tests: Update misc/pep8.py to 1.4.5

Version in our tree is a wee bit outdated. For example,
later work will introduce an utf8 encoded python source
file which our old pep8.py does not yet understand (yeah,
it really was *that* ancient)...

Updated from:
https://raw.github.com/jcrocholl/pep8/1.4.5/pep8.py

Takes 552c1f1525e37a30376790151c1ba437776682c5,
f941537d1c0a40f0906490ed160db6c79af572d3,
5a4afe2a77d0ff7d9fea13dd93c3304a6ca993de and
a17f157e19bd6792c00321c8020dca5e5a281f45 into account...

https://bugzilla.gnome.org/show_bug.cgi?id=699535

26 files changed, 1765 insertions, 682 deletions

diff --git a/Makefile.am b/Makefile.am
index 202464fa..c04c05a0 100644
--- a/Makefile.am
+++ b/Makefile.am
@@ -55,10 +55,16 @@ EXTRA_DIST += 			\
 	misc/pep8.py		\
 	misc/pyflakes.py
 
+# Default pep8.py --exclude + emacs backup files
+PEP8_EXCLUDES=--exclude='.svn,CVS,.bzr,.hg,.git,__pycache__,.\#*'
+
 check-local:
 	gtester --verbose $(GTESTER_PROGS)
 	@echo "TEST: PEP-8 INQUISITION"
-	@find $(top_srcdir)/giscanner -name \*.py | sort | uniq | xargs $(PYTHON) $(top_srcdir)/misc/pep8.py --repeat --exclude=config.py
+	@find $(top_srcdir)/giscanner -name \*.py | sort | uniq | xargs \
+		$(PYTHON) $(top_srcdir)/misc/pep8.py --max-line-length=99 --ignore=E128 $(PEP8_EXCLUDES)
+	@find $(top_srcdir)/tests -name \*.py | sort | uniq | xargs \
+		$(PYTHON) $(top_srcdir)/misc/pep8.py --ignore=E127,E501 $(PEP8_EXCLUDES)
 	@echo "TEST: Annotation pattern programs"
 	PYTHONPATH="$(top_builddir):$(top_srcdir)" $(PYTHON) $(top_srcdir)/tests/scanner/annotationparser/test_patterns.py
 	@echo "TEST: GTK-Doc Annotation Parser"
diff --git a/giscanner/annotationmain.py b/giscanner/annotationmain.py
index 4df6e831..304f5a32 100644
--- a/giscanner/annotationmain.py
+++ b/giscanner/annotationmain.py
@@ -26,6 +26,7 @@ from giscanner.scannermain import (get_preprocessor_option_group,
                                    create_source_scanner,
                                    process_packages)
 
+
 def annotation_main(args):
     parser = optparse.OptionParser('%prog [options] sources')
 
diff --git a/giscanner/annotationparser.py b/giscanner/annotationparser.py
index c0b439a1..a0657dc4 100644
--- a/giscanner/annotationparser.py
+++ b/giscanner/annotationparser.py
@@ -141,7 +141,8 @@ OPT_TRANSFER_FLOATING = 'floating'
 # Program matching the start of a comment block.
 #
 # Results in 0 symbolic groups.
-COMMENT_START_RE = re.compile(r'''
+COMMENT_START_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     /                                        # 1 forward slash character
@@ -157,7 +158,8 @@ COMMENT_START_RE = re.compile(r'''
 #
 # Results in 1 symbolic group:
 #    - group 1 = description
-COMMENT_END_RE = re.compile(r'''
+COMMENT_END_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     (?P<description>.*?)                     # description text
@@ -173,7 +175,8 @@ COMMENT_END_RE = re.compile(r'''
 # line inside a comment block.
 #
 # Results in 0 symbolic groups.
-COMMENT_ASTERISK_RE = re.compile(r'''
+COMMENT_ASTERISK_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     \*                                       # 1 asterisk character
@@ -189,7 +192,8 @@ COMMENT_ASTERISK_RE = re.compile(r'''
 #
 # Results in 1 symbolic group:
 #   - group 1 = indentation
-COMMENT_INDENTATION_RE = re.compile(r'''
+COMMENT_INDENTATION_RE = re.compile(
+    r'''
     ^
     (?P<indentation>[^\S\n\r]*)              # 0 or more whitespace characters
     .*
@@ -200,7 +204,8 @@ COMMENT_INDENTATION_RE = re.compile(r'''
 # Program matching an empty line.
 #
 # Results in 0 symbolic groups.
-EMPTY_LINE_RE = re.compile(r'''
+EMPTY_LINE_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     $                                        # end
@@ -212,7 +217,8 @@ EMPTY_LINE_RE = re.compile(r'''
 # Results in 2 symbolic groups:
 #   - group 1 = colon
 #   - group 2 = section_name
-SECTION_RE = re.compile(r'''
+SECTION_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     SECTION                                  # SECTION
@@ -231,7 +237,8 @@ SECTION_RE = re.compile(r'''
 #   - group 1 = symbol_name
 #   - group 2 = colon
 #   - group 3 = annotations
-SYMBOL_RE = re.compile(r'''
+SYMBOL_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     (?P<symbol_name>[\w-]*\w)                # symbol name
@@ -251,7 +258,8 @@ SYMBOL_RE = re.compile(r'''
 #   - group 2 = property_name
 #   - group 3 = colon
 #   - group 4 = annotations
-PROPERTY_RE = re.compile(r'''
+PROPERTY_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     (?P<class_name>[\w]+)                    # class name
@@ -275,7 +283,8 @@ PROPERTY_RE = re.compile(r'''
 #   - group 2 = signal_name
 #   - group 3 = colon
 #   - group 4 = annotations
-SIGNAL_RE = re.compile(r'''
+SIGNAL_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     (?P<class_name>[\w]+)                    # class name
@@ -299,7 +308,8 @@ SIGNAL_RE = re.compile(r'''
 #   - group 2 = annotations
 #   - group 3 = colon
 #   - group 4 = description
-PARAMETER_RE = re.compile(r'''
+PARAMETER_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     @                                        # @ character
@@ -324,7 +334,8 @@ PARAMETER_RE = re.compile(r'''
 #   - group 3 = colon
 #   - group 4 = description
 _all_tags = '|'.join(_ALL_TAGS).replace(' ', '\\ ')
-TAG_RE = re.compile(r'''
+TAG_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     (?P<tag_name>''' + _all_tags + r''')     # tag name
@@ -348,7 +359,8 @@ TAG_RE = re.compile(r'''
 #   - group 2 = annotations
 #   - group 3 = colon
 #   - group 4 = description
-MULTILINE_ANNOTATION_CONTINUATION_RE = re.compile(r'''
+MULTILINE_ANNOTATION_CONTINUATION_RE = re.compile(
+    r'''
     ^                                        # start
     [^\S\n\r]*                               # 0 or more whitespace characters
     (?P<annotations>(?:\(.*?\)[^\S\n\r]*)*)  # annotations
@@ -444,8 +456,8 @@ class DocTag(object):
                 s = 'one value'
             else:
                 s = '%d values' % (n_params, )
-            if ((n_params > 0 and (value is None or value.length() != n_params)) or
-                n_params == 0 and value is not None):
+            if ((n_params > 0 and (value is None or value.length() != n_params))
+            or n_params == 0 and value is not None):
                 if value is None:
                     length = 0
                 else:
@@ -814,7 +826,7 @@ class AnnotationParser(object):
             if description:
                 comment_lines[-1] = (line_offset, description)
                 position = message.Position(filename, lineno + line_offset)
-                marker = ' '*result.end('description') + '^'
+                marker = ' ' * result.end('description') + '^'
                 message.warn("Comments should end with */ on a new line:\n%s\n%s" %
                              (line, marker),
                              position)
@@ -922,7 +934,7 @@ class AnnotationParser(object):
                     if 'colon' in result.groupdict() and result.group('colon') != ':':
                         colon_start = result.start('colon')
                         colon_column = column_offset + colon_start
-                        marker = ' '*colon_column + '^'
+                        marker = ' ' * colon_column + '^'
                         message.warn("missing ':' at column %s:\n%s\n%s" %
                                      (colon_column + 1, original_line, marker),
                                      position)
@@ -942,7 +954,7 @@ class AnnotationParser(object):
                     # right thing to do because sooner or later some long
                     # descriptions will contain something matching an identifier
                     # pattern by accident.
-                    marker = ' '*column_offset + '^'
+                    marker = ' ' * column_offset + '^'
                     message.warn('ignoring unrecognized GTK-Doc comment block, identifier not '
                                  'found:\n%s\n%s' % (original_line, marker),
                                  position)
@@ -965,7 +977,7 @@ class AnnotationParser(object):
 
