diff --git a/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40PostingsReader.java b/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40PostingsReader.java
index 0f4212d9f97..c51df5f572f 100644
--- a/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40PostingsReader.java
+++ b/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40PostingsReader.java
@@ -61,10 +61,14 @@ public class Lucene40PostingsReader extends PostingsReaderBase {
   public Lucene40PostingsReader(Directory dir, FieldInfos fieldInfos, SegmentInfo segmentInfo, IOContext ioContext, String segmentSuffix) throws IOException {
     freqIn = dir.openInput(IndexFileNames.segmentFileName(segmentInfo.name, segmentSuffix, Lucene40PostingsFormat.FREQ_EXTENSION),
                            ioContext);
-    // nocommit don't consult FieldInfos here... then nuke
-    // fieldInfos arg...
-    // nocommit we can assert FIS.hasProx == our hasProx here...
-    // this.segment = segmentInfo.name;
+    // TODO: hasProx should (somehow!) become codec private,
+    // but it's tricky because 1) FIS.hasProx is global (it
+    // could be all fields that have prox are written by a
+    // different codec), 2) the field may have had prox in
+    // the past but all docs w/ that field were deleted.
+    // Really we'd need to init prxOut lazily on write, and
+    // then somewhere record that we actually wrote it so we
+    // know whether to open on read:
     if (fieldInfos.hasProx()) {
       boolean success = false;
       try {
diff --git a/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfoWriter.java b/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfoWriter.java
index f39b86e7e2e..a85f1811fa3 100644
--- a/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfoWriter.java
+++ b/lucene/core/src/java/org/apache/lucene/codecs/lucene40/Lucene40SegmentInfoWriter.java
@@ -40,8 +40,7 @@ public class Lucene40SegmentInfoWriter extends SegmentInfoWriter {
   @Override
   public void write(Directory dir, SegmentInfo si, FieldInfos fis, IOContext ioContext) throws IOException {
     final String fileName = IndexFileNames.segmentFileName(si.name, "", Lucene40SegmentInfoFormat.SI_EXTENSION);
-    assert si.getFiles() != null;
-    si.getFiles().add(fileName);
+    si.addFile(fileName);
 
     final IndexOutput output = dir.createOutput(fileName, ioContext);
 
@@ -57,7 +56,7 @@ public class Lucene40SegmentInfoWriter extends SegmentInfoWriter {
       output.writeByte((byte) (si.getUseCompoundFile() ? SegmentInfo.YES : SegmentInfo.NO));
       output.writeStringStringMap(si.getDiagnostics());
       output.writeStringStringMap(si.attributes());
-      output.writeStringSet(si.getFiles());
+      output.writeStringSet(si.files());
 
       success = true;
     } finally {
diff --git a/lucene/core/src/java/org/apache/lucene/codecs/lucene40/values/DocValuesWriterBase.java b/lucene/core/src/java/org/apache/lucene/codecs/lucene40/values/DocValuesWriterBase.java
index b8d2c0a7f6e..146342161bd 100644
--- a/lucene/core/src/java/org/apache/lucene/codecs/lucene40/values/DocValuesWriterBase.java
+++ b/lucene/core/src/java/org/apache/lucene/codecs/lucene40/values/DocValuesWriterBase.java
@@ -67,7 +67,6 @@ public abstract class DocValuesWriterBase extends PerDocConsumer {
    *                         docvalues of type {@link Type#BYTES_FIXED_SORTED} and {@link Type#BYTES_VAR_SORTED}.
    */
   protected DocValuesWriterBase(PerDocWriteState state, boolean fasterButMoreRam) {
-    // nocommit save away SegmentInfo instead?
     this.segmentName = state.segmentInfo.name;
     this.bytesUsed = state.bytesUsed;
     this.context = state.context;
diff --git a/lucene/core/src/java/org/apache/lucene/codecs/simpletext/SimpleTextSegmentInfoWriter.java b/lucene/core/src/java/org/apache/lucene/codecs/simpletext/SimpleTextSegmentInfoWriter.java
index 11f1c4148de..8fed6db5aff 100644
--- a/lucene/core/src/java/org/apache/lucene/codecs/simpletext/SimpleTextSegmentInfoWriter.java
+++ b/lucene/core/src/java/org/apache/lucene/codecs/simpletext/SimpleTextSegmentInfoWriter.java
@@ -59,7 +59,7 @@ public class SimpleTextSegmentInfoWriter extends SegmentInfoWriter {
   public void write(Directory dir, SegmentInfo si, FieldInfos fis, IOContext ioContext) throws IOException {
 
