fix(copyright): restore detector authors and cover tests in CI

[abraemer]

Summary

• fix refine_final_authors so it no longer drops valid authors that earlier detector heuristics already accepted, including maintainer lines with trailing handles and structured metadata names like gRPC authors, Meta, The libunwind project, and S2Geometry
• add focused regression tests at the postprocess layer and restore the hidden copyright::detector unit-test target to green
• fix the CI coverage gap: the job named Rust Library Tests was only running cargo test --bin provenant, so detector library tests were never exercised; this PR renames that job/step accurately and adds a dedicated cargo test -p provenant-cli --lib --release copyright::detector:: step

Issues

• Covers: hidden copyright detector regressions in maintainer/metadata author extraction and the CI gap that let those library tests break without detection

Scope and exclusions

• Included: the minimal author postprocess fix, focused regression tests, targeted CI coverage for copyright::detector, and clearer unit-test job naming
• Explicit exclusions: broadening CI to the full cargo test -p provenant-cli --lib --release target. Local validation surfaced unrelated pre-existing non-detector lib-target failures, so this PR fixes the missed copyright coverage without expanding into those separate failures.

Intentional differences from Python

• None. This restores intended detector behavior by preserving valid author detections that the Rust heuristics had already accepted upstream.

Follow-up work

• Created or intentionally deferred: investigate the unrelated failing non-detector --lib tests separately, then consider broadening CI to full library-target coverage once that suite is green

Expected-output fixture changes

• Files changed: none
• Why the new expected output is correct: this change only fixes author retention in detector unit-tested code paths and CI coverage; it does not intentionally change golden or expected output fixtures