diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
index 5d28a1d2072..8cf32e91431 100644
--- a/lucene/CHANGES.txt
+++ b/lucene/CHANGES.txt
@@ -19,6 +19,10 @@ Changes in backwards compatibility policy
   (Nikola Tanković, Uwe Schindler, Chris Male, Mike McCandless,
   Robert Muir)
 
+* LUCENE-4924: DocIdSetIterator.docID() must now return -1 when the iterator is
+  not positioned. This change affects all classes that inherit from
+  DocIdSetIterator, including DocsEnum and DocsAndPositionsEnum. (Adrien Grand)
+
 New Features
 
 * LUCENE-4747: Move to Java 7 as minimum Java version.
diff --git a/lucene/core/src/java/org/apache/lucene/search/DocIdSet.java b/lucene/core/src/java/org/apache/lucene/search/DocIdSet.java
index 6df23b17bd2..901fe7f2364 100644
--- a/lucene/core/src/java/org/apache/lucene/search/DocIdSet.java
+++ b/lucene/core/src/java/org/apache/lucene/search/DocIdSet.java
@@ -29,20 +29,36 @@ public abstract class DocIdSet {
   /** An empty {@code DocIdSet} instance for easy use, e.g. in Filters that hit no documents. */
   public static final DocIdSet EMPTY_DOCIDSET = new DocIdSet() {
     
-    private final DocIdSetIterator iterator = new DocIdSetIterator() {
-      @Override
-      public int advance(int target) { return NO_MORE_DOCS; }
-      @Override
-      public int docID() { return NO_MORE_DOCS; }
-      @Override
-      public int nextDoc() { return NO_MORE_DOCS; }
-      @Override
-      public long cost() { return 0; }
-    };
-    
     @Override
     public DocIdSetIterator iterator() {
-      return iterator;
+      return new DocIdSetIterator() {
+        boolean exhausted = false;
+
+        @Override
+        public int advance(int target) {
+          assert !exhausted;
+          assert target >= 0;
+          exhausted = true;
+          return NO_MORE_DOCS;
+        }
+
+        @Override
+        public int docID() {
+          return exhausted ? NO_MORE_DOCS : -1;
+        }
+
+        @Override
+        public int nextDoc() {
+          assert !exhausted;
+          exhausted = true;
+          return NO_MORE_DOCS;
+        }
+
+        @Override
+        public long cost() {
+          return 0;
+        }
+      };
     }
     
     @Override
@@ -56,7 +72,7 @@ public abstract class DocIdSet {
       return null;
     }
   };
-    
+
   /** Provides a {@link DocIdSetIterator} to access the set.
    * This implementation can return <code>null</code> or
    * <code>{@linkplain #EMPTY_DOCIDSET}.iterator()</code> if there
diff --git a/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java b/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java
index 6aa870ee411..e1e328b7129 100644
--- a/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java
+++ b/lucene/core/src/java/org/apache/lucene/search/DocIdSetIterator.java
@@ -37,7 +37,7 @@ public abstract class DocIdSetIterator {
   /**
    * Returns the following:
    * <ul>
-   * <li>-1 or {@link #NO_MORE_DOCS} if {@link #nextDoc()} or
+   * <li><code>-1</code> if {@link #nextDoc()} or
    * {@link #advance(int)} were not called yet.
    * <li>{@link #NO_MORE_DOCS} if the iterator has exhausted.
    * <li>Otherwise it should return the doc ID it is currently on.
@@ -96,8 +96,7 @@ public abstract class DocIdSetIterator {
   /** Slow (linear) implementation of {@link #advance} relying on
    *  {@link #nextDoc()} to advance beyond the target position. */
   protected final int slowAdvance(int target) throws IOException {
-    assert docID() == NO_MORE_DOCS // can happen when the enum is not positioned yet
-        || docID() < target;
+    assert docID() < target;
     int doc;
     do {
       doc = nextDoc();
diff --git a/lucene/core/src/java/org/apache/lucene/search/Scorer.java b/lucene/core/src/java/org/apache/lucene/search/Scorer.java
index 8ece8c3cc6d..47fef12cebe 100644
--- a/lucene/core/src/java/org/apache/lucene/search/Scorer.java
+++ b/lucene/core/src/java/org/apache/lucene/search/Scorer.java
@@ -58,7 +58,7 @@ public abstract class Scorer extends DocsEnum {
    * @param collector The collector to which all matching documents are passed.
    */
   public void score(Collector collector) throws IOException {
-    assert docID() == -1 || docID() == NO_MORE_DOCS; // not started
+    assert docID() == -1; // not started
     collector.setScorer(this);
     int doc;
     while ((doc = nextDoc()) != NO_MORE_DOCS) {
diff --git a/lucene/core/src/java/org/apache/lucene/search/spans/SpanScorer.java b/lucene/core/src/java/org/apache/lucene/search/spans/SpanScorer.java
index b1428c3615c..a362763febf 100644
--- a/lucene/core/src/java/org/apache/lucene/search/spans/SpanScorer.java
+++ b/lucene/core/src/java/org/apache/lucene/search/spans/SpanScorer.java
@@ -42,12 +42,8 @@ public class SpanScorer extends Scorer {
     this.docScorer = docScorer;
     this.spans = spans;
 
