Apache Lucene open-source search software
Move CHANGES.txt entry to 10.5.0
Shubham Chaudhary4dacd6cf9a67
diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txtindex 34b393064cf..a8f217a1870 100644--- a/lucene/CHANGES.txt+++ b/lucene/CHANGES.txt@@ -158,13 +158,6 @@ OptimizationsBug Fixes----------------------* GITHUB#15901: Fix undercounting of RAM used by vectors buffered in in-memory segments.- BufferingKnnVectorsWriter hardcoded Float.BYTES for all vector encodings, overcounting byte- vectors by 4x. Scalar quantized vector writers (Lucene104, Lucene99, Lucene102) did not- account for rawVectorDelegate RAM, making byte vector fields completely invisible in RAM- reporting. Also added missing dimensionSums array accounting in quantized writers.- (Prithvi S)-* GITHUB#14049: Randomize KNN codec params in RandomCodec. Fixes scalar quantization div-by-zerowhen all values are identical. (Mike Sokolov)@@ -517,6 +510,12 @@ Bug Fixes* GITHUB#16046: Fix double-counting of underlying Automaton in CompiledAutomaton#ramBytesUsed. (Eugene Rizhkov)+* GITHUB#15901: Fix undercounting of RAM used by vectors buffered in in-memory segments.+ BufferingKnnVectorsWriter hardcoded Float.BYTES for all vector encodings, overcounting byte+ vectors by 4x. Scalar quantized vector writers (Lucene104, Lucene99, Lucene102) did not+ account for rawVectorDelegate RAM, making byte vector fields completely invisible in RAM+ reporting. Also added missing dimensionSums array accounting in quantized writers. (Prithvi S)+Other---------------------* GITHUB#15586: Document that scoring and ranking may change across major Lucene versions, and that applications requiring stable ranking should explicitly configure Similarity. (Parveen Saini)