Updated Javadoc

src/main/java/org/jboss/netty/handler/stream/ChunkedInput.java

@@ -18,7 +18,7 @@ package org.jboss.netty.handler.stream;
 import org.jboss.netty.buffer.ChannelBuffer;
 
 /**
- * A large data stream which is consumed by {@link ChunkedWriteHandler}.
+ * A data stream of indefinite length which is consumed by {@link ChunkedWriteHandler}.
  *
  * @author <a href="http://www.jboss.org/netty/">The Netty Project</a>
  * @author <a href="http://gleamynode.net/">Trustin Lee</a>
@@ -40,7 +40,9 @@ public interface ChunkedInput {
      * Fetches a chunked data from the stream.  The returned chunk is usually
      * a {@link ChannelBuffer}, but you could extend an existing implementation
      * to convert the {@link ChannelBuffer} into a different type that your
-     * handler or encoder understands.
+     * handler or encoder understands.  Once this method returns the last chunk
+     * and thus the stream has reached at its end, any subsequent {@link #isEndOfInput()}
+     * call must return {@code false}.
      *
      * @return the fetched chunk, which is usually {@link ChannelBuffer}.
      *         {@code null} if there is no data left in the stream.

src/main/java/org/jboss/netty/handler/stream/ChunkedWriteHandler.java

@@ -62,8 +62,8 @@ import org.jboss.netty.util.internal.LinkedTransferQueue;
  * <h3>Sending a stream which generates a chunk intermittently</h3>
  *
  * Some {@link ChunkedInput} generates a chunk on a certain event or timing.
- * Such {@link ChunkedInput} implementation often returns {@code false} on
- * {@link ChunkedInput#hasNextChunk()}, resulting in the indefinitely suspended
+ * Such {@link ChunkedInput} implementation often returns {@code null} on
+ * {@link ChunkedInput#nextChunk()}, resulting in the indefinitely suspended
  * transfer.  To resume the transfer when a new chunk is available, you have to
  * call {@link #resumeTransfer()}.
  *