-    if (this.spans.next()) {
-      doc = -1;
-    } else {
-      doc = NO_MORE_DOCS;
-      more = false;
-    }
+    doc = -1;
+    more = spans.next();
   }
 
   @Override
diff --git a/lucene/core/src/test/org/apache/lucene/codecs/lucene41/TestBlockPostingsFormat3.java b/lucene/core/src/test/org/apache/lucene/codecs/lucene41/TestBlockPostingsFormat3.java
index cff6b0aa608..19d8a1a0a50 100644
--- a/lucene/core/src/test/org/apache/lucene/codecs/lucene41/TestBlockPostingsFormat3.java
+++ b/lucene/core/src/test/org/apache/lucene/codecs/lucene41/TestBlockPostingsFormat3.java
@@ -394,8 +394,8 @@ public class TestBlockPostingsFormat3 extends LuceneTestCase {
       assertNull(rightDocs);
       return;
     }
-    assertTrue(leftDocs.docID() == -1 || leftDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
-    assertTrue(rightDocs.docID() == -1 || rightDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, leftDocs.docID());
+    assertEquals(-1, rightDocs.docID());
     int docid;
     while ((docid = leftDocs.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
       assertEquals(docid, rightDocs.nextDoc());
@@ -417,8 +417,8 @@ public class TestBlockPostingsFormat3 extends LuceneTestCase {
       assertNull(rightDocs);
       return;
     }
-    assertTrue(leftDocs.docID() == -1 || leftDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
-    assertTrue(rightDocs.docID() == -1 || rightDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, leftDocs.docID());
+    assertEquals(-1, rightDocs.docID());
     int docid;
     while ((docid = leftDocs.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
       assertEquals(docid, rightDocs.nextDoc());
diff --git a/lucene/core/src/test/org/apache/lucene/index/TestDocsAndPositions.java b/lucene/core/src/test/org/apache/lucene/index/TestDocsAndPositions.java
index c0d297ab084..80d0d7375ba 100644
--- a/lucene/core/src/test/org/apache/lucene/index/TestDocsAndPositions.java
+++ b/lucene/core/src/test/org/apache/lucene/index/TestDocsAndPositions.java
@@ -336,7 +336,7 @@ public class TestDocsAndPositions extends LuceneTestCase {
     AtomicReader r = getOnlySegmentReader(reader);
     DocsEnum disi = _TestUtil.docs(random(), r, "foo", new BytesRef("bar"), null, null, DocsEnum.FLAG_NONE);
     int docid = disi.docID();
-    assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, docid);
     assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
     
     // now reuse and check again
@@ -344,7 +344,7 @@ public class TestDocsAndPositions extends LuceneTestCase {
     assertTrue(te.seekExact(new BytesRef("bar"), true));
     disi = _TestUtil.docs(random(), te, null, disi, DocsEnum.FLAG_NONE);
     docid = disi.docID();
-    assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, docid);
     assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
     writer.close();
     r.close();
@@ -361,7 +361,7 @@ public class TestDocsAndPositions extends LuceneTestCase {
     AtomicReader r = getOnlySegmentReader(reader);
     DocsAndPositionsEnum disi = r.termPositionsEnum(new Term("foo", "bar"));
     int docid = disi.docID();
-    assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, docid);
     assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
     