                 if in_part != PART_PARAMETERS:
                     column = result.start('parameter_name') + column_offset
-                    marker = ' '*column + '^'
+                    marker = ' ' * column + '^'
                     message.warn("'@%s' parameter unexpected at this location:\n%s\n%s" %
                                  (param_name, original_line, marker),
                                  position)
@@ -983,7 +995,7 @@ class AnnotationParser(object):
                                      position)
                 elif param_name in comment_block.params.keys():
                     column = result.start('parameter_name') + column_offset
-                    marker = ' '*column + '^'
+                    marker = ' ' * column + '^'
                     message.warn("multiple '@%s' parameters for identifier '%s':\n%s\n%s" %
                                  (param_name, comment_block.name, original_line, marker),
                                  position)
@@ -1007,8 +1019,7 @@ class AnnotationParser(object):
             # identifier (when there are no parameters) and encounter an empty
             # line, we must be parsing the comment block description.
             ####################################################################
-            if (EMPTY_LINE_RE.match(line)
-            and in_part in [PART_IDENTIFIER, PART_PARAMETERS]):
+            if (EMPTY_LINE_RE.match(line) and in_part in [PART_IDENTIFIER, PART_PARAMETERS]):
                 in_part = PART_DESCRIPTION
                 part_indent = line_indent
                 continue
@@ -1022,7 +1033,7 @@ class AnnotationParser(object):
                 tag_annotations = result.group('annotations')
                 tag_description = result.group('description')
 
-                marker = ' '*(result.start('tag_name') + column_offset) + '^'
+                marker = ' ' * (result.start('tag_name') + column_offset) + '^'
 
                 # Deprecated GTK-Doc Description: tag
                 if tag_name.lower() == TAG_DESCRIPTION:
@@ -1047,7 +1058,7 @@ class AnnotationParser(object):
 
                 if in_part != PART_TAGS:
                     column = result.start('tag_name') + column_offset
-                    marker = ' '*column + '^'
+                    marker = ' ' * column + '^'
                     message.warn("'%s:' tag unexpected at this location:\n%s\n%s" %
                                  (tag_name, original_line, marker),
                                  position)
@@ -1071,7 +1082,7 @@ class AnnotationParser(object):
                 else:
                     if tag_name.lower() in comment_block.tags.keys():
                         column = result.start('tag_name') + column_offset
-                        marker = ' '*column + '^'
+                        marker = ' ' * column + '^'
                         message.warn("multiple '%s:' tags for identifier '%s':\n%s\n%s" %
                                      (tag_name, comment_block.name, original_line, marker),
                                      position)
@@ -1149,7 +1160,7 @@ class AnnotationParser(object):
             part.value = ''
 
     def _validate_multiline_annotation_continuation(self, line, original_line,
-                                                          column_offset, position):
+                                                    column_offset, position):
         '''
         Validate parameters and tags (except the first line) and generate
         warnings about invalid annotations spanning multiple lines.
@@ -1163,7 +1174,7 @@ class AnnotationParser(object):
         result = MULTILINE_ANNOTATION_CONTINUATION_RE.match(line)
         if result:
             column = result.start('annotations') + column_offset
-            marker = ' '*column + '^'
+            marker = ' ' * column + '^'
             message.warn('ignoring invalid multiline annotation continuation:\n'
                          '%s\n%s' % (original_line, marker),
                          position)
@@ -1179,7 +1190,7 @@ class AnnotationParser(object):
 
         for i, c in enumerate(value):
             if c == '(' and opened == -1:
-                opened = i+1
+                opened = i + 1
             if c == ')' and opened != -1:
                 segment = value[opened:i]
                 parts = segment.split(' ', 1)
diff --git a/giscanner/ast.py b/giscanner/ast.py
index 25184886..4c54b548 100644
--- a/giscanner/ast.py
+++ b/giscanner/ast.py
@@ -28,15 +28,17 @@ from .collections import OrderedDict
 from .message import Position
 from .utils import to_underscores
 
+
 class Type(object):
-    """A Type can be either:
-* A reference to a node (target_giname)
-* A reference to a "fundamental" type like 'utf8'
-* A "foreign" type - this can be any string."
-If none are specified, then it's in an "unresolved" state.  An
-unresolved type can have two data sources; a "ctype" which comes
-from a C type string, or a gtype_name (from g_type_name()).
-""" # '''
+    """
+    A Type can be either:
+    * A reference to a node (target_giname)
+    * A reference to a "fundamental" type like 'utf8'
+    * A "foreign" type - this can be any string."
+    If none are specified, then it's in an "unresolved" state.  An
+    unresolved type can have two data sources; a "ctype" which comes
+    from a C type string, or a gtype_name (from g_type_name()).
+    """
 
     def __init__(self,
                  ctype=None,
@@ -125,11 +127,12 @@ in contrast to the other create_type() functions."""
     def __cmp__(self, other):
         if self.target_fundamental:
             return cmp(self.target_fundamental, other.target_fundamental)
-        if self.target_giname:
+        elif self.target_giname:
             return cmp(self.target_giname, other.target_giname)
-        if self.target_foreign:
+        elif self.target_foreign:
             return cmp(self.target_foreign, other.target_foreign)
-        return cmp(self.ctype, other.ctype)
+        else:
+            return cmp(self.ctype, other.ctype)
 
     def is_equiv(self, typeval):
         """Return True if the specified types are compatible at
@@ -170,6 +173,7 @@ in contrast to the other create_type() functions."""
             data = ''
         return '%s(%sctype=%s)' % (self.__class__.__name__, data, self.ctype)
 
+
 class TypeUnknown(Type):
     def __init__(self):
         Type.__init__(self, _target_unknown=True)
@@ -351,9 +355,7 @@ SIGNAL_MUST_COLLECT = 'must-collect'
 
 
 class Namespace(object):
-    def __init__(self, name, version,
-                 identifier_prefixes=None,
-                 symbol_prefixes=None):
+    def __init__(self, name, version, identifier_prefixes=None, symbol_prefixes=None):
         self.name = name
         self.version = version
         if identifier_prefixes is not None:
@@ -367,15 +369,15 @@ class Namespace(object):
             self.symbol_prefixes = [to_underscores(p).lower() for p in ps]
         # cache upper-cased versions
         self._ucase_symbol_prefixes = [p.upper() for p in self.symbol_prefixes]
-        self.names = OrderedDict() # Maps from GIName -> node
-        self.aliases = {} # Maps from GIName -> GIName
-        self.type_names = {} # Maps from GTName -> node
-        self.ctypes = {} # Maps from CType -> node
-        self.symbols = {} # Maps from function symbols -> Function
-        self.includes = set() # Include
-        self.shared_libraries = [] # str
-        self.c_includes = [] # str
-        self.exported_packages = [] # str
+        self.names = OrderedDict()   # Maps from GIName -> node
+        self.aliases = {}            # Maps from GIName -> GIName
+        self.type_names = {}         # Maps from GTName -> node
+        self.ctypes = {}             # Maps from CType -> node
+        self.symbols = {}            # Maps from function symbols -> Function
+        self.includes = set()        # Include
+        self.shared_libraries = []   # str
+        self.c_includes = []         # str
+        self.exported_packages = []  # str
 
     def type_from_name(self, name, ctype=None):
         """Backwards compatibility method for older .gir files, which
@@ -416,7 +418,7 @@ but adds it to things like ctypes, symbols, and type_names.
         if isinstance(node, (Class, Interface)):
             for m in chain(node.signals, node.properties):
                 m.namespace = self
-        if isinstance(node, Enum) or isinstance(node, Bitfield):
+        if isinstance(node, (Enum, Bitfield)):
             for fn in node.static_methods:
                 if not isinstance(fn, Function):
                     continue
@@ -482,6 +484,7 @@ functions via get_by_symbol()."""
         for node in self.itervalues():
             node.walk(callback, [])
 
