You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@lucene.apache.org by si...@apache.org on 2012/01/19 15:06:10 UTC

svn commit: r1233380 - in /lucene/dev/trunk/lucene/src: java/org/apache/lucene/codecs/lucene3x/ java/org/apache/lucene/codecs/lucene40/ test-framework/java/org/apache/lucene/codecs/preflexrw/

Author: simonw
Date: Thu Jan 19 14:06:10 2012
New Revision: 1233380

URL: http://svn.apache.org/viewvc?rev=1233380&view=rev
Log:
LUCENE-3707: Add a Lucene3x private SegmentInfosFormat implementation

Added:
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosFormat.java
      - copied, changed from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosReader.java
      - copied, changed from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java
    lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosFormat.java
      - copied, changed from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java
    lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosWriter.java
      - copied, changed from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java
Modified:
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xCodec.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xFieldInfosFormat.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xNormsFormat.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xPostingsFormat.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xTermVectorsFormat.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java
    lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java
    lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWCodec.java

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xCodec.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xCodec.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xCodec.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xCodec.java Thu Jan 19 14:06:10 2012
@@ -30,9 +30,7 @@ import org.apache.lucene.codecs.Postings
 import org.apache.lucene.codecs.SegmentInfosFormat;
 import org.apache.lucene.codecs.StoredFieldsFormat;
 import org.apache.lucene.codecs.TermVectorsFormat;
-import org.apache.lucene.codecs.lucene40.Lucene40SegmentInfosFormat;
 import org.apache.lucene.codecs.lucene40.Lucene40StoredFieldsFormat;
-import org.apache.lucene.codecs.lucene40.Lucene40TermVectorsFormat;
 import org.apache.lucene.index.PerDocWriteState;
 import org.apache.lucene.index.SegmentInfo;
 import org.apache.lucene.index.SegmentReadState;
@@ -55,12 +53,8 @@ public class Lucene3xCodec extends Codec
   
   private final FieldInfosFormat fieldInfosFormat = new Lucene3xFieldInfosFormat();
 
-  // TODO: this should really be a different impl
-  // also if we want preflex to *really* be read-only it should throw exception for the writer?
-  // this way IR.commit fails on delete/undelete/setNorm/etc ?
-  private final SegmentInfosFormat infosFormat = new Lucene40SegmentInfosFormat();
+  private final SegmentInfosFormat infosFormat = new Lucene3xSegmentInfosFormat();
   
-  // TODO: this should really be a different impl
   private final NormsFormat normsFormat = new Lucene3xNormsFormat();
   
   // 3.x doesn't support docvalues

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xFieldInfosFormat.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xFieldInfosFormat.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xFieldInfosFormat.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xFieldInfosFormat.java Thu Jan 19 14:06:10 2012
@@ -27,8 +27,12 @@ import org.apache.lucene.index.SegmentIn
 import org.apache.lucene.store.Directory;
 
 /**
+ * Lucene3x ReadOnly FieldInfosFromat implementation
+ * @deprecated (4.0) This is only used to read indexes created
+ * before 4.0.
  * @lucene.experimental
  */
+@Deprecated
 public class Lucene3xFieldInfosFormat extends FieldInfosFormat {
   private final FieldInfosReader reader = new Lucene3xFieldInfosReader();
   
@@ -39,7 +43,7 @@ public class Lucene3xFieldInfosFormat ex
 
   @Override
   public FieldInfosWriter getFieldInfosWriter() throws IOException {
-    throw new IllegalArgumentException("this codec can only be used for reading");
+    throw new UnsupportedOperationException("this codec can only be used for reading");
   }
 
   @Override

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xNormsFormat.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xNormsFormat.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xNormsFormat.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xNormsFormat.java Thu Jan 19 14:06:10 2012
@@ -29,10 +29,12 @@ import org.apache.lucene.index.SegmentRe
 import org.apache.lucene.store.Directory;
 
