[thekevinbot]

Step 2 of 2 in the packages/python/dirsql lane of the #1020 ratchet endgame. Stacked on #1077 (claude/1066-p1-resolve-extension) — the new has_bare_name module imports is_bare_name, which #1077 moves to its own file. GitHub retargets this to main when #1077 merges.

resolve_config_extensions.py held four module-scope functions and the gate flagged three at max_lines = 1. The holder is a free choice, so the module's named export resolve_config_extension_specs keeps the file and the helpers move out:

Moved	To	Carried with it
_load_toml_module	dirsql/toml_module.py	the _toml binding it produces (its only reader is _load_extension_entries)
_load_extension_entries	dirsql/load_extension_entries.py	—
_has_bare_name	dirsql/has_bare_name.py	—

3 flagged, 3 moves plus the _toml binding riding along — no extra modules needed.

resolve_configs_extension_specs.py imports _has_bare_name and _load_extension_entries; it now takes each from its own module rather than reaching through the singular resolver, matching how it already imports _resolve_entries. Its colocated test gains three __module__ assertions pinning where each helper is bound from, so the co-change is a real assertion rather than a whitespace edit.

With this the whole packages/python/dirsql tree reports zero violations at max_lines = 1. Verified by temporarily setting packages/python/testing-conventions.toml to 1 and re-running the gate; the config change is not in this diff (per #1020, the committed ceiling stays at 12 until the final PR deletes the key).

Red/green does not apply: pure moves, no behavior change.

PARITY.md needs no change. None of these functions is in dirsql.__all__. PARITY.md names resolve_config_extension_specs and resolve_configs_extension_specs; both keep their names and their modules. No parity drift — file layout inside one SDK.

Gates

just preflight: 0 failing pairs, 3 skipped (the three packaging pairs, which need a built artifact).

• unit mutation, diff-scoped against the stack parent (--base claude/1066-p1-resolve-extension): 20 mutants tested, 0 survivors
• unit coverage: 100% line + branch across packages/python/dirsql (307 stmts, 90 branches, 0 miss)
• unit one-function-per-file at max_lines = 1: the whole tree clean
• unit lint, colocated-test (incl. co-change), integration lint, declared-deps: clean
• Unit suite 134 passed; hermetic integration 26 passed

Two tests were added rather than relocated, both to keep mutation at zero on newly-moved lines: an OSError-on-open case for the unreadable-config arm, and a traversal-segment case pinning that the base directory is normalised.

Manual run (uv run python -c ... against the package, with a .dirsql.toml naming sqlite_vec):

bare name reached the resolver: could not resolve extension package 'sqlite_vec': not installed
missing config -> None
plural, empty -> None

E2E Verification

• Ran e2e suites locally for every affected SDK
• Python SDK e2e: pass
• TypeScript SDK e2e: N/A
• Rust core e2e (if applicable): N/A
• packages/python/e2e-attestations/<branch>.json receipt written if packages/python changed (just e2e-attest-python)
• packages/ts/e2e-attestations/<branch>.json receipt written if packages/ts changed (just e2e-attest-ts)
• Command(s) run: cargo build -p dirsql --features cli, just e2e-attest-python
• Result summary: 37 passed in 15.31s. Receipt at packages/python/e2e-attestations/claude-1079-p2-resolve-config-extensions.json; git show --stat on the attest commit confirms it adds one file and deletes no sibling receipts (e2e attest silently deletes other branches' attestation receipts #1041).

Changelog / Migrations

• Changelog fragment added under <root>/<pkg>/changelog.d/ for each changed package (or: skip-changelog trailer on a commit with reason)
• Migration fragment added under <root>/<pkg>/migrations.d/ (or: not required -- additive/bugfix only)

packages/python/changelog.d/2026-09-05-config-extension-loader-split.md. No migration fragment: pure refactor, no public API or runtime behavior change.

Fixes #1079