+
 class Include(object):
 
     def __init__(self, name, version):
@@ -504,6 +507,7 @@ class Include(object):
     def __str__(self):
         return '%s-%s' % (self.name, self.version)
 
+
 class Annotated(object):
     """An object which has a few generic metadata
 properties."""
@@ -511,12 +515,13 @@ properties."""
         self.version = None
         self.skip = False
         self.introspectable = True
-        self.attributes = [] # (key, value)*
+        self.attributes = []    # (key, value)*
         self.stability = None
         self.deprecated = None
         self.deprecated_version = None
         self.doc = None
 
+
 class Node(Annotated):
     """A node is a type of object which is uniquely identified by its
 (namespace, name) pair.  When combined with a ., this is called a
@@ -527,7 +532,7 @@ GIName.  It's possible for nodes to contain or point to other nodes."""
 
     def __init__(self, name=None):
         Annotated.__init__(self)
-        self.namespace = None # Should be set later by Namespace.append()
+        self.namespace = None   # Should be set later by Namespace.append()
         self.name = name
         self.foreign = False
         self.file_positions = set()
@@ -596,8 +601,8 @@ class Callable(Node):
         self.retval = retval
         self.parameters = parameters
         self.throws = not not throws
-        self.instance_parameter = None # Parameter
-        self.parent = None # A Class or Interface
+        self.instance_parameter = None  # Parameter
+        self.parent = None  # A Class or Interface
 
     # Returns all parameters, including the instance parameter
     @property
@@ -627,10 +632,10 @@ class Function(Callable):
         self.symbol = symbol
         self.is_method = False
         self.is_constructor = False
-        self.shadowed_by = None # C symbol string
-        self.shadows = None # C symbol string
-        self.moved_to = None # namespaced function name string
-        self.internal_skipped = False # if True, this func will not be written to GIR
+        self.shadowed_by = None         # C symbol string
+        self.shadows = None             # C symbol string
+        self.moved_to = None            # namespaced function name string
+        self.internal_skipped = False   # if True, this func will not be written to GIR
 
     def clone(self):
         clone = copy.copy(self)
@@ -641,8 +646,7 @@ class Function(Callable):
 
     def is_type_meta_function(self):
         # Named correctly
-        if not (self.name.endswith('_get_type') or
-                self.name.endswith('_get_gtype')):
+        if not (self.name.endswith('_get_type') or self.name.endswith('_get_gtype')):
             return False
 
         # Doesn't have any parameters
@@ -651,14 +655,13 @@ class Function(Callable):
 
         # Returns GType
         rettype = self.retval.type
-        if (not rettype.is_equiv(TYPE_GTYPE) and
-           rettype.target_giname != 'Gtk.Type'):
-            message.warn("function '%s' returns '%r', not a GType" %
-                         (self.name, rettype))
+        if (not rettype.is_equiv(TYPE_GTYPE) and rettype.target_giname != 'Gtk.Type'):
+            message.warn("function '%s' returns '%r', not a GType" % (self.name, rettype))
             return False
 
         return True
 
+
 class ErrorQuarkFunction(Function):
 
     def __init__(self, name, retval, parameters, throws, symbol, error_domain):
@@ -679,7 +682,6 @@ class VFunction(Callable):
         return obj
 
 
-
 class Varargs(Type):
 
     def __init__(self):
@@ -715,6 +717,7 @@ class Array(Type):
         arr.size = self.size
         return arr
 
+
 class List(Type):
 
     def __init__(self, name, element_type, **kwargs):
@@ -727,6 +730,7 @@ class List(Type):
     def clone(self):
         return List(self.name, self.element_type)
 
+
 class Map(Type):
 
     def __init__(self, key_type, value_type, **kwargs):
@@ -739,6 +743,7 @@ class Map(Type):
     def clone(self):
         return Map(self.key_type, self.value_type)
 
+
 class Alias(Node):
 
     def __init__(self, name, target, ctype=None):
@@ -881,6 +886,7 @@ class Compound(Node, Registered):
             if field.anonymous_node is not None:
                 field.anonymous_node.walk(callback, chain)
 
+
 class Field(Annotated):
 
     def __init__(self, name, typenode, readable, writable, bits=None,
@@ -894,7 +900,7 @@ class Field(Annotated):
         self.bits = bits
         self.anonymous_node = anonymous_node
         self.private = False
-        self.parent = None # a compound
+        self.parent = None  # a compound
 
     def __cmp__(self, other):
         return cmp(self.name, other.name)
@@ -1090,7 +1096,7 @@ class Property(Node):
             self.transfer = PARAM_TRANSFER_NONE
         else:
             self.transfer = transfer
-        self.parent = None # A Class or Interface
+        self.parent = None  # A Class or Interface
 
 
 class Callback(Callable):
diff --git a/giscanner/cachestore.py b/giscanner/cachestore.py
index 5f66b66a..ad4c7a36 100644
--- a/giscanner/cachestore.py
+++ b/giscanner/cachestore.py
@@ -31,6 +31,7 @@ import giscanner
 
 _CACHE_VERSION_FILENAME = '.cache-version'
 
+
 def _get_versionhash():
     toplevel = os.path.dirname(giscanner.__file__)
     # Use pyc instead of py to avoid extra IO
@@ -40,6 +41,7 @@ def _get_versionhash():
     mtimes = (str(os.stat(source).st_mtime) for source in sources)
     return hashlib.sha1(''.join(mtimes)).hexdigest()
 
+
 def _get_cachedir():
     if 'GI_SCANNER_DISABLE_CACHE' in os.environ:
         return None
@@ -150,8 +152,7 @@ class CacheStore(object):
         if store_filename is None:
             return
 
-        if (os.path.exists(store_filename) and
-            self._cache_is_valid(store_filename, filename)):
+        if (os.path.exists(store_filename) and self._cache_is_valid(store_filename, filename)):
             return None
 
         tmp_fd, tmp_filename = tempfile.mkstemp(prefix='g-ir-scanner-cache-')
diff --git a/giscanner/codegen.py b/giscanner/codegen.py
index b73a7da3..e9ed9415 100644
--- a/giscanner/codegen.py
+++ b/giscanner/codegen.py
@@ -24,6 +24,7 @@ from contextlib import contextmanager
 
 from . import ast
 
+
 class CCodeGenerator(object):
     def __init__(self, namespace, out_h_filename, out_c_filename):
         self.out_h_filename = out_h_filename
@@ -36,15 +37,16 @@ class CCodeGenerator(object):
         return '%s_%s' % (self.namespace.symbol_prefixes[0], name)
 
     def _typecontainer_to_ctype(self, param):
-        if (isinstance(param, ast.Parameter) and
-            param.direction in (ast.PARAM_DIRECTION_OUT,
-                                ast.PARAM_DIRECTION_INOUT)):
+        if (isinstance(param, ast.Parameter)
+        and param.direction in (ast.PARAM_DIRECTION_OUT, ast.PARAM_DIRECTION_INOUT)):
             suffix = '*'
         else:
             suffix = ''
-        if (param.type.is_equiv((ast.TYPE_STRING, ast.TYPE_FILENAME)) and
-            param.transfer == ast.PARAM_TRANSFER_NONE):
+
+        if (param.type.is_equiv((ast.TYPE_STRING, ast.TYPE_FILENAME))
+        and param.transfer == ast.PARAM_TRANSFER_NONE):
             return "const gchar*" + suffix
+
         return param.type.ctype + suffix
 
     def _write_prelude(self, out, func):
diff --git a/giscanner/docmain.py b/giscanner/docmain.py
index afd509ff..3287dcc6 100644
--- a/giscanner/docmain.py
+++ b/giscanner/docmain.py
@@ -24,6 +24,7 @@ import optparse
 from .docwriter import DocWriter
 from .transformer import Transformer
 
+
 def doc_main(args):
     parser = optparse.OptionParser('%prog [options] GIR-file')
 
diff --git a/giscanner/docwriter.py b/giscanner/docwriter.py
index b3716ec6..982ab37c 100644
--- a/giscanner/docwriter.py
+++ b/giscanner/docwriter.py
@@ -31,6 +31,7 @@ from mako.lookup import TemplateLookup
 from . import ast, xmlwriter
 from .utils import to_underscores
 
+
 def make_page_id(node, recursive=False):
     if isinstance(node, ast.Namespace):
         if recursive:
@@ -51,6 +52,7 @@ def make_page_id(node, recursive=False):
     else:
         return '%s.%s' % (make_page_id(parent, recursive=True), node.name)
 
