Add `DiffCleanupSemantic` into the mix when generated diffs (PR review,
commit view and issue/comment history). This avoids trying to produce a
optimal diff and tries to reduce the amount of edits, by combing them
into larger edits, which is nicer and easier to 'look at'. There's no
need for a perfect minimal diff, as the output isn't being parsed by a
computer, it's parsed by people.
Ref: https://codeberg.org/forgejo/forgejo/issues/4996
Change all license headers to comply with REUSE specification.
Fix #16132
Co-authored-by: flynnnnnnnnnn <flynnnnnnnnnn@github>
Co-authored-by: John Olheiser <john.olheiser@gmail.com>
Use Unicode placeholders to replace HTML tags and HTML entities first, then do diff, then recover the HTML tags and HTML entities. Now the code diff with highlight has stable behavior, and won't emit broken tags.
