From 49631091d60c2bca111cebe9ca501d951df585eb Mon Sep 17 00:00:00 2001
From: Michael McCandless <mikemccand@apache.org>
Date: Tue, 4 May 2010 18:38:18 +0000
Subject: [PATCH] LUCENE-2422: don't reuse byte[]/char[] for String IO in
 IndexInput/Output

git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/trunk@940994 13f79535-47bb-0310-9956-ffa450edef68
---
 lucene/CHANGES.txt                                  |  4 ++++
 .../src/java/org/apache/lucene/store/DataInput.java | 13 ++-----------
 .../java/org/apache/lucene/store/DataOutput.java    |  3 +--
 3 files changed, 7 insertions(+), 13 deletions(-)

diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
index c00c5e8bb68..70878968b52 100644
--- a/lucene/CHANGES.txt
+++ b/lucene/CHANGES.txt
@@ -365,6 +365,10 @@ Bug fixes
 * LUCENE-2074: Reduce buffer size of lexer back to default on reset.
   (Ruben Laguna, Shai Erera via Uwe Schindler)
   
+* LUCENE-2422: Don't reuse byte[] in IndexInput/Output -- it gains
+  little performance, and ties up possibly large amounts of memory for
+  apps that index large docs.  (Ross Woolf via Mike McCandless)
+
 New features
 
 * LUCENE-2128: Parallelized fetching document frequencies during weight
diff --git a/lucene/src/java/org/apache/lucene/store/DataInput.java b/lucene/src/java/org/apache/lucene/store/DataInput.java
index e175be6ae71..e779bac01a3 100644
--- a/lucene/src/java/org/apache/lucene/store/DataInput.java
+++ b/lucene/src/java/org/apache/lucene/store/DataInput.java
@@ -29,8 +29,6 @@ import org.apache.lucene.util.RamUsageEstimator;
  * data types.
  */
 public abstract class DataInput implements Cloneable {
-  private byte[] bytes;                           // used by readString()
-  private char[] chars;                           // used by readModifiedUTF8String()
   private boolean preUTF8Strings;                 // true if we are reading old (modified UTF8) string format
 
   /** Reads and returns a single byte.
@@ -131,18 +129,14 @@ public abstract class DataInput implements Cloneable {
     if (preUTF8Strings)
       return readModifiedUTF8String();
     int length = readVInt();
-    if (bytes == null || length > bytes.length) {
-      bytes = new byte[ArrayUtil.oversize(length, 1)];
-    }
+    final byte[] bytes = new byte[length];
     readBytes(bytes, 0, length);
     return new String(bytes, 0, length, "UTF-8");
   }
 
   private String readModifiedUTF8String() throws IOException {
     int length = readVInt();
-    if (chars == null || length > chars.length) {
-      chars = new char[ArrayUtil.oversize(length, RamUsageEstimator.NUM_BYTES_CHAR)];
-    }
+    final char[] chars = new char[length];
     readChars(chars, 0, length);
     return new String(chars, 0, length);
   }
@@ -219,9 +213,6 @@ public abstract class DataInput implements Cloneable {
       clone = (DataInput)super.clone();
     } catch (CloneNotSupportedException e) {}
 
-    clone.bytes = null;
-    clone.chars = null;
-
     return clone;
   }
 
diff --git a/lucene/src/java/org/apache/lucene/store/DataOutput.java b/lucene/src/java/org/apache/lucene/store/DataOutput.java
index bb463009ecc..e44d9cb33e6 100644
--- a/lucene/src/java/org/apache/lucene/store/DataOutput.java
+++ b/lucene/src/java/org/apache/lucene/store/DataOutput.java
@@ -29,8 +29,6 @@ import org.apache.lucene.util.UnicodeUtil;
  */
 public abstract class DataOutput {
 
-  private BytesRef utf8Result = new BytesRef(10);
-
   /** Writes a single byte.
    * @see IndexInput#readByte()
    */
@@ -101,6 +99,7 @@ public abstract class DataOutput {
    * @see DataInput#readString()
    */
   public void writeString(String s) throws IOException {
+    final BytesRef utf8Result = new BytesRef(10);
     UnicodeUtil.UTF16toUTF8(s, 0, s.length(), utf8Result);
     writeVInt(utf8Result.length);
     writeBytes(utf8Result.bytes, 0, utf8Result.length);