As implemented, the trim filter does not update offsets when it trims a
token. Is this intentional, or just has not been important to anyone?
I'm trying to build links from an input text using the TokenStream. For
example, i have:
subject = " aaa -- bbb (ccc)";
With the current impl, I get:
<a href="/get/subject:aaa/"> aaa </a>--<a href="/get/subject:bbb/"> bbb
</a>(<a href="/get/subject:ccc/">ccc</a>)
I would like to get:
<a href="/get/subject:aaa/">aaa</a> -- <a href="/get/subject:bbb/">
bbb</a> (<a href="/get/subject:ccc/">ccc</a>)
ryan
