fix(copyright): handle Javadoc @author with a trailing homepage anchor

[mstykow]

Summary

Follow-up to #1281, closing the remaining @author residual on eclipse-vertx/vert.x.

#1281 recovered @author <a href="url">Name</a> (name inside the anchor). vert.x also uses the form where the name is plain text and an http anchor follows it as a handle/homepage link:

* @author Francesco Guardiani <a href="https://slinkydeveloper.github.io/">@slinkydeveloper</a>

The trailing URL + @handle still broke extraction (no author). This drops the trailing http(s) anchor (href and link text) from an author-marked line when a plain-text name precedes it, leaving just the name. Scoped to an author marker; the name-inside-anchor form and copyright/contact anchors are unaffected.

Testing

• cargo test --lib copyright::detector::tests_false_positives::test_javadoc_author_with_html_anchor_is_extracted — extended with the trailing-anchor case.
• cargo test --features golden-tests --test copyright_golden — 36/36 pass.
• cargo test --lib copyright::prepare — 103 pass.

Compare evidence

eclipse-vertx/vert.x @ 78ade62: recovers Francesco Guardiani (and the other trailing-anchor @author files), closing the 4-file residual noted in #1282.

🤖 Generated with Claude Code

[greptile-apps]

Greptile Summary

This PR extends Javadoc @author extraction to handle two additional patterns: a trailing HTML handle anchor (@author Name <a href="…">@handle</a>) and a bare trailing handle (@author Name @handle), closing a residual extraction gap against eclipse-vertx/vert.x. Two complementary layers are introduced: a new static regex in prepare.rs that strips the HTML anchor before author detection runs, and a new strip_trailing_at_handle function inserted early in the refine_author pipeline that removes bare social handles post-extraction.

• prepare.rs: AUTHOR_TRAILING_HTTP_ANCHOR_RE fires before the existing AUTHOR_HTTP_ANCHOR_RE, requiring a letter immediately after the marker so it never fires on the name-inside-anchor form handled by fix(copyright): recover Javadoc @author names and drop bare XML tags #1281.
• refiner/author.rs: strip_trailing_at_handle removes a trailing @handle token when the remaining prefix is non-empty, alphabetic, and contains no @ — correctly leaving email-based author strings untouched.
• tests_false_positives.rs: Two new end-to-end cases are added to test_javadoc_author_with_html_anchor_is_extracted, covering both the HTML-wrapped handle and the bare handle patterns.

Confidence Score: 5/5

Safe to merge — the change is a narrowly scoped fix to two well-isolated extraction layers, both guarded by letter/at-sign checks that prevent misfires against the existing name-inside-anchor form.

Both new code paths carry explicit guards that prevent them from firing on the patterns handled by the previous PR. The regex ordering in prepare.rs is correct: the new trailing-anchor regex runs first and the letter guard ensures the name-inside-anchor form is unaffected. The refiner function is guarded against email-bearing or empty prefixes. End-to-end tests cover both new patterns through the full detection pipeline.

No files require special attention.

Important Files Changed

Filename	Overview
src/copyright/prepare.rs	Adds AUTHOR_TRAILING_HTTP_ANCHOR_RE to strip trailing HTML handle/homepage anchors from @author lines; applied before the existing AUTHOR_HTTP_ANCHOR_RE to avoid interference. Letter-guard on the capture group correctly prevents misfires against the name-inside-anchor form.
src/copyright/refiner/author.rs	Adds strip_trailing_at_handle as a second layer for bare @handle tokens; placed early in refine_author before collective-contributor truncation. Guards against empty prefix, email-bearing prefix (contains '@'), and non-alphabetic-only prefix are correct and consistent with the sibling function.
src/copyright/detector/tests_false_positives.rs	Extends test_javadoc_author_with_html_anchor_is_extracted with both new patterns (HTML-wrapped handle and bare handle); tests are end-to-end via detect_copyrights_from_text, covering both the prepare.rs and refiner layers.

_{Reviews (3): Last reviewed commit: "docs(copyright): correct the trailing-an..." | Re-trigger Greptile}