Hi Michael, I don't think that there is a convention about the packaging or structure of contrib. For the ports, a java-like directory hierarchy can be used (like Contrib\Queries\Search\src & Contrib\Queries\Search\test for BooleanFilter & TermFilter)
For others, something like Contrib/Search/SegmentCache Contrib/Index/Enumerators/FieldEnumerators & TermVectorEnumerator etc. I would also like to see all the enumerators in the core (although George wouldn't like that idea :). DIGY -----Original Message----- From: Michael Garski [mailto:mgar...@myspace-inc.com] Sent: Wednesday, February 17, 2010 4:49 AM To: lucene-net-dev@lucene.apache.org Subject: Contrib Section I have a few items nearly ready for the contrib section. Here's a list of what I have that I can contribute at this time: 1) FieldEnumerators : wrappers around TermEnum/TermDocs to provide foreach syntax support, including those for numeric fields. 2) TermVectorEnumerator : gives foreach syntax to allow for enumerating over all of the term vectors in an index. 3) BooleanFilter : combine multiple filters in various logical combinations (port of Java contrib.) 4) TermsFilter : creates a filter from several terms as a logical OR; note: makes use of the FieldEnumerators when setting the doc set bits (port of Java contrib) 5) SegmentCache : a generic cache that works with the segment-level searching in Lucene 2.9 to cache any data at the segment level in the same manner as the field cache and CachingWrapperFilter. I've used it for caching aggregate data on the documents in a segment for use in diagnostics and for caching TermVectors. Is there an accepted approach as to how to package these up other than adding the standard Apache header? They are currently in one project with another project containing Visual Studio tests (I can convert to NUnit if necessary). I need to beef up the documentation comments on these items a bit as well to cover usage and include a reference to the original item if ported from Java. Michael
