The advanced query syntax parser here was cribbed from the original implemenation inside the OpenSearch server. mod_perl, due to being too clever by half, required a call to decode_utf8 to shove raw octets into the correct shape. Becuase the data coming to this implementation comes from opensrf, which is unicode end-to-end via JSON, it's somebody else's job to make sure the characters are properly encoded, and once they get to us they're already proper UTF-8 character strings. Thus, no need for (and things broken by) the call to decode_utf8.
Why is that explanation so long? So that we don't do it again ... :)
git-svn-id: svn://svn.open-ils.org/ILS/trunk@9142
dcc99617-32d9-48b4-a31d-
7c20da2025e4