tests: correct two inaccurate statements in surveyed scenarios

[cargo-affected-bot]

Two accuracy fixes in the functional scenarios today's nightly survey covered. Both are test-file-only; no behavior change.

clean.rs — the post-clean assertion accepted a message that path can't emit. clean_then_status_reports_no_coverage asserted contains("no coverage data yet") || contains("no coverage data found"). The second alternative is the DB-file-absent branch (status.rs:46-50), and the scenario has already asserted three lines earlier that the file still exists — clean clears via SQL DELETE, not unlink. So only the CacheMiss::NoCoverage branch (status.rs:121-125) can fire, which is exactly what the comment above the assertion already says. The || widened the assertion to cover a path the scenario structurally excludes. Dropped it; verified the single-message form still passes.

db_has_function_ranges.rs — module doc called CRATE_ROOT_SENTINEL_END pub. It is pub(crate) (coverage.rs:97), per CLAUDE.md's "Items are pub(crate), never pub". The reason the scenario hard-codes i64::MAX is unchanged and now reads as two independent reasons rather than a wrong one.

No regression test: the first change removes an over-broad assertion (a test of a test), the second is a comment. Verified locally with cargo test --bins (112 passed), cargo test --test functional clean_ and --test functional collect_writes_non_sentinel (both pass), and cargo clippy --all-targets clean.