+
 def get_node_kind(node):
     if isinstance(node, ast.Namespace):
         node_kind = 'namespace'
@@ -78,6 +80,7 @@ def get_node_kind(node):
 
     return node_kind
 
+
 class TemplatedScanner(object):
     def __init__(self, specs):
         self.specs = self.unmangle_specs(specs)
@@ -141,6 +144,7 @@ class TemplatedScanner(object):
         if pos < len(text):
             yield ('other', text[pos:], None)
 
+
 class DocstringScanner(TemplatedScanner):
     def __init__(self):
         specs = [
@@ -156,6 +160,7 @@ class DocstringScanner(TemplatedScanner):
 
         super(DocstringScanner, self).__init__(specs)
 
+
 class DocFormatter(object):
     def __init__(self, transformer):
         self._transformer = transformer
@@ -358,6 +363,7 @@ class DocFormatter(object):
         parent_chain.reverse()
         return parent_chain
 
+
 class DocFormatterC(DocFormatter):
     language = "C"
     mime_type = "text/x-csrc"
@@ -380,7 +386,7 @@ class DocFormatterC(DocFormatter):
             return getattr(node, 'ctype')
 
     def format_function_name(self, func):
-        if isinstance(func, (ast.Function)):
+        if isinstance(func, ast.Function):
             return func.symbol
         else:
             return func.name
@@ -388,6 +394,7 @@ class DocFormatterC(DocFormatter):
     def get_parameters(self, node):
         return node.all_parameters
 
+
 class DocFormatterIntrospectableBase(DocFormatter):
     def should_render_node(self, node):
         if isinstance(node, ast.Record) and node.is_gtype_struct_for is not None:
@@ -398,6 +405,7 @@ class DocFormatterIntrospectableBase(DocFormatter):
 
         return super(DocFormatterIntrospectableBase, self).should_render_node(node)
 
+
 class DocFormatterPython(DocFormatterIntrospectableBase):
     language = "Python"
     mime_type = "text/python"
@@ -418,7 +426,7 @@ class DocFormatterPython(DocFormatterIntrospectableBase):
         if getattr(node, "is_method", False):
             return True
 
-        if isinstance(node, (ast.VFunction)):
+        if isinstance(node, ast.VFunction):
             return True
 
         return False
@@ -451,8 +459,7 @@ class DocFormatterPython(DocFormatterIntrospectableBase):
             "GParam": "GLib.Param",
             "PyObject": "object",
             "GStrv": "[str]",
-            "GVariant": "GLib.Variant",
-            }
+            "GVariant": "GLib.Variant"}
 
         return fundamental_types.get(name, name)
 
@@ -476,6 +483,7 @@ class DocFormatterPython(DocFormatterIntrospectableBase):
     def get_parameters(self, node):
         return node.all_parameters
 
+
 class DocFormatterGjs(DocFormatterIntrospectableBase):
     language = "Gjs"
     mime_type = "text/x-gjs"
@@ -490,7 +498,7 @@ class DocFormatterGjs(DocFormatterIntrospectableBase):
         if getattr(node, "is_method", False):
             return True
 
-        if isinstance(node, (ast.VFunction)):
+        if isinstance(node, ast.VFunction):
             return True
 
         return False
@@ -514,8 +522,7 @@ class DocFormatterGjs(DocFormatterIntrospectableBase):
             "GParam": "GLib.Param",
             "PyObject": "Object",
             "GStrv": "[String]",
-            "GVariant": "GLib.Variant",
-            }
+            "GVariant": "GLib.Variant"}
 
         return fundamental_types.get(name, name)
 
@@ -558,12 +565,14 @@ class DocFormatterGjs(DocFormatterIntrospectableBase):
                 params.append(param)
         return params
 
+
 LANGUAGES = {
     "c": DocFormatterC,
     "python": DocFormatterPython,
     "gjs": DocFormatterGjs,
 }
 
+
 class DocWriter(object):
     def __init__(self, transformer, language):
         self._transformer = transformer
diff --git a/giscanner/dumper.py b/giscanner/dumper.py
index b9f529ea..157b24da 100644
--- a/giscanner/dumper.py
+++ b/giscanner/dumper.py
@@ -223,7 +223,7 @@ class DumpCompiler(object):
         # The Microsoft compiler uses different option flags for
         # compilation result output
         if self._pkgconfig_msvc_flags:
-            args.extend(['-c', '-Fe'+output, '-Fo'+output])
+            args.extend(['-c', '-Fe' + output, '-Fo' + output])
         else:
             args.extend(['-c', '-o', output])
         for source in sources:
@@ -254,7 +254,7 @@ class DumpCompiler(object):
         # We can use -o for the Microsoft compiler/linker,
         # but it is considered deprecated usage with that
         if self._pkgconfig_msvc_flags:
-            args.extend(['-Fe'+output])
+            args.extend(['-Fe' + output])
         else:
             args.extend(['-o', output])
         if libtool:
@@ -313,9 +313,9 @@ class DumpCompiler(object):
                 if self._options.namespace_version:
                     args.append(str.lower(self._options.namespace_name) +
                                 '-' +
-                                self._options.namespace_version+'.lib')
+                                self._options.namespace_version + '.lib')
                 else:
-                    args.append(str.lower(self._options.namespace_name)+'.lib')
+                    args.append(str.lower(self._options.namespace_name) + '.lib')
             else:
                 args.append('-Wl,-rpath=.')
 
@@ -329,7 +329,7 @@ class DumpCompiler(object):
             # to .lib files, not the .dll as the --library option specifies the
             # .dll(s) the .gir file refers to
             if self._pkgconfig_msvc_flags == '':
-                if library.endswith(".la"): # explicitly specified libtool library
+                if library.endswith(".la"):  # explicitly specified libtool library
                     args.append(library)
                 else:
                     args.append('-l' + library)
@@ -358,11 +358,12 @@ class DumpCompiler(object):
             # The --library option on Windows pass in the .dll file(s) the
             # .gir files refer to, so don't link to them on Visual C++
             if self._pkgconfig_msvc_flags == '':
-                if library.endswith(".la"): # explicitly specified libtool library
+                if library.endswith(".la"):  # explicitly specified libtool library
                     args.append(library)
                 else:
                     args.append('-l' + library)
 
+
 def compile_introspection_binary(options, get_type_functions,
                                  error_quark_functions):
     dc = DumpCompiler(options, get_type_functions, error_quark_functions)
diff --git a/giscanner/gdumpparser.py b/giscanner/gdumpparser.py
index 79525030..568777bd 100644
--- a/giscanner/gdumpparser.py
+++ b/giscanner/gdumpparser.py
@@ -203,8 +203,7 @@ blob containing data gleaned from GObject's primitive introspection."""
 
     def _initparse_gobject_record(self, record):
         if (record.name.startswith('ParamSpec')
-              and not record.name in ('ParamSpecPool', 'ParamSpecClass',
-                                      'ParamSpecTypeInfo')):
+        and not record.name in ('ParamSpecPool', 'ParamSpecClass', 'ParamSpecTypeInfo')):
             parent = None
             if record.name != 'ParamSpec':
                 parent = ast.Type(target_giname='GObject.ParamSpec')
@@ -280,7 +279,6 @@ blob containing data gleaned from GObject's primitive introspection."""
                                       member.attrib['name'],
                                       member.attrib['nick']))
 
