Re: mRFC 0024: Full text indexing in Midgard
Jukka Zitting <[email protected]>
| Newsgroups | gmane.comp.web.midgard.devel |
|---|---|
| Message-ID | <[email protected]> |
Hi, Different options for linking Lucene functionality with midgard-core as described in mRFC 0024: * lucene4j (http://incubator.apache.org/lucene4c/) Not yet stable, and not too actively maintained. Based on APR. Author seems to think that using Lucene Java with gcj is a better option. * CLucene (http://clucene.sourceforge.net/) C++ library, would require C++ code in midgard-core. * Lucene Java (http://lucene.apache.org/java/) with gcj (http://gcc.gnu.org/java/) The "official" Lucene version compiled with gcj to a native library. Accessible from midgard-core through the standard JNI methods. * Remote calls to a separate indexer process Used by MidCOM indexer. After evaluating all the above options I've come to the conclusion that using Lucene Java with gcj seems like the best choise for midgard-core. We could distribute the Lucene jar along the midgard-core sources to avoid any external dependencies besides gcj. We could even use the same mechanism for the indexer process, and avoid the JRE dependency alltogether. Unless anyone objects, I will proceed with the gcj integration. BR, Jukka Zitting -- Yukatan - http://yukatan.fi/ - [email protected] Software craftmanship, JCR consulting, and Java development