 /**
- * Read-Only Lucene 3.x Norms Format
- * 
+ * Lucene3x ReadOnly NormsFormat implementation
+ * @deprecated (4.0) This is only used to read indexes created
+ * before 4.0.
  * @lucene.experimental
  */
+@Deprecated
 public class Lucene3xNormsFormat extends NormsFormat {
 
 
@@ -49,7 +51,7 @@ public class Lucene3xNormsFormat extends
 
   @Override
   public PerDocConsumer docsConsumer(PerDocWriteState state) throws IOException {
-    throw new IllegalArgumentException("this codec can only be used for reading");
+    throw new UnsupportedOperationException("this codec can only be used for reading");
   }
 
   @Override

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xPostingsFormat.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xPostingsFormat.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xPostingsFormat.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xPostingsFormat.java Thu Jan 19 14:06:10 2012
@@ -57,7 +57,7 @@ public class Lucene3xPostingsFormat exte
   
   @Override
   public FieldsConsumer fieldsConsumer(SegmentWriteState state) throws IOException {
-    throw new IllegalArgumentException("this codec can only be used for reading");
+    throw new UnsupportedOperationException("this codec can only be used for reading");
   }
 
   @Override

Copied: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosFormat.java (from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java)
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosFormat.java?p2=lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosFormat.java&p1=lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java&r1=1233248&r2=1233380&rev=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosFormat.java Thu Jan 19 14:06:10 2012
@@ -1,4 +1,4 @@
-package org.apache.lucene.codecs.lucene40;
+package org.apache.lucene.codecs.lucene3x;
 
 import org.apache.lucene.codecs.SegmentInfosFormat;
 import org.apache.lucene.codecs.SegmentInfosReader;
@@ -22,11 +22,14 @@ import org.apache.lucene.codecs.SegmentI
  */
 
 /**
+ * Lucene3x ReadOnly SegmentInfosFormat implementation
+ * @deprecated (4.0) This is only used to read indexes created
+ * before 4.0.
  * @lucene.experimental
  */
-public class Lucene40SegmentInfosFormat extends SegmentInfosFormat {
-  private final SegmentInfosReader reader = new Lucene40SegmentInfosReader();
-  private final SegmentInfosWriter writer = new Lucene40SegmentInfosWriter();
+@Deprecated
+public class Lucene3xSegmentInfosFormat extends SegmentInfosFormat {
+  private final SegmentInfosReader reader = new Lucene3xSegmentInfosReader();
   
   @Override
   public SegmentInfosReader getSegmentInfosReader() {
@@ -35,6 +38,6 @@ public class Lucene40SegmentInfosFormat 
 
   @Override
   public SegmentInfosWriter getSegmentInfosWriter() {
-    return writer;
+    throw new UnsupportedOperationException("this codec can only be used for reading");
   }
 }

Copied: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosReader.java (from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java)
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosReader.java?p2=lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosReader.java&p1=lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java&r1=1233248&r2=1233380&rev=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xSegmentInfosReader.java Thu Jan 19 14:06:10 2012
@@ -1,4 +1,4 @@
-package org.apache.lucene.codecs.lucene40;
+package org.apache.lucene.codecs.lucene3x;
 
 /**
  * Licensed to the Apache Software Foundation (ASF) under one or more
@@ -23,7 +23,8 @@ import java.util.Map;
 
 import org.apache.lucene.codecs.Codec;
 import org.apache.lucene.codecs.SegmentInfosReader;
-import org.apache.lucene.codecs.lucene40.Lucene40TermVectorsReader;
+// TODO we need a 3x version of stored fields reader
+import org.apache.lucene.codecs.lucene40.Lucene40StoredFieldsReader;
 import org.apache.lucene.index.IndexFileNames;
 import org.apache.lucene.index.IndexFormatTooOldException;
 import org.apache.lucene.index.SegmentInfo;
@@ -34,19 +35,11 @@ import org.apache.lucene.store.Directory
 import org.apache.lucene.store.IOContext;
 
 /**
- * Default implementation of {@link SegmentInfosReader}.
+ * Lucene 3x implementation of {@link SegmentInfosReader}.
  * @lucene.experimental
  */
-public class Lucene40SegmentInfosReader extends SegmentInfosReader {
+public class Lucene3xSegmentInfosReader extends SegmentInfosReader {
 
