Hi, Lewis. Given your exposition, I agree with you. Just comment that I don't agree with patch GORA-206 because it is only a patch for -obviously- GORA-206 and for me does not seems to be reusable in GORA-223, but as told, your exposition is completely consistent and reasoned.
+1 Regards, Alfonso Nishikawa 2013/6/6 Lewis John Mcgibbney <[email protected]> > Hi, > My justification for the above is simple. We addressed GORA-174, which was > to handle ["string", "null"] unions in Avro schemas. > AFAICT, full support for Alfonso's additional functionality e.g. > > - Support of ["null",type] (a.k.a. optional field). > - Support for mutitypes(3+) unions. > - Support of nested unions. > - Support of recursive optional records. > - Support of unions as value in maps and arrays. > - Serialization of topmost optional fields of the main record in "raw": > topmost ["null","type"] (optional field) will be persisted like if it was > ["type"] (and non-existant column === null). This ensures data form 0.2.1 > can be read. > > Has now been implemented in all modules bar gora-cassandra, hence the > opening of GORA-223 [0]. > > I do not think it is a productive way for us to spend time forward and > back-porting patches between codebases. I've not had the best of > experiences with this in the past. > > I propose to skip four failing tests in gora-cassandra e.g. > testGetRecursive, testGetDoubleRecursive, testGetNested and finally, > testGet3UnionField. > > The hope is that we can regroup the effort to focus on smaller more > manageable tasks for the 0.4 development drive. > > Any thoughts folks? > > One last thing. Great work on this one to everyone. I contributed very > little having spent a good bit of time looking over the various commits > learnt a lot. > > Lewis > > [0] https://issues.apache.org/jira/browse/GORA-223 > > > > > > > > > > -- > *Lewis* > -- "Drinking bloody marys all night will make you feel like a corpse in the morning."