     // now reuse and check again
@@ -369,7 +369,7 @@ public class TestDocsAndPositions extends LuceneTestCase {
     assertTrue(te.seekExact(new BytesRef("bar"), true));
     disi = te.docsAndPositions(null, disi);
     docid = disi.docID();
-    assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, docid);
     assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
     writer.close();
     r.close();
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 1fbbdd72f68..a57ace2a7ce 100644
--- a/lucene/core/src/test/org/apache/lucene/index/TestTermVectorsReader.java
+++ b/lucene/core/src/test/org/apache/lucene/index/TestTermVectorsReader.java
@@ -229,7 +229,7 @@ public class TestTermVectorsReader extends LuceneTestCase {
         docsEnum = _TestUtil.docs(random(), termsEnum, null, docsEnum, DocsEnum.FLAG_NONE);
         assertNotNull(docsEnum);
         int doc = docsEnum.docID();
-        assertTrue(doc == -1 || doc == DocIdSetIterator.NO_MORE_DOCS);
+        assertEquals(-1, doc);
         assertTrue(docsEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
         assertEquals(DocIdSetIterator.NO_MORE_DOCS, docsEnum.nextDoc());
       }
@@ -256,7 +256,7 @@ public class TestTermVectorsReader extends LuceneTestCase {
       dpEnum = termsEnum.docsAndPositions(null, dpEnum);
       assertNotNull(dpEnum);
       int doc = dpEnum.docID();
-      assertTrue(doc == -1 || doc == DocIdSetIterator.NO_MORE_DOCS);
+      assertEquals(-1, doc);
       assertTrue(dpEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
       assertEquals(dpEnum.freq(), positions[i].length);
       for (int j = 0; j < positions[i].length; j++) {
@@ -266,7 +266,7 @@ public class TestTermVectorsReader extends LuceneTestCase {
 
       dpEnum = termsEnum.docsAndPositions(null, dpEnum);
       doc = dpEnum.docID();
-      assertTrue(doc == -1 || doc == DocIdSetIterator.NO_MORE_DOCS);
+      assertEquals(-1, doc);
       assertTrue(dpEnum.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
       assertNotNull(dpEnum);
       assertEquals(dpEnum.freq(), positions[i].length);
diff --git a/lucene/join/src/java/org/apache/lucene/search/join/TermsIncludingScoreQuery.java b/lucene/join/src/java/org/apache/lucene/search/join/TermsIncludingScoreQuery.java
index df3520aef4f..3d1f91268e5 100644
--- a/lucene/join/src/java/org/apache/lucene/search/join/TermsIncludingScoreQuery.java
+++ b/lucene/join/src/java/org/apache/lucene/search/join/TermsIncludingScoreQuery.java
@@ -193,29 +193,22 @@ class TermsIncludingScoreQuery extends Query {
     DocsEnum docsEnum;
     DocsEnum reuse;
     int scoreUpto;
+    int doc;
 
     SVInnerScorer(Weight weight, Bits acceptDocs, TermsEnum termsEnum, long cost) {
       super(weight);
       this.acceptDocs = acceptDocs;
       this.termsEnum = termsEnum;
       this.cost = cost;
+      this.doc = -1;
     }
 
     @Override
     public void score(Collector collector) throws IOException {
-      score(collector, NO_MORE_DOCS, nextDocOutOfOrder());
-    }
-
-    @Override
-    public boolean score(Collector collector, int max, int firstDocID)
-        throws IOException {
-      assert collector.acceptsDocsOutOfOrder();
       collector.setScorer(this);
-      int doc;
-      for (doc = firstDocID; doc < max; doc = nextDocOutOfOrder()) {
+      for (int doc = nextDocOutOfOrder(); doc != NO_MORE_DOCS; doc = nextDocOutOfOrder()) {
         collector.collect(doc);
       }
-      return doc != NO_MORE_DOCS;
     }
 
     @Override
@@ -229,7 +222,7 @@ class TermsIncludingScoreQuery extends Query {
 
     @Override
     public int docID() {
-      return docsEnum != null ? docsEnum.docID() : DocIdSetIterator.NO_MORE_DOCS;
+      return doc;
     }
 
     int nextDocOutOfOrder() throws IOException {
@@ -238,13 +231,13 @@ class TermsIncludingScoreQuery extends Query {
         if (docId == DocIdSetIterator.NO_MORE_DOCS) {
           docsEnum = null;
         } else {
-          return docId;
+          return doc = docId;
         }
       }
 
       do {
         if (upto == terms.size()) {
-          return DocIdSetIterator.NO_MORE_DOCS;
+          return doc = DocIdSetIterator.NO_MORE_DOCS;
         }
 
         scoreUpto = upto;
@@ -253,7 +246,7 @@ class TermsIncludingScoreQuery extends Query {
         }
       } while (docsEnum == null);
 