-  // TODO: shove all backwards code to preflex!
-  // this is a little tricky, because of IR.commit(), two options:
-  // 1. PreFlex writes 4.x SIS format, but reads both 3.x and 4.x
-  //    (and maybe RW always only writes the 3.x one? for that to work well,
-  //     we have to move .fnx file to codec too, not too bad but more work).
-  //     or we just have crappier RW testing like today.
-  // 2. PreFlex writes 3.x SIS format, and only reads 3.x
-  //    (in this case we have to move .fnx file to codec as well)
   @Override
   public void read(Directory directory, String segmentsFileName, ChecksumIndexInput input, SegmentInfos infos, IOContext context) throws IOException { 
     infos.version = input.readLong(); // read version
@@ -115,11 +108,9 @@ public class Lucene40SegmentInfosReader 
       docStoreIsCompoundFile = false;
     }
 
-    if (format > SegmentInfos.FORMAT_4_0) {
-      // pre-4.0 indexes write a byte if there is a single norms file
-      byte b = input.readByte();
-      assert 1 == b;
-    }
+    // pre-4.0 indexes write a byte if there is a single norms file
+    byte b = input.readByte();
+    assert 1 == b : "expected 1 but was: "+ b + " format: " + format;
 
     final int numNormGen = input.readInt();
     final Map<Integer,Long> normGen;
@@ -128,12 +119,7 @@ public class Lucene40SegmentInfosReader 
     } else {
       normGen = new HashMap<Integer, Long>();
       for(int j=0;j<numNormGen;j++) {
-        int fieldNumber = j;
-        if (format <= SegmentInfos.FORMAT_4_0) {
-          fieldNumber = input.readInt();
-        }
-
-        normGen.put(fieldNumber, input.readLong());
+        normGen.put(j, input.readLong());
       }
     }
     final boolean isCompoundFile = input.readByte() == SegmentInfo.YES;
@@ -143,13 +129,7 @@ public class Lucene40SegmentInfosReader 
 
     final int hasProx = input.readByte();
 
-    final Codec codec;
-    // note: if the codec is not available: Codec.forName will throw an exception.
-    if (format <= SegmentInfos.FORMAT_4_0) {
-      codec = Codec.forName(input.readString());
-    } else {
-      codec = Codec.forName("Lucene3x");
-    }
+    final Codec codec = Codec.forName("Lucene3x");
     final Map<String,String> diagnostics = input.readStringStringMap();
 
     final int hasVectors;
@@ -175,8 +155,7 @@ public class Lucene40SegmentInfosReader 
         dirToTest = dir;
       }
       try {
-        // TODO: remove this manual file check or push to preflex codec
-        hasVectors = dirToTest.fileExists(IndexFileNames.segmentFileName(storesSegment, "", Lucene40TermVectorsReader.VECTORS_INDEX_EXTENSION)) ? SegmentInfo.YES : SegmentInfo.NO;
+        hasVectors = dirToTest.fileExists(IndexFileNames.segmentFileName(storesSegment, "", Lucene3xTermVectorsReader.VECTORS_INDEX_EXTENSION)) ? SegmentInfo.YES : SegmentInfo.NO;
       } finally {
         if (isCompoundFile) {
           dirToTest.close();

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xTermVectorsFormat.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xTermVectorsFormat.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xTermVectorsFormat.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene3x/Lucene3xTermVectorsFormat.java Thu Jan 19 14:06:10 2012
@@ -28,6 +28,13 @@ import org.apache.lucene.index.SegmentIn
 import org.apache.lucene.store.Directory;
 import org.apache.lucene.store.IOContext;
 
+/**
+ * Lucene3x ReadOnly TermVectorsFormat implementation
+ * @deprecated (4.0) This is only used to read indexes created
+ * before 4.0.
+ * @lucene.experimental
+ */
+@Deprecated
 public class Lucene3xTermVectorsFormat extends TermVectorsFormat {
 
