Summary

This PR introduces a new extension point, processEqualities(), inside DiffRowGenerator to address the HTML-escaping issue described in #219.

Previously, equal (unchanged) segments were passed through HTML normalization only once.
Because < and > were normalized to &amp;lt; and &amp;gt;, equalities that began with & caused incorrect inline diff grouping (e.g., &amp;lt; vs &amp;gt;). This produced invalid HTML when inlineDiffByWord(true) was used.

What This PR Changes

✔ Adds a new protected hook:

protected String processEqualities(String text)

✔ All equality chunks now pass through this hook before becoming DiffRow(Tag.EQUAL, ...).
✔ Default behavior is no modification (fully backward compatible).
✔ Users can override this hook to customize HTML or text transformations after diffing, solving the core issue.

Why this Fix Works

The problem in #219 occurs because the diff logic compares normalized text, but the final output does not normalize equal segments the same way as diffs.

By exposing processEqualities(), we give users a consistent post-diff transformation path.
This ensures that equal text is handled symmetrically to diffed text and prevents HTML-breaking cases like &amp;lt; / &amp;gt; being partially diffed.

Backward Compatibility

• Existing behavior is preserved.
• No breaking API changes.
• The feature is opt-in — users override it only if needed.

Tests Included

✔ Added DiffRowGeneratorEqualitiesTest
✔ Verifies that:

• Equalities now route through processEqualities()
• Inline diffs still work correctly
• The issue Line normalization before diff #219 reproduction scenario no longer produces broken HTML

Fixes

Fixes #219