     String segFileName = IndexFileNames.segmentFileName(si.name, "", SimpleTextSegmentInfoFormat.SI_EXTENSION);
-    si.getFiles().add(segFileName);
+    si.addFile(segFileName);
 
     boolean success = false;
     IndexOutput output = dir.createOutput(segFileName,  ioContext);
@@ -115,7 +115,7 @@ public class SimpleTextSegmentInfoWriter extends SegmentInfoWriter {
         }
       }
 
-      Set<String> files = si.getFiles();
+      Set<String> files = si.files();
       int numFiles = files == null ? 0 : files.size();
       SimpleTextUtil.write(output, SI_NUM_FILES);
       SimpleTextUtil.write(output, Integer.toString(numFiles), scratch);
diff --git a/lucene/core/src/java/org/apache/lucene/index/CheckIndex.java b/lucene/core/src/java/org/apache/lucene/index/CheckIndex.java
index 395f50fdcfb..1ca275d31df 100644
--- a/lucene/core/src/java/org/apache/lucene/index/CheckIndex.java
+++ b/lucene/core/src/java/org/apache/lucene/index/CheckIndex.java
@@ -479,8 +479,8 @@ public class CheckIndex {
         segInfoStat.codec = codec;
         msg("    compound=" + info.info.getUseCompoundFile());
         segInfoStat.compound = info.info.getUseCompoundFile();
-        msg("    numFiles=" + info.info.files().size());
-        segInfoStat.numFiles = info.info.files().size();
+        msg("    numFiles=" + info.files().size());
+        segInfoStat.numFiles = info.files().size();
         segInfoStat.sizeMB = info.sizeInBytes()/(1024.*1024.);
         msg("    size (MB)=" + nf.format(segInfoStat.sizeMB));
         Map<String,String> diagnostics = info.info.getDiagnostics();
diff --git a/lucene/core/src/java/org/apache/lucene/index/DocumentsWriterPerThread.java b/lucene/core/src/java/org/apache/lucene/index/DocumentsWriterPerThread.java
index d37775c3ba4..df14290d7a0 100644
--- a/lucene/core/src/java/org/apache/lucene/index/DocumentsWriterPerThread.java
+++ b/lucene/core/src/java/org/apache/lucene/index/DocumentsWriterPerThread.java
@@ -489,6 +489,7 @@ class DocumentsWriterPerThread {
       segmentInfo.docCount = flushState.numDocs;
       segmentInfo.setFiles(new HashSet<String>(directory.getCreatedFiles()));
 
+      final SegmentInfoPerCommit segmentInfoPerCommit = new SegmentInfoPerCommit(segmentInfo, 0, -1L);
       if (infoStream.isEnabled("DWPT")) {
         infoStream.message("DWPT", "new segment has " + (flushState.liveDocs == null ? 0 : (flushState.numDocs - flushState.delCountOnFlush)) + " deleted docs");
         infoStream.message("DWPT", "new segment has " +
@@ -497,7 +498,7 @@ class DocumentsWriterPerThread {
                            (flushState.fieldInfos.hasDocValues() ? "docValues" : "no docValues") + "; " + 
                            (flushState.fieldInfos.hasProx() ? "prox" : "no prox") + "; " + 
                            (flushState.fieldInfos.hasFreq() ? "freqs" : "no freqs"));
-        infoStream.message("DWPT", "flushedFiles=" + segmentInfo.files());
+        infoStream.message("DWPT", "flushedFiles=" + segmentInfoPerCommit.files());
         infoStream.message("DWPT", "flushed codec=" + codec);
       }
 
@@ -522,7 +523,7 @@ class DocumentsWriterPerThread {
 
       assert segmentInfo != null;
 