   @Override
@@ -37,8 +44,7 @@ public class Lucene3xTermVectorsFormat e
 
   @Override
   public TermVectorsWriter vectorsWriter(Directory directory, String segment, IOContext context) throws IOException {
-    // TODO all these IAEs in preflex should be UOEs?
-    throw new IllegalArgumentException("this codec can only be used for reading");
+    throw new UnsupportedOperationException("this codec can only be used for reading");
   }
 
   @Override

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosReader.java Thu Jan 19 14:06:10 2012
@@ -23,13 +23,10 @@ import java.util.Map;
 
 import org.apache.lucene.codecs.Codec;
 import org.apache.lucene.codecs.SegmentInfosReader;
-import org.apache.lucene.codecs.lucene40.Lucene40TermVectorsReader;
-import org.apache.lucene.index.IndexFileNames;
 import org.apache.lucene.index.IndexFormatTooOldException;
 import org.apache.lucene.index.SegmentInfo;
 import org.apache.lucene.index.SegmentInfos;
 import org.apache.lucene.store.ChecksumIndexInput;
-import org.apache.lucene.store.CompoundFileDirectory;
 import org.apache.lucene.store.Directory;
 import org.apache.lucene.store.IOContext;
 
@@ -39,88 +36,37 @@ import org.apache.lucene.store.IOContext
  */
 public class Lucene40SegmentInfosReader extends SegmentInfosReader {
 
-  // TODO: shove all backwards code to preflex!
-  // this is a little tricky, because of IR.commit(), two options:
-  // 1. PreFlex writes 4.x SIS format, but reads both 3.x and 4.x
-  //    (and maybe RW always only writes the 3.x one? for that to work well,
-  //     we have to move .fnx file to codec too, not too bad but more work).
-  //     or we just have crappier RW testing like today.
-  // 2. PreFlex writes 3.x SIS format, and only reads 3.x
-  //    (in this case we have to move .fnx file to codec as well)
   @Override
   public void read(Directory directory, String segmentsFileName, ChecksumIndexInput input, SegmentInfos infos, IOContext context) throws IOException { 
     infos.version = input.readLong(); // read version
     infos.counter = input.readInt(); // read counter
     final int format = infos.getFormat();
+    assert format <= SegmentInfos.FORMAT_4_0;
     for (int i = input.readInt(); i > 0; i--) { // read segmentInfos
       SegmentInfo si = readSegmentInfo(directory, format, input);
-      if (si.getVersion() == null) {
-        // Could be a 3.0 - try to open the doc stores - if it fails, it's a
-        // 2.x segment, and an IndexFormatTooOldException will be thrown,
-        // which is what we want.
-        Directory dir = directory;
-        if (si.getDocStoreOffset() != -1) {
-          if (si.getDocStoreIsCompoundFile()) {
-            dir = new CompoundFileDirectory(dir, IndexFileNames.segmentFileName(
-                si.getDocStoreSegment(), "",
-                IndexFileNames.COMPOUND_FILE_STORE_EXTENSION), context, false);
-          }
-        } else if (si.getUseCompoundFile()) {
-          dir = new CompoundFileDirectory(dir, IndexFileNames.segmentFileName(
-              si.name, "", IndexFileNames.COMPOUND_FILE_EXTENSION), context, false);
-        }
-
-        try {
-          Lucene40StoredFieldsReader.checkCodeVersion(dir, si.getDocStoreSegment());
-        } finally {
-          // If we opened the directory, close it
-          if (dir != directory) dir.close();
-        }
-          
-        // Above call succeeded, so it's a 3.0 segment. Upgrade it so the next
-        // time the segment is read, its version won't be null and we won't
-        // need to open FieldsReader every time for each such segment.
-        si.setVersion("3.0");
-      } else if (si.getVersion().equals("2.x")) {
-        // If it's a 3x index touched by 3.1+ code, then segments record their
-        // version, whether they are 2.x ones or not. We detect that and throw
-        // appropriate exception.
-        throw new IndexFormatTooOldException("segment " + si.name + " in resource " + input, si.getVersion());
-      }
+      assert si.getVersion() != null;
       infos.add(si);
     }
       
