From 9194d36a64b3a00e1408c340c0305b201d05331c Mon Sep 17 00:00:00 2001
From: Shay Banon <kimchy@gmail.com>
Date: Mon, 30 Jan 2012 21:23:15 +0200
Subject: [PATCH] improve compressed string construction, allow to construct it
 from utf8 bytes and use it where applicable

---
 .../cluster/metadata/AliasMetaData.java       |  3 ++-
 .../cluster/metadata/MappingMetaData.java     |  4 +++-
 .../common/compress/CompressedString.java     | 21 +++++++++++++++++++
 .../index/mapper/DocumentMapper.java          |  2 +-
 4 files changed, 27 insertions(+), 3 deletions(-)

diff --git a/src/main/java/org/elasticsearch/cluster/metadata/AliasMetaData.java b/src/main/java/org/elasticsearch/cluster/metadata/AliasMetaData.java
index f0f890b6813..1ac5f56e740 100644
--- a/src/main/java/org/elasticsearch/cluster/metadata/AliasMetaData.java
+++ b/src/main/java/org/elasticsearch/cluster/metadata/AliasMetaData.java
@@ -168,7 +168,8 @@ public class AliasMetaData {
                 return this;
             }
             try {
-                this.filter = new CompressedString(XContentFactory.jsonBuilder().map(filter).string());
+                XContentBuilder builder = XContentFactory.jsonBuilder().map(filter);
+                this.filter = new CompressedString(builder.underlyingBytes(), 0, builder.underlyingBytesLength());
                 return this;
             } catch (IOException e) {
                 throw new ElasticSearchGenerationException("Failed to build json for alias request", e);
diff --git a/src/main/java/org/elasticsearch/cluster/metadata/MappingMetaData.java b/src/main/java/org/elasticsearch/cluster/metadata/MappingMetaData.java
index 15f9aeeeeb4..77ee0555581 100644
--- a/src/main/java/org/elasticsearch/cluster/metadata/MappingMetaData.java
+++ b/src/main/java/org/elasticsearch/cluster/metadata/MappingMetaData.java
@@ -28,6 +28,7 @@ import org.elasticsearch.common.io.stream.StreamInput;
 import org.elasticsearch.common.io.stream.StreamOutput;
 import org.elasticsearch.common.joda.FormatDateTimeFormatter;
 import org.elasticsearch.common.joda.Joda;
+import org.elasticsearch.common.xcontent.XContentBuilder;
 import org.elasticsearch.common.xcontent.XContentFactory;
 import org.elasticsearch.common.xcontent.XContentHelper;
 import org.elasticsearch.common.xcontent.XContentParser;
@@ -279,7 +280,8 @@ public class MappingMetaData {
 
     public MappingMetaData(String type, Map<String, Object> mapping) throws IOException {
         this.type = type;
-        this.source = new CompressedString(XContentFactory.jsonBuilder().map(mapping).string());
+        XContentBuilder mappingBuilder = XContentFactory.jsonBuilder().map(mapping);
+        this.source = new CompressedString(mappingBuilder.underlyingBytes(), 0, mappingBuilder.underlyingBytesLength());
         Map<String, Object> withoutType = mapping;
         if (mapping.size() == 1 && mapping.containsKey(type)) {
             withoutType = (Map<String, Object>) mapping.get(type);
diff --git a/src/main/java/org/elasticsearch/common/compress/CompressedString.java b/src/main/java/org/elasticsearch/common/compress/CompressedString.java
index b6230ade2ba..0fd564cd09b 100644
--- a/src/main/java/org/elasticsearch/common/compress/CompressedString.java
+++ b/src/main/java/org/elasticsearch/common/compress/CompressedString.java
@@ -21,6 +21,7 @@ package org.elasticsearch.common.compress;
 
 import org.apache.lucene.util.UnicodeUtil;
 import org.elasticsearch.common.Unicode;
+import org.elasticsearch.common.compress.lzf.LZF;
 import org.elasticsearch.common.compress.lzf.LZFDecoder;
 import org.elasticsearch.common.compress.lzf.LZFEncoder;
 import org.elasticsearch.common.io.stream.StreamInput;
@@ -40,10 +41,30 @@ public class CompressedString implements Streamable {
     CompressedString() {
     }
 
+    /**
+     * Constructor assuming the data provided is compressed (UTF8). It uses the provided
+     * array without copying it.
+     */
     public CompressedString(byte[] compressed) {
         this.bytes = compressed;
     }
 
+    /**
+     * Constructs a new compressed string, assuming the bytes are UTF8, by copying it over.
+     *
+     * @param data   The byte array
+     * @param offset Offset into the byte array
+     * @param length The length of the data
+     * @throws IOException
+     */
+    public CompressedString(byte[] data, int offset, int length) throws IOException {
+        if (LZF.isCompressed(data, offset, length)) {
+            this.bytes = Arrays.copyOfRange(data, offset, offset + length);
+        } else {
+            this.bytes = LZFEncoder.encode(data, offset, length);
+        }
+    }
+
     public CompressedString(String str) throws IOException {
         UnicodeUtil.UTF8Result result = Unicode.unsafeFromStringAsUtf8(str);
         this.bytes = LZFEncoder.encode(result.result, result.length);
diff --git a/src/main/java/org/elasticsearch/index/mapper/DocumentMapper.java b/src/main/java/org/elasticsearch/index/mapper/DocumentMapper.java
index c06867044bc..41791d6dc78 100644
--- a/src/main/java/org/elasticsearch/index/mapper/DocumentMapper.java
+++ b/src/main/java/org/elasticsearch/index/mapper/DocumentMapper.java
@@ -573,7 +573,7 @@ public class DocumentMapper implements ToXContent {
             builder.startObject();
             toXContent(builder, ToXContent.EMPTY_PARAMS);
             builder.endObject();
-            this.mappingSource = new CompressedString(builder.string());
+            this.mappingSource = new CompressedString(builder.underlyingBytes(), 0, builder.underlyingBytesLength());
         } catch (Exception e) {
             throw new FailedToGenerateSourceMapperException(e.getMessage(), e);
         }