-
         if xmlnode.tag == 'flags':
             klass = ast.Bitfield
         else:
@@ -437,7 +435,7 @@ different --identifier-prefix.""" % (xmlnode.attrib['name'], self._namespace.ide
                 if i == 0:
                     argname = 'object'
                 else:
-                    argname = 'p%s' % (i-1, )
+                    argname = 'p%s' % (i - 1, )
                 pctype = parameter.attrib['type']
                 ptype = ast.Type.create_from_gtype_name(pctype)
                 param = ast.Parameter(argname, ptype)
@@ -526,8 +524,7 @@ different --identifier-prefix.""" % (xmlnode.attrib['name'], self._namespace.ide
             return False
 
     def _strip_class_suffix(self, name):
-        if (name.endswith('Class') or
-            name.endswith('Iface')):
+        if (name.endswith('Class') or name.endswith('Iface')):
             return name[:-5]
         elif name.endswith('Interface'):
             return name[:-9]
diff --git a/giscanner/girparser.py b/giscanner/girparser.py
index 63a3fd02..2538036a 100644
--- a/giscanner/girparser.py
+++ b/giscanner/girparser.py
@@ -101,9 +101,8 @@ class GIRParser(object):
         assert root.tag == _corens('repository')
         version = root.attrib['version']
         if version != COMPATIBLE_GIR_VERSION:
-            raise SystemExit("%s: Incompatible version %s (supported: %s)" \
-                             % (self._get_current_file(),
-                                version, COMPATIBLE_GIR_VERSION))
+            raise SystemExit("%s: Incompatible version %s (supported: %s)" %
+                             (self._get_current_file(), version, COMPATIBLE_GIR_VERSION))
 
         for node in root.getchildren():
             if node.tag == _corens('include'):
@@ -122,9 +121,9 @@ class GIRParser(object):
         if symbol_prefixes:
             symbol_prefixes = symbol_prefixes.split(',')
         self._namespace = ast.Namespace(ns.attrib['name'],
-                                    ns.attrib['version'],
-                                    identifier_prefixes=identifier_prefixes,
-                                    symbol_prefixes=symbol_prefixes)
+                                        ns.attrib['version'],
+                                        identifier_prefixes=identifier_prefixes,
+                                        symbol_prefixes=symbol_prefixes)
         if 'shared-library' in ns.attrib:
             self._namespace.shared_libraries = ns.attrib['shared-library'].split(',')
         self._namespace.includes = self._includes
@@ -140,8 +139,7 @@ class GIRParser(object):
             _corens('interface'): self._parse_object_interface,
             _corens('record'): self._parse_record,
             _corens('union'): self._parse_union,
-            _glibns('boxed'): self._parse_boxed,
-            }
+            _glibns('boxed'): self._parse_boxed}
 
         if not self._types_only:
             parser_methods[_corens('constant')] = self._parse_constant
@@ -153,8 +151,7 @@ class GIRParser(object):
                 method(node)
 
     def _parse_include(self, node):
-        include = ast.Include(node.attrib['name'],
-                          node.attrib['version'])
+        include = ast.Include(node.attrib['name'], node.attrib['version'])
         self._includes.add(include)
 
     def _parse_pkgconfig_package(self, node):
@@ -165,9 +162,7 @@ class GIRParser(object):
 
     def _parse_alias(self, node):
         typeval = self._parse_type(node)
-        alias = ast.Alias(node.attrib['name'],
-                      typeval,
-                      node.attrib.get(_cns('type')))
+        alias = ast.Alias(node.attrib['name'], typeval, node.attrib.get(_cns('type')))
         self._parse_generic_attribs(node, alias)
         self._namespace.append(alias)
 
@@ -427,7 +422,8 @@ class GIRParser(object):
                 return ast.Type(ctype=ctype)
             elif name in ['GLib.List', 'GLib.SList']:
                 subchild = self._find_first_child(typenode,
-                               map(_corens, ('callback', 'array', 'varargs', 'type')))
+                                                  map(_corens, ('callback', 'array',
+                                                                'varargs', 'type')))
                 if subchild is not None:
                     element_type = self._parse_type(typenode)
                 else:
@@ -438,9 +434,7 @@ class GIRParser(object):
                 subchildren_types = map(self._parse_type_simple, subchildren)
                 while len(subchildren_types) < 2:
                     subchildren_types.append(ast.TYPE_ANY)
-                return ast.Map(subchildren_types[0],
-                           subchildren_types[1],
-                           ctype=ctype)
+                return ast.Map(subchildren_types[0], subchildren_types[1], ctype=ctype)
             else:
                 return self._namespace.type_from_name(name, ctype)
         else:
@@ -462,8 +456,8 @@ class GIRParser(object):
         lenidx = typenode.attrib.get('length')
         if lenidx is not None:
             idx = int(lenidx)
-            assert idx < len(parent.parameters), "%r %d >= %d" \
-                      % (parent, idx, len(parent.parameters))
+            assert idx < len(parent.parameters), "%r %d >= %d" % (parent, idx,
+                                                                  len(parent.parameters))
             typeval.length_param_name = parent.parameters[idx].argname
 
     def _parse_boxed(self, node):
@@ -509,11 +503,11 @@ class GIRParser(object):
             assert node.tag == _corens('field'), node.tag
             type_node = self._parse_type(node)
         field = ast.Field(node.attrib.get('name'),
-                      type_node,
-                      node.attrib.get('readable') != '0',
-                      node.attrib.get('writable') == '1',
-                      node.attrib.get('bits'),
-                      anonymous_node=anonymous_node)
+                          type_node,
+                          node.attrib.get('readable') != '0',
+                          node.attrib.get('writable') == '1',
+                          node.attrib.get('bits'),
+                          anonymous_node=anonymous_node)
         field.private = node.attrib.get('private') == '1'
         field.parent = parent
         self._parse_generic_attribs(node, field)
@@ -521,12 +515,12 @@ class GIRParser(object):
 
     def _parse_property(self, node, parent):
         prop = ast.Property(node.attrib['name'],
-                        self._parse_type(node),
-                        node.attrib.get('readable') != '0',
-                        node.attrib.get('writable') == '1',
-                        node.attrib.get('construct') == '1',
-                        node.attrib.get('construct-only') == '1',
-                        node.attrib.get('transfer-ownership'))
+                            self._parse_type(node),
+                            node.attrib.get('readable') != '0',
+                            node.attrib.get('writable') == '1',
+                            node.attrib.get('construct') == '1',
+                            node.attrib.get('construct-only') == '1',
+                            node.attrib.get('transfer-ownership'))
         self._parse_generic_attribs(node, prop)
         prop.parent = parent
         return prop
diff --git a/giscanner/girwriter.py b/giscanner/girwriter.py
index 2578a331..e7af2533 100644
--- a/giscanner/girwriter.py
+++ b/giscanner/girwriter.py
@@ -28,14 +28,15 @@ from .xmlwriter import XMLWriter
 # Compatible changes we just make inline
 COMPATIBLE_GIR_VERSION = '1.2'
 
+
 class GIRWriter(XMLWriter):
 
     def __init__(self, namespace):
         super(GIRWriter, self).__init__()
         self.write_comment(
-'''This file was automatically generated from C sources - DO NOT EDIT!
-To affect the contents of this file, edit the original C definitions,
-and/or use gtk-doc annotations. ''')
+            'This file was automatically generated from C sources - DO NOT EDIT!\n'
+            'To affect the contents of this file, edit the original C definitions,\n'
+            'and/or use gtk-doc annotations. ')
         self._write_repository(namespace)
 
     def _write_repository(self, namespace):
@@ -43,8 +44,7 @@ and/or use gtk-doc annotations. ''')
             ('version', COMPATIBLE_GIR_VERSION),
             ('xmlns', 'http://www.gtk.org/introspection/core/1.0'),
             ('xmlns:c', 'http://www.gtk.org/introspection/c/1.0'),
-            ('xmlns:glib', 'http://www.gtk.org/introspection/glib/1.0'),
-            ]
+            ('xmlns:glib', 'http://www.gtk.org/introspection/glib/1.0')]
         with self.tagcontext('repository', attrs):
             for include in sorted(namespace.includes):
                 self._write_include(include)
@@ -292,8 +292,8 @@ and/or use gtk-doc annotations. ''')
                 attrs.append(('fixed-size', '%d' % (ntype.size, )))
             if ntype.length_param_name is not None:
                 assert function
-                attrs.insert(0, ('length', '%d'
-                            % (function.get_parameter_index(ntype.length_param_name, ))))
+                length = function.get_parameter_index(ntype.length_param_name)
+                attrs.insert(0, ('length', '%d' % (length, )))
 
             with self.tagcontext('array', attrs):
                 self._write_type(ntype.element_type)
@@ -482,7 +482,7 @@ and/or use gtk-doc annotations. ''')
         attrs = list(extra_attrs)
         if record.name is not None:
             attrs.append(('name', record.name))
-        if record.ctype is not None: # the record might be anonymous
+        if record.ctype is not None:  # the record might be anonymous
             attrs.append(('c:type', record.ctype))
         if record.disguised:
             attrs.append(('disguised', '1'))
@@ -513,7 +513,7 @@ and/or use gtk-doc annotations. ''')
         attrs = []
         if union.name is not None:
             attrs.append(('name', union.name))
-        if union.ctype is not None: # the union might be anonymous
+        if union.ctype is not None:  # the union might be anonymous
             attrs.append(('c:type', union.ctype))
         self._append_version(union, attrs)
         self._append_node_generic(union, attrs)
@@ -544,8 +544,7 @@ and/or use gtk-doc annotations. ''')
             elif isinstance(field.anonymous_node, ast.Union):
                 self._write_union(field.anonymous_node)
             else:
-                raise AssertionError("Unknown field anonymous: %r" \
-                                         % (field.anonymous_node, ))
+                raise AssertionError("Unknown field anonymous: %r" % (field.anonymous_node, ))
         else:
             attrs = [('name', field.name)]
             self._append_node_generic(field, attrs)
diff --git a/giscanner/introspectablepass.py b/giscanner/introspectablepass.py
index 3e295391..3d67c73e 100644
--- a/giscanner/introspectablepass.py
+++ b/giscanner/introspectablepass.py
@@ -21,6 +21,7 @@ from . import ast
 from . import message
 from .annotationparser import TAG_RETURNS
 
