[gyb] Use lambda function to work around PEP 3114

[PEP 3114](https://www.python.org/dev/peps/pep-3114/) renamed
`iterator.next()` (Python 2) to `iterator.__next__()` (Python 3). The
recommended solution to make the code work in both Python 2 and 3 is to
call the global `next` function.

To use this recommended global `next` function this patch uses a lambda
function to supply `tokenize.generate_tokens` a callable function as it was previously.

This should be functionally equivalent to the old code with the added
benefit of working on both Python 2 and 3.

Author: RLovelett

utils/gyb.py

@@ -304,7 +304,7 @@ def splitGybLines(sourceLines):
     dedents = 0
     try:
         for tokenKind, tokenText, tokenStart, (tokenEndLine, tokenEndCol), lineText \
-            in tokenize.generate_tokens(sourceLines.__iter__().next):
+            in tokenize.generate_tokens(lambda i = iter(sourceLines): next(i)):
 
             if tokenKind in (tokenize.COMMENT, tokenize.ENDMARKER): 
                 continue
@@ -347,7 +347,7 @@ def codeStartsWithDedentKeyword(sourceLines):
     """
     tokenText = None
     for tokenKind, tokenText, _, _, _ \
-        in tokenize.generate_tokens(sourceLines.__iter__().next):
+        in tokenize.generate_tokens(lambda i = iter(sourceLines): next(i)):
 
         if tokenKind != tokenize.COMMENT and tokenText.strip() != '':
             break