Why: because its faster by multiple orders of magnitude.
Stemming is supported. Stemming language can be set per feed and in the search dialog for specific query (this is the best I could figure out so far given the limitations I'm dealing with).
You don't need to change default language to search for w/e unicode stuff, it's only for word stemming.
Also, you will need to reindex your old stuff. This unfortunately is a pretty slow process. Stemming is currently hardcoded to English when importing old data.
Code: Select all