     infos.userData = input.readStringStringMap();
   }
   
-  // if we make a preflex impl we can remove a lot of this hair...
   public SegmentInfo readSegmentInfo(Directory dir, int format, ChecksumIndexInput input) throws IOException {
-    final String version;
-    if (format <= SegmentInfos.FORMAT_3_1) {
-      version = input.readString();
-    } else {
-      version = null;
-    }
+    final String version = input.readString();
     final String name = input.readString();
     final int docCount = input.readInt();
     final long delGen = input.readLong();
+    // this is still written in 4.0 if we open a 3.x and upgrade the SI
     final int docStoreOffset = input.readInt();
     final String docStoreSegment;
     final boolean docStoreIsCompoundFile;
-    if (docStoreOffset != -1) {
+    if (docStoreOffset != -1) { 
       docStoreSegment = input.readString();
       docStoreIsCompoundFile = input.readByte() == SegmentInfo.YES;
     } else {
       docStoreSegment = name;
       docStoreIsCompoundFile = false;
     }
-
-    if (format > SegmentInfos.FORMAT_4_0) {
-      // pre-4.0 indexes write a byte if there is a single norms file
-      byte b = input.readByte();
-      assert 1 == b;
-    }
-
     final int numNormGen = input.readInt();
     final Map<Integer,Long> normGen;
     if (numNormGen == SegmentInfo.NO) {
@@ -128,61 +74,17 @@ public class Lucene40SegmentInfosReader 
     } else {
       normGen = new HashMap<Integer, Long>();
       for(int j=0;j<numNormGen;j++) {
-        int fieldNumber = j;
-        if (format <= SegmentInfos.FORMAT_4_0) {
-          fieldNumber = input.readInt();
-        }
-
-        normGen.put(fieldNumber, input.readLong());
+        normGen.put(input.readInt(), input.readLong());
       }
     }
     final boolean isCompoundFile = input.readByte() == SegmentInfo.YES;
 
     final int delCount = input.readInt();
     assert delCount <= docCount;
-
     final int hasProx = input.readByte();
-
-    final Codec codec;
-    // note: if the codec is not available: Codec.forName will throw an exception.
-    if (format <= SegmentInfos.FORMAT_4_0) {
-      codec = Codec.forName(input.readString());
-    } else {
-      codec = Codec.forName("Lucene3x");
-    }
+    final Codec codec = Codec.forName(input.readString());
     final Map<String,String> diagnostics = input.readStringStringMap();
-
-    final int hasVectors;
-    if (format <= SegmentInfos.FORMAT_HAS_VECTORS) {
-      hasVectors = input.readByte();
-    } else {
-      final String storesSegment;
-      final String ext;
-      final boolean storeIsCompoundFile;
-      if (docStoreOffset != -1) {
-        storesSegment = docStoreSegment;
-        storeIsCompoundFile = docStoreIsCompoundFile;
-        ext = IndexFileNames.COMPOUND_FILE_STORE_EXTENSION;
-      } else {
-        storesSegment = name;
-        storeIsCompoundFile = isCompoundFile;
-        ext = IndexFileNames.COMPOUND_FILE_EXTENSION;
-      }
-      final Directory dirToTest;
-      if (storeIsCompoundFile) {
-        dirToTest = new CompoundFileDirectory(dir, IndexFileNames.segmentFileName(storesSegment, "", ext), IOContext.READONCE, false);
-      } else {
-        dirToTest = dir;
-      }
-      try {
-        // TODO: remove this manual file check or push to preflex codec
-        hasVectors = dirToTest.fileExists(IndexFileNames.segmentFileName(storesSegment, "", Lucene40TermVectorsReader.VECTORS_INDEX_EXTENSION)) ? SegmentInfo.YES : SegmentInfo.NO;
-      } finally {
-        if (isCompoundFile) {
-          dirToTest.close();
-        }
-      }
-    }
+    final int hasVectors = input.readByte();
     
     return new SegmentInfo(dir, version, name, docCount, delGen, docStoreOffset,
       docStoreSegment, docStoreIsCompoundFile, normGen, isCompoundFile,

