I know we removed string interning in trunk, but it might come as a surprise to users trying out trunk on an existing index that this codec is still interning their field names.
Does Lucene3x codec really need to intern field names? Does it rely upon this and is this actually even working correctly today (especially since the interning does not happen in Term etc)? Maybe it only appears to be working since our tests always use string constants for field names... can we change LuceneTestCase.newField() to sometimes take the input string and use a non-interned equivalent? and what about at query time? -- lucidimagination.com --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
