Move the compatibility tool's inline tests out of its pinned lib.rs

[lamemustafa]

Follows docs/proposed-rust-module-conventions.md item 2, as in #440–#442, for the tests module of tools/bridge-tally-compatibility/src/lib.rs, the tool that validates the compatibility surface. It moves to lib_tests.rs beside it via #[path].

file	before	after
lib.rs	2,853	1,686
lib_tests.rs	—	1,162

No production code followed the test block.

Checked, not asserted

• No production change. Before rustfmt, lib.rs rebuilt byte-identical to HEAD from the new parent plus the re-indented child. After rustfmt, the child differs from that reconstruction only in layout (below).
• rustfmt made one non-whitespace change, a trailing comma. All 169 string and char literals are identical.
• No test lost. The test lists are identical before and after: 24 lib tests by default, and 30 with bills-native-outstandings-probe-receipt. The full crate passes under both, with no warnings.
• The one test that needed resealing. Before the reseal, real_tree_has_complete_migration_and_report_surface_coverage failed with surface_file_changed, which is expected after changing a pinned file. It passes after the reseal. reseal --verify exits 0, and exactly one pinned hash changed.
• Clippy -D warnings is clean for CI's tools-workspace step and for its probe-receipt feature step.
• Scripts that read these files as text:
  • check-tally-live-read-boundary.mjs allowlists this lib.rs for probe identifiers. The moved tests contain none, and the check passes.
  • The request-builder hazard scanner now reads lib_tests.rs as production code (see Skip extracted test modules in the hazard scanner by how they load, not by name #443). It still reports exactly its 10 pinned violations.
  • The other scanning scripts pass.
  • The frontend suite passes on a clean tree: 239 node, 113 vitest, 10 Playwright. (The merge-driver test refuses to run while surface files have uncommitted changes.)
• Windows. The one #[cfg(windows)] test couldn't run on the macOS host that did this move. CI's windows-latest native job runs the tools workspace tests.

What leaves the seal, and why this file is different

lib_tests.rs is unpinned. That follows #416's rule and #440–#442: tests decide nothing about what Bridge posts or emits.

The loss is sharper here than in earlier extractions. real_tree_has_complete_migration_and_report_surface_coverage holds the only executable check of the reserve bound:

assert!(MAX_SURFACE_FILES - surface.files.len() <= RESERVED_SURFACE_FILES);

The cap itself is still enforced in the pinned lib.rs. But editing or deleting that assertion no longer changes a pinned hash. If you want it sealed, pinning lib_tests.rs takes one entry, a cap raise with a named reason, and matching changes to the guard.

Correction to the remaining-work picture

An earlier version of this description called this the last large inline test block in a pinned file. It isn't. I had ranked files by production lines and truncated the list. Ranked by inline test lines, 70 pinned files still hold 18,384 lines of top-level inline #[cfg(test)] modules. This one is third, behind source_draft/catalog.rs (1,779) and tally/connector.rs (1,317).

Coordination

#444 also edits this lib.rs, raising the cap to 251 in the doc comment and constant near the top. The two edits don't overlap in the source. Both regenerate the surface manifest, and #444 adds pins, so whichever lands second must merge the pin list and reseal, not take either side.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.