Zhang Chao

use ByteArrayComparator replace Arrays.compareUnsigned

8

this is lucene optimize for comparisons

Use group-varint encode the positions

8

Thanks the suggestion from @jpountz , as discussed in https://github.com/apache/lucene/issues/12826 This PR use group-varint to encode some vint values if `storeOffsets` is true, it's still using class `GroupVIntReader` and `GroupVIntWriter`,...

This change will slightly improve the performance of `DataOutput#writeGroupVInts` and some other methods related to `XXXRefBuilder#append`. Here is a JMH benchmark for `Util.toIntsRef`, using java 21 on my MAC (intel...

### Description bisect shows df6bd25ce44ead0b38cffca6ef9bc3a9ce63dbea is the first bad commit. ``` org.apache.lucene.codecs.lucene90.TestLucene90FieldInfosFormat > testRandom FAILED java.lang.IllegalArgumentException: parent document and soft-deletes field can't be the same field "" at __randomizedtesting.SeedInfo.seed([E9414A90E55BE2D:7CD831A6BF35085E]:0) at...

### Description As discussed in https://github.com/apache/lucene/pull/13104, the current API allows for changing the configuration of token streams after construction(such as `ShingleFilter`), but it should be set only during construction. Perhaps...

Advance to first position of 1 in BitSet before iterating the lead in BitSetConjunctionDISI?

2

### Description In `BitSetConjunctionDISI`, we iterate over the `DocIdSetIterator` of lead, and then lookup the BitSets to get conjunction. currently, we use `bitSet.length()` to limit the termination of the iteration,...

### Description ``` org.apache.lucene.spatial3d.TestGeo3DPoint > testRandomBig FAILED java.lang.AssertionError: FAIL: id=23785 should not have matched but did shape=GeoStandardPath: {planetmodel=PlanetModel.WGS84, width=2.5638182999069305E-5(0.0014689596802307306), points={[[lat=2.4457272005608357E-47, lon=0.0([X=1.0011188539901221, Y=0.0, Z=2.4484636121979335E-47])], [lat=-1.0718146506637272, lon=0.017453291479645996([X=0.47775442364187093, Y=0.008339233987253964, Z=-0.8767777439379159])], [lat=1.2798499685240037, lon=-0.4737449978411123([X=0.2547649811064312, Y=-0.13061449414092574,...

I observed several `int[]` allocation related to `BKDPointTree`, we can use `DocIdsWriter#scratch` as the docIDs buffer for `BKDReaderDocIDSetIterator`.