Modified: lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java (original)
+++ lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java Thu Jan 19 14:06:10 2012
@@ -43,6 +43,10 @@ public class Lucene40SegmentInfosWriter 
     IndexOutput out = createOutput(dir, segmentFileName, new IOContext(new FlushInfo(infos.size(), infos.totalDocCount())));
     boolean success = false;
     try {
+      /*
+       * TODO its not ideal that we write the format and the codecID inside the
+       * codec private classes but we read it in SegmentInfos.
+       */
       out.writeInt(SegmentInfos.FORMAT_CURRENT); // write FORMAT
       out.writeString(codecID); // write codecID
       out.writeLong(infos.version);
@@ -69,7 +73,7 @@ public class Lucene40SegmentInfosWriter 
     output.writeString(si.name);
     output.writeInt(si.docCount);
     output.writeLong(si.getDelGen());
-
+    // we still need to write this in 4.0 since we can open a 3.x with shared docStores
     output.writeInt(si.getDocStoreOffset());
     if (si.getDocStoreOffset() != -1) {
       output.writeString(si.getDocStoreSegment());

Modified: lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWCodec.java
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWCodec.java?rev=1233380&r1=1233379&r2=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWCodec.java (original)
+++ lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWCodec.java Thu Jan 19 14:06:10 2012
@@ -20,6 +20,7 @@ package org.apache.lucene.codecs.preflex
 import org.apache.lucene.codecs.FieldInfosFormat;
 import org.apache.lucene.codecs.NormsFormat;
 import org.apache.lucene.codecs.PostingsFormat;
+import org.apache.lucene.codecs.SegmentInfosFormat;
 import org.apache.lucene.codecs.TermVectorsFormat;
 import org.apache.lucene.codecs.lucene3x.Lucene3xCodec;
 import org.apache.lucene.util.LuceneTestCase;
@@ -33,6 +34,7 @@ public class PreFlexRWCodec extends Luce
   private final NormsFormat norms = new PreFlexRWNormsFormat();
   private final FieldInfosFormat fieldInfos = new PreFlexRWFieldInfosFormat();
   private final TermVectorsFormat termVectors = new PreFlexRWTermVectorsFormat();
+  private final SegmentInfosFormat segmentInfos = new PreFlexRWSegmentInfosFormat();
   
   @Override
   public PostingsFormat postingsFormat() {
@@ -53,6 +55,15 @@ public class PreFlexRWCodec extends Luce
   }
 
   @Override
+  public SegmentInfosFormat segmentInfosFormat() {
+    if (LuceneTestCase.PREFLEX_IMPERSONATION_IS_ACTIVE) {
+      return segmentInfos ;
+    } else {
+      return super.segmentInfosFormat();
+    }
+  }
+
+  @Override
   public FieldInfosFormat fieldInfosFormat() {
     if (LuceneTestCase.PREFLEX_IMPERSONATION_IS_ACTIVE) {
       return fieldInfos;

Copied: lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosFormat.java (from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java)
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosFormat.java?p2=lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosFormat.java&p1=lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java&r1=1233248&r2=1233380&rev=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosFormat.java (original)
+++ lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosFormat.java Thu Jan 19 14:06:10 2012
@@ -1,8 +1,7 @@
-package org.apache.lucene.codecs.lucene40;
+package org.apache.lucene.codecs.preflexrw;
 
-import org.apache.lucene.codecs.SegmentInfosFormat;
-import org.apache.lucene.codecs.SegmentInfosReader;
 import org.apache.lucene.codecs.SegmentInfosWriter;
+import org.apache.lucene.codecs.lucene3x.Lucene3xSegmentInfosFormat;
 