+
 class IntrospectablePass(object):
 
     def __init__(self, transformer, blocks):
@@ -79,7 +80,7 @@ class IntrospectablePass(object):
 
         if not node.type.resolved:
             self._parameter_warning(parent, node,
-"Unresolved type: %r" % (node.type.unresolved_string, ))
+                                    "Unresolved type: %r" % (node.type.unresolved_string, ))
             parent.introspectable = False
             return
 
@@ -88,22 +89,23 @@ class IntrospectablePass(object):
             return
 
         if (isinstance(node.type, (ast.List, ast.Array))
-            and node.type.element_type == ast.TYPE_ANY):
+        and node.type.element_type == ast.TYPE_ANY):
             self._parameter_warning(parent, node, "Missing (element-type) annotation")
             parent.introspectable = False
             return
 
         if (is_parameter
-            and isinstance(target, ast.Callback)
-            and not node.type.target_giname in ('GLib.DestroyNotify',
-                                                'Gio.AsyncReadyCallback')
-            and node.scope is None):
-                self._parameter_warning(parent, node,
-                    ("Missing (scope) annotation for callback" +
-                     " without GDestroyNotify (valid: %s, %s)")
-                     % (ast.PARAM_SCOPE_CALL, ast.PARAM_SCOPE_ASYNC))
-                parent.introspectable = False
-                return
+        and isinstance(target, ast.Callback)
+        and not node.type.target_giname in ('GLib.DestroyNotify', 'Gio.AsyncReadyCallback')
+        and node.scope is None):
+            self._parameter_warning(
+                parent,
+                node,
+                "Missing (scope) annotation for callback without "
+                "GDestroyNotify (valid: %s, %s)" % (ast.PARAM_SCOPE_CALL, ast.PARAM_SCOPE_ASYNC))
+
+            parent.introspectable = False
+            return
 
         if is_return and isinstance(target, ast.Callback):
             self._parameter_warning(parent, node, "Callbacks cannot be return values; use (skip)")
@@ -111,12 +113,14 @@ class IntrospectablePass(object):
             return
 
         if (is_return
-            and isinstance(target, (ast.Record, ast.Union))
-            and target.get_type is None
-            and not target.foreign):
+        and isinstance(target, (ast.Record, ast.Union))
+        and target.get_type is None
+        and not target.foreign):
             if node.transfer != ast.PARAM_TRANSFER_NONE:
-                self._parameter_warning(parent, node,
-"Invalid non-constant return of bare structure or union; register as boxed type or (skip)")
+                self._parameter_warning(
+                    parent, node,
+                    "Invalid non-constant return of bare structure or union; "
+                    "register as boxed type or (skip)")
                 parent.introspectable = False
             return
 
@@ -143,10 +147,10 @@ class IntrospectablePass(object):
             # These are not introspectable pending us adding
             # larger type tags to the typelib (in theory these could
             # be 128 bit or larger)
-            if typeval.is_equiv((ast.TYPE_LONG_LONG, ast.TYPE_LONG_ULONG,
-                                 ast.TYPE_LONG_DOUBLE)):
+            elif typeval.is_equiv((ast.TYPE_LONG_LONG, ast.TYPE_LONG_ULONG, ast.TYPE_LONG_DOUBLE)):
                 return False
-            return True
+            else:
+                return True
         target = self._transformer.lookup_typenode(typeval)
         if not target:
             return False
diff --git a/giscanner/maintransformer.py b/giscanner/maintransformer.py
index 8292f8eb..6acc5947 100644
--- a/giscanner/maintransformer.py
+++ b/giscanner/maintransformer.py
@@ -37,6 +37,7 @@ from .annotationparser import (OPT_ALLOW_NONE, OPT_ARRAY, OPT_ATTRIBUTE,
                                OPT_TRANSFER_NONE, OPT_TRANSFER_FLOATING)
 from .utils import to_underscores_noprefix
 
+
 class MainTransformer(object):
 
     def __init__(self, transformer, blocks):
@@ -105,7 +106,7 @@ class MainTransformer(object):
 
     def _pass_fixup_hidden_fields(self, node, chain):
         """Hide all callbacks starting with _; the typical
-usage is void (*_gtk_reserved1)(void);"""
+        usage is void (*_gtk_reserved1)(void);"""
         if isinstance(node, (ast.Class, ast.Interface, ast.Record, ast.Union)):
             for field in node.fields:
                 if (field
@@ -204,7 +205,7 @@ usage is void (*_gtk_reserved1)(void);"""
         if isinstance(node, ast.Function):
             self._apply_annotations_function(node, chain)
         if isinstance(node, ast.Callback):
-            self._apply_annotations_callable(node, chain, block = self._get_block(node))
+            self._apply_annotations_callable(node, chain, block=self._get_block(node))
         if isinstance(node, (ast.Class, ast.Interface, ast.Union, ast.Enum,
                              ast.Bitfield, ast.Callback)):
             self._apply_annotations_annotated(node, self._get_block(node))
@@ -257,7 +258,7 @@ usage is void (*_gtk_reserved1)(void);"""
             Use resolver() on each identifier, and combiner() on the parts of
             each complete type. (top_combiner is used on the top-most type.)"""
             bits = re.split(r'([,<>()])', type_str, 1)
-            first, sep, rest = [bits[0], '', ''] if (len(bits)==1) else bits
+            first, sep, rest = [bits[0], '', ''] if (len(bits) == 1) else bits
             args = [resolver(first)]
             if sep == '<' or sep == '(':
                 lastsep = '>' if (sep == '<') else ')'
@@ -268,9 +269,11 @@ usage is void (*_gtk_reserved1)(void);"""
             else:
                 rest = sep + rest
             return top_combiner(*args), rest
+
         def resolver(ident):
             res = self._transformer.create_type_from_user_string(ident)
             return res
+
         def combiner(base, *rest):
             if not rest:
                 return base
@@ -281,6 +284,7 @@ usage is void (*_gtk_reserved1)(void);"""
             message.warn(
                 "Too many parameters in type specification %r" % (type_str, ))
             return base
+
         def top_combiner(base, *rest):
             if type_node is not None and isinstance(type_node, ast.Type):
                 base.is_const = type_node.is_const
@@ -327,24 +331,23 @@ usage is void (*_gtk_reserved1)(void);"""
         return block.position
 
     def _check_array_element_type(self, array, options):
+        array_type = array.array_type
+        element_type = array.element_type
+
         # GPtrArrays are allowed to contain non basic types
         # (except enums and flags) or basic types that are
         # as big as a gpointer
-        if array.array_type == ast.Array.GLIB_PTRARRAY and \
-           ((array.element_type in ast.BASIC_GIR_TYPES
-             and not array.element_type in ast.POINTER_TYPES) or
-            isinstance(array.element_type, ast.Enum) or
-            isinstance(array.element_type, ast.Bitfield)):
-            message.warn("invalid (element-type) for a GPtrArray, "
-                        "must be a pointer", options.position)
+        if array_type == ast.Array.GLIB_PTRARRAY:
+            if ((element_type in ast.BASIC_GIR_TYPES and not element_type in ast.POINTER_TYPES)
+            or isinstance(element_type, (ast.Enum, ast.Bitfield))):
+                message.warn("invalid (element-type) for a GPtrArray, "
+                             "must be a pointer", options.position)
 