-      FlushedSegment fs = new FlushedSegment(new SegmentInfoPerCommit(segmentInfo, 0, -1L), flushState.fieldInfos,
+      FlushedSegment fs = new FlushedSegment(segmentInfoPerCommit, flushState.fieldInfos,
                                              segmentDeletes, flushState.liveDocs, flushState.delCountOnFlush);
       doAfterFlush();
       success = true;
diff --git a/lucene/core/src/java/org/apache/lucene/index/IndexWriter.java b/lucene/core/src/java/org/apache/lucene/index/IndexWriter.java
index 7891bf41da2..18449ebd6bb 100644
--- a/lucene/core/src/java/org/apache/lucene/index/IndexWriter.java
+++ b/lucene/core/src/java/org/apache/lucene/index/IndexWriter.java
@@ -1505,15 +1505,11 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
       // merge:
       for(final MergePolicy.OneMerge merge  : pendingMerges) {
         merge.maxNumSegments = maxNumSegments;
-        // nocommit: remove this, except it causes
-        // TestExternalCodecs.testPerFieldCodec failures:
         segmentsToMerge.put(merge.info, Boolean.TRUE);
       }
 
       for (final MergePolicy.OneMerge merge: runningMerges) {
         merge.maxNumSegments = maxNumSegments;
-        // nocommit: remove this, except it causes
-        // TestExternalCodecs.testPerFieldCodec failures:
         segmentsToMerge.put(merge.info, Boolean.TRUE);
       }
     }
@@ -2046,7 +2042,7 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
       // above:
       codec.segmentInfoFormat().getSegmentInfosWriter().write(directory, newSegment.info, flushedSegment.fieldInfos, context);
 
-      // nocommit ideally we would freeze newSegment here!!
+      // TODO: ideally we would freeze newSegment here!!
       // because any changes after writing the .si will be
       // lost... 
 
@@ -2341,7 +2337,7 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
       // and 2) .si reflects useCompoundFile=true change
       // above:
       codec.segmentInfoFormat().getSegmentInfosWriter().write(trackingDir, info, mergeState.fieldInfos, context);
-      info.getFiles().addAll(trackingDir.getCreatedFiles());
+      info.addFiles(trackingDir.getCreatedFiles());
 
       // Register the new segment
       synchronized(this) {
@@ -2410,7 +2406,8 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
 
     Set<String> segFiles = new HashSet<String>();
 
-    // Build up new segment's file names:
+    // Build up new segment's file names.  Must do this
+    // before writing SegmentInfo:
     for (String file: info.files()) {
       final String newFileName;
       if (codecDocStoreFiles.contains(file)) {
@@ -2421,17 +2418,20 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
       segFiles.add(newFileName);
     }
     newInfo.setFiles(segFiles);
-    
+
     // We must rewrite the SI file because it references
     // segment name (its own name, if its 3.x, and doc
     // store segment name):
+    TrackingDirectoryWrapper trackingDir = new TrackingDirectoryWrapper(directory);
     try {
-      newInfo.getCodec().segmentInfoFormat().getSegmentInfosWriter().write(directory, newInfo, null, context);
+      newInfo.getCodec().segmentInfoFormat().getSegmentInfosWriter().write(trackingDir, newInfo, null, context);
     } catch (UnsupportedOperationException uoe) {
       // OK: 3x codec cannot write a new SI file;
       // SegmentInfos will write this on commit
     }
 
+    final Collection<String> siFiles = trackingDir.getCreatedFiles();
+
     // Copy the segment's files
     for (String file: info.files()) {
 
@@ -2446,14 +2446,12 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
         newFileName = segName + IndexFileNames.stripSegmentName(file);
       }
 
-      // nocommit hack
-      //if (siFileNames != null && siFileNames.contains(newFileName)) {
-      if (newFileName.endsWith(".si")) {
+      if (siFiles.contains(newFileName)) {
         // We already rewrote this above
         continue;
       }
 
-      assert !directory.fileExists(newFileName): "file \"" + newFileName + "\" already exists";
+      assert !directory.fileExists(newFileName): "file \"" + newFileName + "\" already exists; siFiles=" + siFiles;
       assert !copiedFiles.contains(file): "file \"" + file + "\" is being copied more than once";
       copiedFiles.add(file);
       info.info.dir.copy(directory, file, newFileName, context);
@@ -3501,7 +3499,6 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
 
       // This is where all the work happens:
       MergeState mergeState = merger.merge();
-      // nocommit use setter and make this a SetOnce:
       merge.info.info.docCount = mergeState.mergedDocCount;
       merge.info.info.setFiles(new HashSet<String>(dirWrapper.getCreatedFiles()));
 
@@ -3597,7 +3594,7 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
         }
       }
 
-      // nocommit ideally we would freeze merge.info here!!
+      // TODO: ideally we would freeze merge.info here!!
       // because any changes after writing the .si will be
       // lost... 
 