-      return docsEnum.nextDoc();
+      return doc = docsEnum.nextDoc();
     }
 
     @Override
diff --git a/lucene/memory/src/test/org/apache/lucene/index/memory/MemoryIndexTest.java b/lucene/memory/src/test/org/apache/lucene/index/memory/MemoryIndexTest.java
index 0a26bea34dd..5b1ff01997a 100644
--- a/lucene/memory/src/test/org/apache/lucene/index/memory/MemoryIndexTest.java
+++ b/lucene/memory/src/test/org/apache/lucene/index/memory/MemoryIndexTest.java
@@ -323,7 +323,7 @@ public class MemoryIndexTest extends BaseTokenStreamTestCase {
     AtomicReader reader = (AtomicReader) memory.createSearcher().getIndexReader();
     DocsEnum disi = _TestUtil.docs(random(), reader, "foo", new BytesRef("bar"), null, null, DocsEnum.FLAG_NONE);
     int docid = disi.docID();
-    assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, docid);
     assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
     
     // now reuse and check again
@@ -331,7 +331,7 @@ public class MemoryIndexTest extends BaseTokenStreamTestCase {
     assertTrue(te.seekExact(new BytesRef("bar"), true));
     disi = te.docs(null, disi, DocsEnum.FLAG_NONE);
     docid = disi.docID();
-    assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(-1, docid);
     assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
     reader.close();
   }
@@ -354,7 +354,7 @@ public class MemoryIndexTest extends BaseTokenStreamTestCase {
       assertEquals(1, reader.terms("foo").getSumTotalTermFreq());
       DocsAndPositionsEnum disi = reader.termPositionsEnum(new Term("foo", "bar"));
       int docid = disi.docID();
-      assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+      assertEquals(-1, docid);
       assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
       assertEquals(0, disi.nextPosition());
       assertEquals(0, disi.startOffset());
@@ -365,7 +365,7 @@ public class MemoryIndexTest extends BaseTokenStreamTestCase {
       assertTrue(te.seekExact(new BytesRef("bar"), true));
       disi = te.docsAndPositions(null, disi);
       docid = disi.docID();
-      assertTrue(docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS);
+      assertEquals(-1, docid);
       assertTrue(disi.nextDoc() != DocIdSetIterator.NO_MORE_DOCS);
       reader.close();
       memory.reset();
diff --git a/lucene/test-framework/src/java/org/apache/lucene/index/AssertingAtomicReader.java b/lucene/test-framework/src/java/org/apache/lucene/index/AssertingAtomicReader.java
index 2a991205fdc..bbcec7aaa40 100644
--- a/lucene/test-framework/src/java/org/apache/lucene/index/AssertingAtomicReader.java
+++ b/lucene/test-framework/src/java/org/apache/lucene/index/AssertingAtomicReader.java
@@ -237,7 +237,7 @@ public class AssertingAtomicReader extends FilterAtomicReader {
       super(in);
       try {
         int docid = in.docID();
-        assert docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS : in.getClass() + ": invalid initial doc id: " + docid;
+        assert docid == -1 : in.getClass() + ": invalid initial doc id: " + docid;
       } catch (UnsupportedOperationException e) {
         if (failOnUnsupportedDocID) {
           throw e;
@@ -256,7 +256,7 @@ public class AssertingAtomicReader extends FilterAtomicReader {
       } else {
         state = DocsEnumState.ITERATING;
       }
-      assert docID() == nextDoc;
+      assert super.docID() == nextDoc;
       return doc = nextDoc;
     }
 
@@ -271,13 +271,15 @@ public class AssertingAtomicReader extends FilterAtomicReader {
       } else {
         state = DocsEnumState.ITERATING;
       }
-      assert docID() == advanced;
+      assert super.docID() == advanced;
       return doc = advanced;
     }
 
-    // NOTE: We don't assert anything for docId(). Specifically DocsEnum javadocs
-    // are ambiguous with DocIdSetIterator here, DocIdSetIterator says its ok
-    // to call this method before nextDoc(), just that it must be -1 or NO_MORE_DOCS!
+    @Override
+    public int docID() {
+      assert doc == super.docID() : " invalid docID() in " + in.getClass() + " " + super.docID() + " instead of " + doc;
+      return doc;
+    }
 