 /**
  * Licensed to the Apache Software Foundation (ASF) under one or more
@@ -24,16 +23,10 @@ import org.apache.lucene.codecs.SegmentI
 /**
  * @lucene.experimental
  */
-public class Lucene40SegmentInfosFormat extends SegmentInfosFormat {
-  private final SegmentInfosReader reader = new Lucene40SegmentInfosReader();
-  private final SegmentInfosWriter writer = new Lucene40SegmentInfosWriter();
+public class PreFlexRWSegmentInfosFormat extends Lucene3xSegmentInfosFormat {
+  private final SegmentInfosWriter writer = new PreFlexRWSegmentInfosWriter();
   
   @Override
-  public SegmentInfosReader getSegmentInfosReader() {
-    return reader;
-  }
-
-  @Override
   public SegmentInfosWriter getSegmentInfosWriter() {
     return writer;
   }

Copied: lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosWriter.java (from r1233248, lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java)
URL: http://svn.apache.org/viewvc/lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosWriter.java?p2=lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosWriter.java&p1=lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java&r1=1233248&r2=1233380&rev=1233380&view=diff
==============================================================================
--- lucene/dev/trunk/lucene/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfosWriter.java (original)
+++ lucene/dev/trunk/lucene/src/test-framework/java/org/apache/lucene/codecs/preflexrw/PreFlexRWSegmentInfosWriter.java Thu Jan 19 14:06:10 2012
@@ -1,4 +1,4 @@
-package org.apache.lucene.codecs.lucene40;
+package org.apache.lucene.codecs.preflexrw;
 
 /**
  * Licensed to the Apache Software Foundation (ASF) under one or more
@@ -32,10 +32,10 @@ import org.apache.lucene.store.IndexOutp
 import org.apache.lucene.util.IOUtils;
 
 /**
- * Default implementation of {@link SegmentInfosWriter}.
+ * PreFlex implementation of {@link SegmentInfosWriter}.
  * @lucene.experimental
  */
-public class Lucene40SegmentInfosWriter extends SegmentInfosWriter {
+public class PreFlexRWSegmentInfosWriter extends SegmentInfosWriter {
 
   @Override
   public IndexOutput writeInfos(Directory dir, String segmentFileName, String codecID, SegmentInfos infos, IOContext context)
@@ -43,8 +43,8 @@ public class Lucene40SegmentInfosWriter 
     IndexOutput out = createOutput(dir, segmentFileName, new IOContext(new FlushInfo(infos.size(), infos.totalDocCount())));
     boolean success = false;
     try {
-      out.writeInt(SegmentInfos.FORMAT_CURRENT); // write FORMAT
-      out.writeString(codecID); // write codecID
+      out.writeInt(SegmentInfos.FORMAT_3_1); // write FORMAT
+      // we don't write a codec - this is 3.x
       out.writeLong(infos.version);
       out.writeInt(infos.counter); // write counter
       out.writeInt(infos.size()); // write infos
@@ -75,6 +75,8 @@ public class Lucene40SegmentInfosWriter 
       output.writeString(si.getDocStoreSegment());
       output.writeByte((byte) (si.getDocStoreIsCompoundFile() ? 1:0));
     }
+    // pre-4.0 indexes write a byte if there is a single norms file
+    output.writeByte((byte) 1);
 
     Map<Integer,Long> normGen = si.getNormGen();
     if (normGen == null) {
@@ -82,7 +84,6 @@ public class Lucene40SegmentInfosWriter 
     } else {
       output.writeInt(normGen.size());
       for (Entry<Integer,Long> entry : normGen.entrySet()) {
-        output.writeInt(entry.getKey());
         output.writeLong(entry.getValue());
       }
     }
@@ -90,7 +91,6 @@ public class Lucene40SegmentInfosWriter 
     output.writeByte((byte) (si.getUseCompoundFile() ? SegmentInfo.YES : SegmentInfo.NO));
     output.writeInt(si.getDelCount());
     output.writeByte((byte) (si.getHasProxInternal()));
-    output.writeString(si.getCodec().getName());
     output.writeStringStringMap(si.getDiagnostics());
     output.writeByte((byte) (si.getHasVectorsInternal()));
   }