Apache Lucene open-source search software
SegmentTermsEnumFrame: avoid unnecessary array allocations that waste memory (#15618)
Misha Dmitriev0a699fa4d2db
diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txtindex 5b64315c514..fdfc63fd795 100644--- a/lucene/CHANGES.txt+++ b/lucene/CHANGES.txt@@ -326,6 +326,8 @@ Optimizations* GITHUB#15592: Optimize DocIdSetIteratorAcceptDocs#cost. (Shubham Chaudhary)+* GITHUB#15618: SegmentTermsEnumFrame: avoid unnecessary array allocations that waste memory (Misha Dmitriev)+Bug Fixes---------------------* GITHUB#14161: PointInSetQuery's constructor now throws IllegalArgumentExceptiondiff --git a/lucene/core/src/java/org/apache/lucene/codecs/lucene103/blocktree/SegmentTermsEnumFrame.java b/lucene/core/src/java/org/apache/lucene/codecs/lucene103/blocktree/SegmentTermsEnumFrame.javaindex 1f3fcfcd05c..b55d9396886 100644--- a/lucene/core/src/java/org/apache/lucene/codecs/lucene103/blocktree/SegmentTermsEnumFrame.java+++ b/lucene/core/src/java/org/apache/lucene/codecs/lucene103/blocktree/SegmentTermsEnumFrame.java@@ -45,13 +45,13 @@ final class SegmentTermsEnumFrame {long fpEnd;long totalSuffixBytes; // for stats- byte[] suffixBytes = new byte[128];+ byte[] suffixBytes;final ByteArrayDataInput suffixesReader = new ByteArrayDataInput();byte[] suffixLengthBytes;final ByteArrayDataInput suffixLengthsReader;- byte[] statBytes = new byte[64];+ byte[] statBytes;int statsSingletonRunLength = 0;final ByteArrayDataInput statsReader = new ByteArrayDataInput();@@ -95,7 +95,7 @@ final class SegmentTermsEnumFrame {final BlockTermState state;// metadata buffer- byte[] bytes = new byte[32];+ byte[] bytes;final ByteArrayDataInput bytesReader = new ByteArrayDataInput();private final SegmentTermsEnum ste;@@ -105,7 +105,6 @@ final class SegmentTermsEnumFrame {this.ord = ord;this.state = ste.fr.parent.postingsReader.newTermState();this.state.totalTermFreq = -1;- suffixLengthBytes = new byte[32];suffixLengthsReader = new ByteArrayDataInput();}@@ -192,7 +191,7 @@ final class SegmentTermsEnumFrame {final long codeL = ste.in.readVLong();isLeafBlock = (codeL & 0x04) != 0;final int numSuffixBytes = (int) (codeL >>> 3);- if (suffixBytes.length < numSuffixBytes) {+ if (suffixBytes == null || suffixBytes.length < numSuffixBytes) {suffixBytes = new byte[ArrayUtil.oversize(numSuffixBytes, 1)];}try {@@ -206,7 +205,7 @@ final class SegmentTermsEnumFrame {int numSuffixLengthBytes = ste.in.readVInt();allEqual = (numSuffixLengthBytes & 0x01) != 0;numSuffixLengthBytes >>>= 1;- if (suffixLengthBytes.length < numSuffixLengthBytes) {+ if (suffixLengthBytes == null || suffixLengthBytes.length < numSuffixLengthBytes) {suffixLengthBytes = new byte[ArrayUtil.oversize(numSuffixLengthBytes, 1)];}if (allEqual) {@@ -227,7 +226,7 @@ final class SegmentTermsEnumFrame {// statsint numBytes = ste.in.readVInt();- if (statBytes.length < numBytes) {+ if (statBytes == null || statBytes.length < numBytes) {statBytes = new byte[ArrayUtil.oversize(numBytes, 1)];}ste.in.readBytes(statBytes, 0, numBytes);@@ -243,7 +242,7 @@ final class SegmentTermsEnumFrame {// that's rare so won't help much// metadatanumBytes = ste.in.readVInt();- if (bytes.length < numBytes) {+ if (bytes == null || bytes.length < numBytes) {bytes = new byte[ArrayUtil.oversize(numBytes, 1)];}ste.in.readBytes(bytes, 0, numBytes);