+1. In addition we should also include common set of test case which can be used to validate the SPI implementations. Also we can leave oak-lucene as is for now and just create new module and implement oak-lucene-v2 based on that. Once it reaches feature parity we can remove oak-lucene bundle Chetan Mehrotra
On Wed, Apr 4, 2018 at 5:43 PM, Thomas Mueller <muel...@adobe.com.invalid> wrote: > +1 > > On 04.04.18, 10:23, "Tommaso Teofili" <tommaso.teof...@gmail.com> wrote: > > Hi all, > > In the context of creating an (abstract) implementation for Oak full text > indexes [1], I'd like to create a new module called _oak-search_. > Such module will contain: > - implementation agnostic utilities for full text search (e.g. aggregation > utilities) > - implementation agnostic SPIs to be extended by implementors (currently > we > expose SPIs in oak-lucene whose signatures include Lucene specific APIs) > - abstract full text editor / query index implementations > - text extraction utilities > > Please share your feedback / opinions / concerns. > > Regards, > Tommaso > > [1] : https://issues.apache.org/jira/browse/OAK-3336 > >