[LUCENE-7042] more cleanup for Point encodings

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 6.0
• Lucene Fields: New
• Assignee: Unassigned
• Reporter: Robert Muir
• Created: 22/Feb/16 22:31
• Updated: 28/Aug/22 14:50
• Resolved: 23/Feb/16 00:33

Description

Followup of LUCENE-7039.

I started consolidating encode/decode functions into XYZField, but its still pretty messy. Specifically, duplicated code because indexer wants packed byte[] and queries want byte[][] and so on.

We can do a bit more cleanup: by defining everything in terms of encodeDimension() and decodeDimension(). This way, the logic is only in one place.

Also the new NumericUtils methods here are confusing I think. Its scary how they take a "dimensional index" when parsing from a byte[]. I think they should just take 'offset' and not try to be fancy, its easier to understand.