@@ -4031,6 +4028,7 @@ public class IndexWriter implements Closeable, TwoPhaseCommit {
       IOUtils.closeWhileHandlingException(prior, cfsDir);
     }
 
+    // Replace all previous files with the CFS/CFE files:
     Set<String> siFiles = new HashSet<String>();
     siFiles.add(fileName);
     siFiles.add(IndexFileNames.segmentFileName(info.name, "", IndexFileNames.COMPOUND_FILE_ENTRIES_EXTENSION));
diff --git a/lucene/core/src/java/org/apache/lucene/index/SegmentInfo.java b/lucene/core/src/java/org/apache/lucene/index/SegmentInfo.java
index e7d01511b3b..030666d5ec9 100644
--- a/lucene/core/src/java/org/apache/lucene/index/SegmentInfo.java
+++ b/lucene/core/src/java/org/apache/lucene/index/SegmentInfo.java
@@ -20,6 +20,7 @@ package org.apache.lucene.index;
 
 import java.io.IOException;
 import java.util.Collection;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.Map;
 import java.util.Set;
@@ -28,9 +29,6 @@ import org.apache.lucene.codecs.Codec;
 import org.apache.lucene.store.Directory;
 import org.apache.lucene.store.TrackingDirectoryWrapper;
 
-// nocommit fix codec api to pass this around so they can
-// store attrs
-
 /**
  * Information about a segment such as it's name, directory, and files related
  * to the segment.
@@ -44,12 +42,9 @@ public final class SegmentInfo {
   public static final int YES = 1;          // e.g. have norms; have deletes;
 
   public final String name;				  // unique name in dir
-  // nocommit make me final:
   public int docCount;				  // number of docs in seg
   public final Directory dir;				  // where segment resides
 
-  // nocommit what other members can we make final?
-
   /*
    * Current generation of each field's norm file. If this array is null,
    * means no separate norms. If this array is not null, its values mean:
@@ -82,7 +77,6 @@ public final class SegmentInfo {
   // The format expected is "x.y" - "2.x" for pre-3.0 indexes (or null), and
   // specific versions afterwards ("3.0", "3.1" etc.).
   // see Constants.LUCENE_MAIN_VERSION.
-  // nocommit final?
   private String version;
 
   void setDiagnostics(Map<String, String> diagnostics) {
@@ -117,9 +111,11 @@ public final class SegmentInfo {
   }
 
   /**
-   * Returns total size in bytes of all of files used by this segment
+   * Returns total size in bytes of all of files used by
+   * this segment.  Note that this will not include any live
+   * docs for the segment; to include that use {@link
+   * SegmentInfoPerCommit.sizeInBytes} instead.
    */
-  // nocommit fails to take live docs into account... hmmm
   public long sizeInBytes() throws IOException {
     if (sizeInBytes == -1) {
       long sum = 0;
@@ -131,10 +127,6 @@ public final class SegmentInfo {
     return sizeInBytes;
   }
 
-  void clearSizeInBytes() {
-    sizeInBytes = -1;
-  }
-
   /**
    * @deprecated separate norms are not supported in >= 4.0
    */
@@ -217,15 +209,11 @@ public final class SegmentInfo {
    * modify it.
    */
 
-  // nocommit remove this temporarily to see who is calling
-  // it ...  very dangerous having this one AND SIPC.files()
-  public Collection<String> files() throws IOException {
-    // nocommit make sure when we are called we really have
-    // files set ...
+  public Set<String> files() throws IOException {
     if (setFiles == null) {
       throw new IllegalStateException("files were not computed yet");
     }
-    return setFiles;
+    return Collections.unmodifiableSet(setFiles);
   }
 
   /** {@inheritDoc} */
@@ -320,19 +308,19 @@ public final class SegmentInfo {
 
   private Set<String> setFiles;
 
-  // nocommit now on building a CFS we erase the files that
-  // are in it... maybe we should somehow preserve it...
   public void setFiles(Set<String> files) {
     setFiles = files;
     sizeInBytes = -1;
   }
 
-  // nocommit remove this!  it's only needed for
-  // clearing/adding the files set...
-  public Set<String> getFiles() {
-    return setFiles;
+  public void addFiles(Collection<String> files) {
+    setFiles.addAll(files);
   }
-  
+
+  public void addFile(String file) {
+    setFiles.add(file);
+  }
+    
   /**
    * Get a codec attribute value, or null if it does not exist
    */
diff --git a/lucene/core/src/java/org/apache/lucene/index/SegmentInfoPerCommit.java b/lucene/core/src/java/org/apache/lucene/index/SegmentInfoPerCommit.java
index 82b4edb75ae..07701a53d40 100644
--- a/lucene/core/src/java/org/apache/lucene/index/SegmentInfoPerCommit.java
+++ b/lucene/core/src/java/org/apache/lucene/index/SegmentInfoPerCommit.java
@@ -41,6 +41,8 @@ public class SegmentInfoPerCommit {
   // are no deletes yet):
   private long delGen;
 
