My *real* suggestion would be to not do it. Write a SolrJ program that uses whatever version of Tika you want to download and use *that* to index rather than try to sort through the various jar dependencies in Solr. It'd be safer.
Otherwise, you're on your own here. Here's some example code: http://www.lucidimagination.com/blog/2012/02/14/indexing-with-solrj/ Best Erick On Sun, Feb 26, 2012 at 9:01 PM, bing <jsuser1...@hotmail.com> wrote: > Hi, Erick, > > My idea is to use Tika0.10 in Dspace1.7.2, which is based on two steps: > > 1. Upgrade Solr1.4.1 to Solr3.3.0 in Dspace1.7.2 > In the following link, upgraded Solr & Lucene 3.3.0 has been resolved. > https://jira.duraspace.org/browse/DS-980 > > 2. Upgrade to Tika0.10 in Solr3.3.0 > In the following link, people has tried to upgrade Tika0.8 to Tika0.9. > http://lucene.472066.n3.nabble.com/upgrading-to-Tika-0-9-on-Solr-1-4-1-td2570526.html > > I was thinking, if both the above two steps can be achieved, then maybe I > can get it done. What is your suggestion? > > Thank you. > > Best Regards, > Bing > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/TikaLanguageIdentifierUpdateProcessorFactory-since-Solr3-5-0-to-be-used-in-Solr3-3-0-tp3771620p3779437.html > Sent from the Solr - User mailing list archive at Nabble.com.
