From: Germán Biozzoli <germanbiozzoli@gm...> - 2010-12-27 19:26:11
I'm trying the 1.70 version of Dspace integrated with Solr.
I've converted the original 1.62 database and created a new dspace170
db with UNICODE encoding.
Considering that my records are accented, I've modified the search
solr's schema including ISOLatin1Filter in every text field, and it's
seems to be OK, I mean that you can search "educacion" and the records
that have "educacion" or "educación" are found. So far so good.
My problem appears trying the inverse, I mean searching "educación",
the resultset is empty. Looking at the tomcat log I can see
INFO: [search] webapp=/solr path=/select
hits=0 status=0 QTime=23
It appears that the search term is not arriving correctly encoded to
Solr, it receives educaciÃ³n.
Looking at my tomcat configuration, I' ve defined
<Connector port="8009" protocol="AJP/1.3" redirectPort="8443" maxThreads="150"
I should suppose then, that there is some kind of communication
encoding problem between the Dspace's form and Solr? Anybody could
point me to the code that I should see to fix it? I'm using the mirage
theme, don't know if this could be a problem (perhaps some form
encoding?). The Dspace page appears to be correctly encoded in UTF-8
and the accented words are displayed OK.
Thanks a lot for the help