+  private volatile long sizeInBytes = -1;
+
   public SegmentInfoPerCommit(SegmentInfo info, int delCount, long delGen) {
     this.info = info;
     this.delCount = delCount;
@@ -53,18 +55,26 @@ public class SegmentInfoPerCommit {
     } else {
       delGen++;
     }
-    info.clearSizeInBytes();
+    sizeInBytes = -1;
   }
 
   public long sizeInBytes() throws IOException {
-    // nocommit add in live docs size
-    return info.sizeInBytes();
+    if (sizeInBytes == -1) {
+      final Collection<String> files = new HashSet<String>();
+      info.getCodec().liveDocsFormat().files(this, files);
+      long sum = info.sizeInBytes();
+      for (final String fileName : files()) {
+        sum += info.dir.fileLength(fileName);
+      }
+      sizeInBytes = sum;
+    }
+
+    return sizeInBytes;
   }
 
   public Collection<String> files() throws IOException {
     Collection<String> files = new HashSet<String>(info.files());
 
-    // nocommit make this take list instead...?
     // Must separately add any live docs files:
     info.getCodec().liveDocsFormat().files(this, files);
 
@@ -81,16 +91,17 @@ public class SegmentInfoPerCommit {
 
   void setBufferedDeletesGen(long v) {
     bufferedDeletesGen = v;
+    sizeInBytes =  -1;
   }
   
   void clearDelGen() {
     delGen = -1;
-    info.clearSizeInBytes();
+    sizeInBytes =  -1;
   }
 
   public void setDelGen(long delGen) {
     this.delGen = delGen;
-    info.clearSizeInBytes();
+    sizeInBytes =  -1;
   }
 
   public boolean hasDeletions() {
diff --git a/lucene/core/src/java/org/apache/lucene/index/SegmentInfos.java b/lucene/core/src/java/org/apache/lucene/index/SegmentInfos.java
index f5075d8bfb6..0b97360d76d 100644
--- a/lucene/core/src/java/org/apache/lucene/index/SegmentInfos.java
+++ b/lucene/core/src/java/org/apache/lucene/index/SegmentInfos.java
@@ -411,7 +411,7 @@ public final class SegmentInfos implements Cloneable, Iterable<SegmentInfoPerCom
 
     // NOTE: this is NOT how 3.x is really written...
     String fileName = IndexFileNames.segmentFileName(si.name, "", Lucene3xSegmentInfoFormat.SI_EXTENSION);
-    si.getFiles().add(fileName);
+    si.addFile(fileName);
 
     //System.out.println("UPGRADE write " + fileName);
     boolean success = false;
@@ -456,7 +456,7 @@ public final class SegmentInfos implements Cloneable, Iterable<SegmentInfoPerCom
       output.writeStringStringMap(si.getDiagnostics());
       // hasVectors (lie):
       output.writeByte((byte) 1);
-      output.writeStringSet(si.getFiles());
+      output.writeStringSet(si.files());
 
       success = true;
     } finally {
diff --git a/lucene/core/src/test/org/apache/lucene/index/TestTermVectorsReader.java b/lucene/core/src/test/org/apache/lucene/index/TestTermVectorsReader.java
index b2706c0f109..67869966a0c 100644
--- a/lucene/core/src/test/org/apache/lucene/index/TestTermVectorsReader.java
+++ b/lucene/core/src/test/org/apache/lucene/index/TestTermVectorsReader.java
@@ -189,16 +189,6 @@ public class TestTermVectorsReader extends LuceneTestCase {
     for (IndexReader r : reader.getSequentialSubReaders()) {
       SegmentInfoPerCommit s = ((SegmentReader) r).getSegmentInfo();
       assertTrue(((SegmentReader) r).getFieldInfos().hasVectors());
-
-      // nocommit
-      /*
-      Set<String> files = new HashSet<String>();
-      s.getCodec().termVectorsFormat().files(s, files);
-      assertFalse(files.isEmpty());
-      for (String file : files) {
-        assertTrue(dir.fileExists(file));
-      }
-      */
     }
     reader.close();
   }