fix(copyright): recognize mixed POD author headings

[mstykow]

Summary

• Treat bounded POD headings that combine AUTHOR(S) with copyright, license, maintenance, or modification-history topics as author sections.
• Accept common separator and casing variants without treating copyright-only or unrelated mixed headings as author evidence.
• Keep the existing bounded section termination rules intact.

Issues

• Covers: False-positives when comparing provenant with scancode-toolkit output #1391

Scope and exclusions

• Included: POD author-section heading classification.
• Explicit exclusions: maintainer credits outside bounded author sections, handled separately.

How to verify

• Scan Perl 5.38.4 with --copyright; pod/perlopentut.pod, pod/perlretut.pod, and cpan/libnet/lib/Net/libnetFAQ.pod should recover their maintainer credits.
• Before/after corpus scans add exactly those three valid findings; Info-ZIP, musl, Dancer2, and Valgrind author output is unchanged, and copyright/holder output is unchanged across all five targets.

Intentional differences from Python

• Provenant recognizes the section structurally and keeps richer contact data where present instead of relying on ScanCode rendering.

Follow-up work

• Created or intentionally deferred: explicit collective maintainer credits outside author sections remain for the next stacked PR.

Expected-output fixture changes

• Files changed: none.
• Why the new expected output is correct: focused unit coverage verifies accepted and rejected heading shapes plus bounded extraction behavior.

[mstykow]

@greptileai review