Turns out Neil didn't intend for *all* of his gen-branch work to get

committed.

tokenize.py:  I like these changes, and have tested them extensively
without even realizing it, so I just updated the docstring and the docs.

tabnanny.py:  Also liked this, but did a little code fiddling.  I should
really rewrite this to *exploit* generators, but that's near the bottom
of my effort/benefit scale so doubt I'll get to it anytime soon (it
would be most useful as a non-trivial example of ideal use of generators;
but test_generators.py has already grown plenty of food-for-thought
examples).

inspect.py:  I'm sure Ping intended for this to continue running even
under 1.5.2, so I reverted this to the last pre-gen-branch version.  The
"bugfix" I checked in in-between was actually repairing a bug *introduced*
by the conversion to generators, so it's OK that the reverted version
doesn't reflect that checkin.

4 files changed, 78 insertions, 46 deletions

diff --git a/Doc/lib/libtokenize.tex b/Doc/lib/libtokenize.tex
index 205407c..6cd9348 100644
--- a/Doc/lib/libtokenize.tex
+++ b/Doc/lib/libtokenize.tex
@@ -12,12 +12,33 @@ source code, implemented in Python.  The scanner in this module
 returns comments as tokens as well, making it useful for implementing
 ``pretty-printers,'' including colorizers for on-screen displays.
 
-The scanner is exposed by a single function:
+The primary entry point is a generator:
 
+\begin{funcdesc}{generate_tokens}{readline}
+  The \function{generate_tokens()} generator requires one argment,
+  \var{readline}, which must be a callable object which
+  provides the same interface as the \method{readline()} method of
+  built-in file objects (see section~\ref{bltin-file-objects}).  Each
+  call to the function should return one line of input as a string.
+
+  The generator produces 5-tuples with these members:
+  the token type;
+  the token string;
+  a 2-tuple \code{(\var{srow}, \var{scol})} of ints specifying the
+  row and column where the token begins in the source;
+  a 2-tuple \code{(\var{erow}, \var{ecol})} of ints specifying the
+  row and column where the token ends in the source;
+  and the line on which the token was found.
+  The line passed is the \emph{logical} line;
+  continuation lines are included.
+  \versionadded{2.2}
+\end{funcdesc}
+
+An older entry point is retained for backward compatibility:
 
 \begin{funcdesc}{tokenize}{readline\optional{, tokeneater}}
   The \function{tokenize()} function accepts two parameters: one
-  representing the input stream, and one providing an output mechanism 
+  representing the input stream, and one providing an output mechanism
   for \function{tokenize()}.
 
   The first parameter, \var{readline}, must be a callable object which
@@ -26,17 +47,13 @@ The scanner is exposed by a single function:
   call to the function should return one line of input as a string.
 
   The second parameter, \var{tokeneater}, must also be a callable
-  object.  It is called with five parameters: the token type, the
-  token string, a tuple \code{(\var{srow}, \var{scol})} specifying the 
-  row and column where the token begins in the source, a tuple
-  \code{(\var{erow}, \var{ecol})} giving the ending position of the
-  token, and the line on which the token was found.  The line passed
-  is the \emph{logical} line; continuation lines are included.
+  object.  It is called once for each token, with five arguments,
+  corresponding to the tuples generated by \function{generate_tokens()}.
 \end{funcdesc}
 
 
-All constants from the \refmodule{token} module are also exported from 
-\module{tokenize}, as are two additional token type values that might be 
+All constants from the \refmodule{token} module are also exported from
+\module{tokenize}, as are two additional token type values that might be
 passed to the \var{tokeneater} function by \function{tokenize()}:
 
 \begin{datadesc}{COMMENT}
diff --git a/Lib/inspect.py b/Lib/inspect.py
index eeb54d2..2d88bc1 100644
--- a/Lib/inspect.py
+++ b/Lib/inspect.py
@@ -349,28 +349,32 @@ class ListReader:
             return self.lines[i]
         else: return ''
 
