Share prepared analysis ownership for queries

[jonathanong]

What changed

Routes data-pw, registry-extension, and fetches through one request-owned AnalysisSession.

Why

These public paths previously created independent inventories, source reads, or parse passes. The request boundary now owns visible files, SourceStore, parsing, and the query-owned registry report projection.

Impact

• data-pw consumes the session's visible inventory and sources
• fetch matching and traversal share one session-backed source/parse path
• registry-extension memoizes success and parse failure per normalized root/file key
• independent registry paths use a DashMap; only same-key builds coordinate through OnceLock

Validation

• registry repeat-success and repeat-failure regressions: 2 passed
• prepared fetch-session regression: 1 passed
• compatibility-wrapper static-fixture regression: 1 passed
• cargo clippy -p no-mistakes --all-targets --all-features -- -D warnings
• cargo fmt --check
• git diff --check
• GitHub Test CI run 31410110320: all tests, Rust coverage, Codecov, benchmarks, lint/package, AST analysis, and JavaScript checks passed

Shepherd Journal

• A resolved automated suggestion proposed wrapping the registry regression in an external request parse-cache scope. That would hide the production ownership gap: public request-scoped AnalysisSession must itself memoize repeated query work. The per-key request-owned report projection fixes the behavior while preserving the unconditional regression.

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@crates/no-mistakes/src/fetches/tests/run_with_base_root_tests.rs`:
- Around line 59-84: Strengthen
prepared_fetch_run_reuses_the_session_for_target_matching_and_traversal by
asserting the session work snapshot contains the fixture’s expected two
normalized source-read and parse-attempt entries before checking each count is
1. Also assert that the report includes the expected /api/users fetch result,
ensuring the test detects bypassed session reads, parsing, or traversal
regressions.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 5760397c-a877-42de-bf9c-fd4a96d260e3

📥 Commits

Reviewing files that changed from the base of the PR and between 287453d and a51b22c.

📒 Files selected for processing (14)

• crates/no-mistakes/src/codebase/ts_source/tests/gitignore.rs
• crates/no-mistakes/src/data_pw_query.rs
• crates/no-mistakes/src/data_pw_query/scan.rs
• crates/no-mistakes/src/data_pw_query/tests.rs
• crates/no-mistakes/src/fetch/file_analysis.rs
• crates/no-mistakes/src/fetch/file_facts.rs
• crates/no-mistakes/src/fetch/import_routes.rs
• crates/no-mistakes/src/fetch/mod.rs
• crates/no-mistakes/src/fetch/route_analysis.rs
• crates/no-mistakes/src/fetches/pipeline/route_analysis.rs
• crates/no-mistakes/src/fetches/pipeline/run.rs
• crates/no-mistakes/src/fetches/tests/run_with_base_root_tests.rs
• crates/no-mistakes/src/registry_extension_query.rs
• crates/no-mistakes/src/registry_extension_query/tests.rs

Memoized registry-extension reports in the request session and strengthened the fetch session regression assertions.