         # GByteArrays have (element-type) guint8 by default
-        if array.array_type == ast.Array.GLIB_BYTEARRAY:
-            if array.element_type == ast.TYPE_ANY:
+        if array_type == ast.Array.GLIB_BYTEARRAY:
+            if element_type == ast.TYPE_ANY:
                 array.element_type = ast.TYPE_UINT8
-            elif not array.element_type in [ast.TYPE_UINT8,
-                                            ast.TYPE_INT8,
-                                            ast.TYPE_CHAR]:
+            elif not element_type in [ast.TYPE_UINT8, ast.TYPE_INT8, ast.TYPE_CHAR]:
                 message.warn("invalid (element-type) for a GByteArray, "
                              "must be one of guint8, gint8 or gchar",
                              options.position)
@@ -454,8 +457,8 @@ usage is void (*_gtk_reserved1)(void);"""
 
     def _get_transfer_default_returntype_basic(self, typeval):
         if (typeval.is_equiv(ast.BASIC_GIR_TYPES)
-            or typeval.is_const
-            or typeval.is_equiv(ast.TYPE_NONE)):
+        or typeval.is_const
+        or typeval.is_equiv(ast.TYPE_NONE)):
             return ast.PARAM_TRANSFER_NONE
         elif typeval.is_equiv(ast.TYPE_STRING):
             # Non-const strings default to FULL
@@ -535,8 +538,7 @@ usage is void (*_gtk_reserved1)(void);"""
 
         caller_allocates = False
         annotated_direction = None
-        if (OPT_INOUT in options or
-            OPT_INOUT_ALT in options):
+        if (OPT_INOUT in options or OPT_INOUT_ALT in options):
             annotated_direction = ast.PARAM_DIRECTION_INOUT
         elif OPT_OUT in options:
             subtype = options[OPT_OUT]
@@ -574,9 +576,9 @@ usage is void (*_gtk_reserved1)(void);"""
 
         self._adjust_container_type(parent, node, options)
 
-        if (OPT_ALLOW_NONE in options or
-            node.type.target_giname == 'Gio.AsyncReadyCallback' or
-            node.type.target_giname == 'Gio.Cancellable'):
+        if (OPT_ALLOW_NONE in options
+        or node.type.target_giname == 'Gio.AsyncReadyCallback'
+        or node.type.target_giname == 'Gio.Cancellable'):
             node.allow_none = True
 
         if tag is not None and tag.comment is not None:
@@ -605,7 +607,7 @@ usage is void (*_gtk_reserved1)(void);"""
             if ': ' in value:
                 colon = value.find(': ')
                 version = value[:colon]
-                desc = value[colon+2:]
+                desc = value[colon + 2:]
             else:
                 desc = value
                 version = None
@@ -718,8 +720,7 @@ usage is void (*_gtk_reserved1)(void);"""
                 (param, ) = unused
                 text = ', should be %r' % (param, )
             else:
-                text = ', should be one of %s' % (
-                ', '.join(repr(p) for p in unused), )
+                text = ', should be one of %s' % (', '.join(repr(p) for p in unused), )
 
             tag = block.params.get(doc_name)
             message.warn(
@@ -795,14 +796,14 @@ usage is void (*_gtk_reserved1)(void);"""
                 # Resolve real parameter names early, so that in later
                 # phase we can refer to them while resolving annotations.
                 for i, param in enumerate(signal.parameters):
-                    param.argname, tag = names[i+1]
+                    param.argname, tag = names[i + 1]
             else:
                 message.warn("incorrect number of parameters in comment block, "
                              "parameter annotations will be ignored.", block.position)
 
         for i, param in enumerate(signal.parameters):
             if names:
-                name, tag = names[i+1]
+                name, tag = names[i + 1]
                 options = getattr(tag, 'options', {})
                 param_type = options.get(OPT_TYPE)
                 if param_type:
@@ -1164,9 +1165,9 @@ method or constructor of some type."""
         origin_node = self._get_constructor_class(func, subsymbol)
         if origin_node is None:
             if func.is_constructor:
-                message.warn_node(func,
-                    "Can't find matching type for constructor; symbol=%r" \
-                    % (func.symbol, ))
+                message.warn_node(
+                    func,
+                    "Can't find matching type for constructor; symbol=%r" % (func.symbol, ))
             return False
 
         # Some sanity checks; only objects and boxeds can have ctors
diff --git a/giscanner/message.py b/giscanner/message.py
index 8a948cd3..3a330afe 100644
--- a/giscanner/message.py
+++ b/giscanner/message.py
@@ -61,7 +61,7 @@ class Position(object):
             return '%s:' % (filename, )
 
     def offset(self, offset):
-        return Position(self.filename, self.line+offset, self.column)
+        return Position(self.filename, self.line + offset, self.column)
 
 
 class MessageLogger(object):
@@ -119,16 +119,14 @@ If the warning is related to a ast.Node type, see log_node()."""
         elif log_type == FATAL:
             error_type = "Fatal"
         if prefix:
-            text = (
-'''%s: %s: %s: %s: %s\n''' % (last_position, error_type, self._namespace.name,
-                            prefix, text))
+            text = ('%s: %s: %s: %s: %s\n' % (last_position, error_type,
+                                              self._namespace.name, prefix, text))
         else:
             if self._namespace:
-                text = (
-'''%s: %s: %s: %s\n''' % (last_position, error_type, self._namespace.name, text))
+                text = ('%s: %s: %s: %s\n' % (last_position, error_type,
+                                              self._namespace.name, text))
             else:
-                text = (
-'''%s: %s: %s\n''' % (last_position, error_type, text))
+                text = ('%s: %s: %s\n' % (last_position, error_type, text))
 
         self._output.write(text)
         if log_type == FATAL:
@@ -169,17 +167,21 @@ def log_node(log_type, node, text, context=None, positions=None):
     ml = MessageLogger.get()
     ml.log_node(log_type, node, text, context=context, positions=positions)
 
+
 def warn(text, positions=None, prefix=None):
     ml = MessageLogger.get()
     ml.log(WARNING, text, positions, prefix)
 
+
 def warn_node(node, text, context=None, positions=None):
     log_node(WARNING, node, text, context=context, positions=positions)
 
+
 def warn_symbol(symbol, text):
     ml = MessageLogger.get()
     ml.log_symbol(WARNING, symbol, text)
 
+
 def fatal(text, positions=None, prefix=None):
     ml = MessageLogger.get()
     ml.log(FATAL, text, positions, prefix)
diff --git a/giscanner/scannermain.py b/giscanner/scannermain.py
index 8ccd1f09..00dc30d6 100755
--- a/giscanner/scannermain.py
+++ b/giscanner/scannermain.py
@@ -42,14 +42,17 @@ from giscanner.sourcescanner import SourceScanner, ALL_EXTS
 from giscanner.transformer import Transformer
 from . import utils
 
+
 def process_cflags_begin(option, opt, value, parser):
     cflags = getattr(parser.values, option.dest)
     while len(parser.rargs) > 0 and parser.rargs[0] != '--cflags-end':
         cflags.append(parser.rargs.pop(0))
 
+
 def process_cflags_end(option, opt, value, parser):
     pass
 
+
 def get_preprocessor_option_group(parser):
     group = optparse.OptionGroup(parser, "Preprocessor options")
     group.add_option("", "--cflags-begin",
@@ -71,6 +74,7 @@ def get_preprocessor_option_group(parser):
     group.add_option("-p", dest="", help="Ignored")
     return group
 
+
 def get_windows_option_group(parser):
     group = optparse.OptionGroup(parser, "Machine Dependent Options")
     group.add_option("-m", help="some machine dependent option",
@@ -79,13 +83,13 @@ def get_windows_option_group(parser):
 
     return group
 
+
 def _get_option_parser():
     parser = optparse.OptionParser('%prog [options] sources')
     parser.add_option('', "--quiet",
                       action="store_true", dest="quiet",
                       default=False,
-                      help="If passed, do not print details of normal" \
-                          + " operation")
+                      help="If passed, do not print details of normal operation")
     parser.add_option("", "--format",
                       action="store", dest="format",
                       default="gir",
@@ -204,6 +208,7 @@ match the namespace prefix.""")
 def _error(msg):
     raise SystemExit('ERROR: %s' % (msg, ))
 
