 key items, strict): the source arm got 12 correct; the store arm got 7 correct, 3 partial and 2 unsupported. Neither arm made a wrong statement. The store's misses are build configuration (the feature-gated iam call), targets (test callers) and manifest facts (Cargo.toml). n = 1 per arm.
• D-CODEGRAPH-RECON-1: Codegraph over real Charon stores, Mermaid reconstruction, review evidence packet #2000 reproduces from its documented commands: rows per crate equal, machine-recovered.mmd byte-identical, Charon recall column identical. The rust-analyzer column was not re-run.

Checks

• In crates/lance-graph-codegraph: cargo fmt -- --check, cargo clippy --all-targets -- -D warnings, cargo test.
• Every new guard was disabled and its test went red (both directions for unread files).
• python3 .claude/tools/codegraph_index.py --check: current.

Open

• Stores per configuration (features, test targets) would close the two unsupported items.
• Targets and opaque modules are declared by the caller; the extractor could record them in the store.

🤖 Generated with Claude Code

https://claude.ai/code/session_01XAqKMLLhjfT6wad5BaeNBS

[coderabbitai]

Pre-merge checks: 1 · 1

Failed checks: 1 inconclusive

Check name	Status	Explanation	Resolution
Title check		This assessment has not completed. The check will be retried on the next pre-merge check run.

Passed checks: 1 passed

Check name	Status	Explanation
Description Check		Check skipped - CodeRabbit’s high-level summary is enabled.

• 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: ✅ Completed

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

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-11T07:23:02.573356Z	b0364e4	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.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b0364e4ecd

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".