Tomcat needs an explicit parameter somewhere to use UTF-8 text. It's on the wiki how to do this.
On Fri, Apr 6, 2012 at 4:41 PM, N. Tucker <ntucker-ml-solr-us...@august20th.com> wrote: > Further info: I can make this work if I stay out of tomcat -- I > download a fresh solr binary distro, copy those five JARs from 'dist' > and 'contrib' into example/solr/lib/, copy my solrconfig.xml and > schema.xml, and run 'java -jar start.jar', and it works fine. But > trying to add those same JARs to my tomcat instance's solrhome/lib > doesn't work. Any ideas how to troubleshoot? > > > On Fri, Apr 6, 2012 at 12:15 PM, N. Tucker > <ntucker-ml-solr-us...@august20th.com> wrote: >> Hello, I'm running into an odd problem trying to use ICUTokenizer >> under a solr installation running under tomcat on ubuntu. It seems >> that all the appropriate JAR files are loaded: >> >> INFO: Adding 'file:/usr/share/solr/lib/lucene-stempel-3.5.0.jar' to >> classloader >> INFO: Adding 'file:/usr/share/solr/lib/lucene-smartcn-3.5.0.jar' to >> classloader >> INFO: Adding 'file:/usr/share/solr/lib/icu4j-4_8_1_1.jar' to classloader >> INFO: Adding 'file:/usr/share/solr/lib/lucene-icu-3.5.0.jar' to classloader >> INFO: Adding 'file:/usr/share/solr/lib/apache-solr-analysis-extras-3.5.0.jar' >> to classloader >> ... but later: ... >> SEVERE: java.lang.NoClassDefFoundError: >> org/apache/lucene/analysis/icu/segmentation/ICUTokenizer >> >> I'm not too clear on the correct way to add the contrib bits other >> than copying them into the 'lib' directory under solrhome. They are >> obviously found there (and I have verified that ICUTokenizer is in >> lucene-icu-3.5.0.jar), but there's still a problem loading the >> ICUTokenizer class. Any tips on troubleshooting this? Are there more >> depenencies that I'm unaware of? -- Lance Norskog goks...@gmail.com