     @Override
     public int freq() throws IOException {
@@ -298,7 +300,7 @@ public class AssertingAtomicReader extends FilterAtomicReader {
     public AssertingDocsAndPositionsEnum(DocsAndPositionsEnum in) {
       super(in);
       int docid = in.docID();
-      assert docid == -1 || docid == DocIdSetIterator.NO_MORE_DOCS : "invalid initial doc id: " + docid;
+      assert docid == -1 : "invalid initial doc id: " + docid;
       doc = -1;
     }
 
@@ -315,7 +317,7 @@ public class AssertingAtomicReader extends FilterAtomicReader {
         state = DocsEnumState.ITERATING;
         positionMax = super.freq();
       }
-      assert docID() == nextDoc;
+      assert super.docID() == nextDoc;
       return doc = nextDoc;
     }
 
@@ -333,10 +335,16 @@ public class AssertingAtomicReader extends FilterAtomicReader {
         state = DocsEnumState.ITERATING;
         positionMax = super.freq();
       }
-      assert docID() == advanced;
+      assert super.docID() == advanced;
       return doc = advanced;
     }
 
+    @Override
+    public int docID() {
+      assert doc == super.docID() : " invalid docID() in " + in.getClass() + " " + super.docID() + " instead of " + doc;
+      return doc;
+    }
+
     @Override
     public int freq() throws IOException {
       assert state != DocsEnumState.START : "freq() called before nextDoc()/advance()";
diff --git a/lucene/test-framework/src/java/org/apache/lucene/index/BasePostingsFormatTestCase.java b/lucene/test-framework/src/java/org/apache/lucene/index/BasePostingsFormatTestCase.java
index de0fcc008c0..fd3f90be826 100644
--- a/lucene/test-framework/src/java/org/apache/lucene/index/BasePostingsFormatTestCase.java
+++ b/lucene/test-framework/src/java/org/apache/lucene/index/BasePostingsFormatTestCase.java
@@ -66,9 +66,6 @@ import org.junit.BeforeClass;
 
 // TODO test when you reuse after skipping a term or two, eg the block reuse case
 
-// TODO hmm contract says .doc() can return NO_MORE_DOCS
-// before nextDoc too...?
-
 /* TODO
   - threads
   - assert doc=-1 before any nextDoc
@@ -699,7 +696,7 @@ public abstract class BasePostingsFormatTestCase extends LuceneTestCase {
 
     assertNotNull("null DocsEnum", docsEnum);
     int initialDocID = docsEnum.docID();
-    assertTrue("inital docID should be -1 or NO_MORE_DOCS: " + docsEnum, initialDocID == -1 || initialDocID == DocsEnum.NO_MORE_DOCS);
+    assertEquals("inital docID should be -1" + docsEnum, -1, initialDocID);
 
     if (VERBOSE) {
       if (prevDocsEnum == null) {
diff --git a/lucene/test-framework/src/java/org/apache/lucene/util/LuceneTestCase.java b/lucene/test-framework/src/java/org/apache/lucene/util/LuceneTestCase.java
index 447d1e465f5..1efa453817f 100644
--- a/lucene/test-framework/src/java/org/apache/lucene/util/LuceneTestCase.java
+++ b/lucene/test-framework/src/java/org/apache/lucene/util/LuceneTestCase.java
@@ -1534,8 +1534,8 @@ public abstract class LuceneTestCase extends Assert {
       assertNull(rightDocs);
       return;
     }
-    assertTrue(info, leftDocs.docID() == -1 || leftDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
-    assertTrue(info, rightDocs.docID() == -1 || rightDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(info, -1, leftDocs.docID());
+    assertEquals(info, -1, rightDocs.docID());
     int docid;
     while ((docid = leftDocs.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
       assertEquals(info, docid, rightDocs.nextDoc());
@@ -1559,8 +1559,8 @@ public abstract class LuceneTestCase extends Assert {
       assertNull(rightDocs);
       return;
     }
-    assertTrue(info, leftDocs.docID() == -1 || leftDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
-    assertTrue(info, rightDocs.docID() == -1 || rightDocs.docID() == DocIdSetIterator.NO_MORE_DOCS);
+    assertEquals(info, -1, leftDocs.docID());
+    assertEquals(info, -1, rightDocs.docID());
     int docid;
     while ((docid = leftDocs.nextDoc()) != DocIdSetIterator.NO_MORE_DOCS) {
       assertEquals(info, docid, rightDocs.nextDoc());