note
salva
That is because <c>\b{wb}</c> matches between those signs.
<p>This seems to solve the issue:
<c>
my @fragments = grep length, split /(\b{wb}\w.*?\b{wb}|\n+)/, $book;
</c>
<p>But my knowledge of Unicode and the <c>\b{wb}</c> semantics is rather limited so that may have other issues.
11141603
11141614