Minor robustification, javadoc improvements
diff --git a/src/main/java/com/fasterxml/jackson/core/TreeNode.java b/src/main/java/com/fasterxml/jackson/core/TreeNode.java
index 1e4f659..8d8f8d7 100644
--- a/src/main/java/com/fasterxml/jackson/core/TreeNode.java
+++ b/src/main/java/com/fasterxml/jackson/core/TreeNode.java
@@ -261,6 +261,15 @@
* Functionally equivalent to first serializing tree using
* {@link ObjectCodec} and then re-parsing but
* more efficient.
+ *<p>
+ * NOTE: constructed parser instance will NOT initially point to a token,
+ * so before passing it to deserializers, it is typically necessary to
+ * advance it to the first available token by calling {@link JsonParser#nextToken()}.
+ *<p>
+ * Also note that calling this method will <b>NOT</b> pass {@link ObjectCodec}
+ * reference, so data-binding callback methods like {@link JsonParser#readValueAs(Class)}
+ * will not work with calling {@link JsonParser#setCodec}).
+ * It is often better to call {@link #traverse(ObjectCodec)} to pass the codec explicitly.
*/
JsonParser traverse();
@@ -268,6 +277,10 @@
* Same as {@link #traverse()}, but additionally passes {@link com.fasterxml.jackson.core.ObjectCodec}
* to use if {@link JsonParser#readValueAs(Class)} is used (otherwise caller must call
* {@link JsonParser#setCodec} on response explicitly).
+ *<p>
+ * NOTE: constructed parser instance will NOT initially point to a token,
+ * so before passing it to deserializers, it is typically necessary to
+ * advance it to the first available token by calling {@link JsonParser#nextToken()}.
*
* @since 2.1
*/