-def getblock(lines):
-    """Extract the block of code at the top of the given list of lines."""
-
-    indent = 0
-    started = 0
-    last = 0
-    tokens = tokenize.generate_tokens(ListReader(lines).readline)
-
-    for (type, token, (srow, scol), (erow, ecol), line) in tokens:
-        if not started:
-            if type == tokenize.NAME:
-                started = 1
+class EndOfBlock(Exception): pass
+
+class BlockFinder:
+    """Provide a tokeneater() method to detect the end of a code block."""
+    def __init__(self):
+        self.indent = 0
+        self.started = 0
+        self.last = 0
+
+    def tokeneater(self, type, token, (srow, scol), (erow, ecol), line):
+        if not self.started:
+            if type == tokenize.NAME: self.started = 1
         elif type == tokenize.NEWLINE:
-            last = srow
+            self.last = srow
         elif type == tokenize.INDENT:
-            indent = indent + 1
+            self.indent = self.indent + 1
         elif type == tokenize.DEDENT:
-            indent = indent - 1
-            if indent == 0:
-                return lines[:last]
-    else:
-        raise ValueError, "unable to find block"
+            self.indent = self.indent - 1
+            if self.indent == 0: raise EndOfBlock, self.last
+
+def getblock(lines):
+    """Extract the block of code at the top of the given list of lines."""
+    try:
+        tokenize.tokenize(ListReader(lines).readline, BlockFinder().tokeneater)
+    except EndOfBlock, eob:
+        return lines[:eob.args[0]]
 
 def getsourcelines(object):
     """Return a list of source lines and starting line number for an object.
diff --git a/Lib/tabnanny.py b/Lib/tabnanny.py
index 8323a33..35e5cfa 100755
--- a/Lib/tabnanny.py
+++ b/Lib/tabnanny.py
@@ -14,6 +14,8 @@ import os
 import sys
 import getopt
 import tokenize
+if not hasattr(tokenize, 'NL'):
+    raise ValueError("tokenize.NL doesn't exist -- tokenize module too old")
 
 __all__ = ["check"]
 
@@ -243,15 +245,11 @@ def format_witnesses(w):
         prefix = prefix + "s"
     return prefix + " " + string.join(firsts, ', ')
 
-# Need Guido's enhancement
-assert hasattr(tokenize, 'NL'), "tokenize module too old"
-
-def process_tokens(tokens,
-                   INDENT=tokenize.INDENT,
-                   DEDENT=tokenize.DEDENT,
-                   NEWLINE=tokenize.NEWLINE,
-                   JUNK=(tokenize.COMMENT, tokenize.NL)):
-
+def process_tokens(tokens):
+    INDENT = tokenize.INDENT
+    DEDENT = tokenize.DEDENT
+    NEWLINE = tokenize.NEWLINE
+    JUNK = tokenize.COMMENT, tokenize.NL
     indents = [Whitespace("")]
     check_equal = 0
 
diff --git a/Lib/tokenize.py b/Lib/tokenize.py
index b3ee4a8..b79cdc0 100644
--- a/Lib/tokenize.py
+++ b/Lib/tokenize.py
@@ -1,13 +1,26 @@
 """Tokenization help for Python programs.
 
-This module exports a function called 'tokenize()' that breaks a stream of
+generate_tokens(readline) is a generator that breaks a stream of
 text into Python tokens.  It accepts a readline-like method which is called
-repeatedly to get the next line of input (or "" for EOF) and a "token-eater"
-function which is called once for each token found.  The latter function is
-passed the token type, a string containing the token, the starting and
-ending (row, column) coordinates of the token, and the original line.  It is
-designed to match the working of the Python tokenizer exactly, except that
-it produces COMMENT tokens for comments and gives type OP for all operators."""
+repeatedly to get the next line of input (or "" for EOF).  It generates
+5-tuples with these members:
+
+    the token type (see token.py)
+    the token (a string)
+    the starting (row, column) indices of the token (a 2-tuple of ints)
+    the ending (row, column) indices of the token (a 2-tuple of ints)
+    the original line (string)
+
+It is designed to match the working of the Python tokenizer exactly, except
+that it produces COMMENT tokens for comments and gives type OP for all
+operators
+
+Older entry points
+    tokenize_loop(readline, tokeneater)
+    tokenize(readline, tokeneater=printtoken)
+are the same, except instead of generating tokens, tokeneater is a callback
+function to which the 5 fields described above are passed as 5 arguments,
+each time a new token is found."""
 
 __author__ = 'Ka-Ping Yee <ping@lfw.org>'
 __credits__ = \
@@ -111,7 +124,7 @@ def tokenize(readline, tokeneater=printtoken):
     except StopTokenizing:
         pass
 
-# backwards compatible interface, probably not used
+# backwards compatible interface
 def tokenize_loop(readline, tokeneater):
     for token_info in generate_tokens(readline):
         apply(tokeneater, token_info)