+
 def passthrough_gir(path, f):
     parser = GIRParser()
     parser.parse(path)
@@ -211,6 +216,7 @@ def passthrough_gir(path, f):
     writer = GIRWriter(parser.get_namespace())
     f.write(writer.get_xml())
 
+
 def test_codegen(optstring):
     (namespace, out_h_filename, out_c_filename) = optstring.split(',')
     if namespace == 'Everything':
@@ -221,6 +227,7 @@ def test_codegen(optstring):
         _error("Invaild namespace %r" % (namespace, ))
     return 0
 
+
 def process_options(output, allowed_flags):
     for option in output.split():
         for flag in allowed_flags:
@@ -229,6 +236,7 @@ def process_options(output, allowed_flags):
             yield option
             break
 
+
 def process_packages(options, packages):
     args = ['pkg-config', '--cflags']
     args.extend(packages)
@@ -248,6 +256,7 @@ def process_packages(options, packages):
     options.cpp_defines.extend(pkg_options.cpp_defines)
     options.cpp_undefines.extend(pkg_options.cpp_undefines)
 
+
 def extract_filenames(args):
     filenames = []
     for arg in args:
@@ -261,6 +270,7 @@ def extract_filenames(args):
             filenames.append(os.path.abspath(arg))
     return filenames
 
+
 def extract_filelist(options):
     filenames = []
     if not os.path.exists(options.filelist):
@@ -271,10 +281,10 @@ def extract_filelist(options):
         # We don't support real C++ parsing yet, but we should be able
         # to understand C API implemented in C++ files.
         filename = line.strip()
-        if (filename.endswith('.c') or filename.endswith('.cpp') or
-            filename.endswith('.cc') or filename.endswith('.cxx') or
-            filename.endswith('.h') or filename.endswith('.hpp') or
-            filename.endswith('.hxx')):
+        if (filename.endswith('.c') or filename.endswith('.cpp')
+        or filename.endswith('.cc') or filename.endswith('.cxx')
+        or filename.endswith('.h') or filename.endswith('.hpp')
+        or filename.endswith('.hxx')):
             if not os.path.exists(filename):
                 _error('%s: Invalid filelist entry-no such file or directory' % (line, ))
             # Make absolute, because we do comparisons inside scannerparser.c
@@ -282,6 +292,7 @@ def extract_filelist(options):
             filenames.append(os.path.abspath(filename))
     return filenames
 
+
 def create_namespace(options):
     if options.strip_prefix:
         print """g-ir-scanner: warning: Option --strip-prefix has been deprecated;
@@ -310,6 +321,7 @@ see --identifier-prefix and --symbol-prefix."""
                      identifier_prefixes=identifier_prefixes,
                      symbol_prefixes=symbol_prefixes)
 
+
 def create_transformer(namespace, options):
     transformer = Transformer(namespace,
                               accept_unprefixed=options.accept_unprefixed)
@@ -331,6 +343,7 @@ def create_transformer(namespace, options):
 
     return transformer
 
+
 def create_binary(transformer, options, args):
     # Transform the C AST nodes into higher level
     # GLib/GObject nodes
@@ -341,7 +354,7 @@ def create_binary(transformer, options, args):
     gdump_parser.init_parse()
 
     if options.program:
-        args=[options.program]
+        args = [options.program]
         args.extend(options.program_args)
         binary = IntrospectionBinary(args)
     else:
@@ -354,6 +367,7 @@ def create_binary(transformer, options, args):
     gdump_parser.parse()
     return shlibs
 
+
 def create_source_scanner(options, args):
     if hasattr(options, 'filelist') and options.filelist:
         filenames = extract_filelist(options)
@@ -371,6 +385,7 @@ def create_source_scanner(options, args):
     ss.parse_macros(filenames)
     return ss
 
+
 def write_output(data, options):
     if options.output == "-":
         output = sys.stdout
@@ -407,6 +422,7 @@ def write_output(data, options):
     except IOError as e:
         _error("while writing output: %s" % (e.strerror, ))
 
+
 def scanner_main(args):
     parser = _get_option_parser()
     (options, args) = parser.parse_args(args)
diff --git a/giscanner/sectionparser.py b/giscanner/sectionparser.py
index 0b013be1..61763e45 100644
--- a/giscanner/sectionparser.py
+++ b/giscanner/sectionparser.py
@@ -19,10 +19,12 @@
 
 import re
 
+
 class SectionsFile(object):
     def __init__(self, sections):
         self.sections = sections
 
+
 class Section(object):
     def __init__(self):
         self.file = None
@@ -30,11 +32,13 @@ class Section(object):
         self.includes = None
         self.subsections = []
 
+
 class Subsection(object):
     def __init__(self, name):
         self.name = name
         self.symbols = []
 
+
 def parse_sections_file(lines):
     sections = []
     current_section = None
diff --git a/giscanner/shlibs.py b/giscanner/shlibs.py
index 4f622db3..1241827d 100644
--- a/giscanner/shlibs.py
+++ b/giscanner/shlibs.py
@@ -26,6 +26,7 @@ import subprocess
 
 from .utils import get_libtool_command, extract_libtool_shlib
 
+
 # For .la files, the situation is easy.
 def _resolve_libtool(options, binary, libraries):
     shlibs = []
@@ -36,6 +37,7 @@ def _resolve_libtool(options, binary, libraries):
 
     return shlibs
 
+
 # Assume ldd output is something vaguely like
 #
 #  libpangoft2-1.0.so.0 => /usr/lib/libpangoft2-1.0.so.0 (0x006c1000)
@@ -51,6 +53,7 @@ def _ldd_library_pattern(library_name):
     return re.compile("(?<![A-Za-z0-9_-])(lib*%s[^A-Za-z0-9_-][^\s\(\)]*)"
                       % re.escape(library_name))
 
+
 # This is a what we do for non-la files. We assume that we are on an
 # ELF-like system where ldd exists and the soname extracted with ldd is
 # a filename that can be opened with dlopen().
@@ -118,6 +121,7 @@ def _resolve_non_libtool(options, binary, libraries):
 
     return shlibs
 
+
 # We want to resolve a set of library names (the <foo> of -l<foo>)
 # against a library to find the shared library name. The shared
 # library name is suppose to be what you pass to dlopen() (or
diff --git a/giscanner/sourcescanner.py b/giscanner/sourcescanner.py
index 736ddbd4..d5c43926 100644
--- a/giscanner/sourcescanner.py
+++ b/giscanner/sourcescanner.py
@@ -93,8 +93,7 @@ def symbol_type_name(symbol_type):
         CSYMBOL_TYPE_UNION: 'union',
         CSYMBOL_TYPE_ENUM: 'enum',
         CSYMBOL_TYPE_TYPEDEF: 'typedef',
-        CSYMBOL_TYPE_MEMBER: 'member',
-        }.get(symbol_type)
+        CSYMBOL_TYPE_MEMBER: 'member'}.get(symbol_type)
 
 
 def ctype_name(ctype):
@@ -108,8 +107,7 @@ def ctype_name(ctype):
         CTYPE_ENUM: 'enum',
         CTYPE_POINTER: 'pointer',
         CTYPE_ARRAY: 'array',
-        CTYPE_FUNCTION: 'function',
-        }.get(ctype)
+        CTYPE_FUNCTION: 'function'}.get(ctype)
 
 
 class SourceType(object):
@@ -266,7 +264,7 @@ class SourceScanner(object):
         return self._scanner.get_comments()
 
     def dump(self):
-        print '-'*30
+        print '-' * 30
         for symbol in self._scanner.get_symbols():
             print symbol.ident, symbol.base_type.name, symbol.type
 
diff --git a/giscanner/testcodegen.py b/giscanner/testcodegen.py
index f304dc7a..1ed247c7 100644
--- a/giscanner/testcodegen.py
+++ b/giscanner/testcodegen.py
@@ -27,12 +27,14 @@ DEFAULT_C_VALUES = {ast.TYPE_ANY: 'NULL',
                     ast.TYPE_FILENAME: '""',
                     ast.TYPE_GTYPE: 'g_object_get_type ()'}
 
+
 def get_default_for_typeval(typeval):
     default = DEFAULT_C_VALUES.get(typeval)
     if default:
         return default
     return "0"
 
+
 def uscore_from_type(typeval):
     if typeval.target_fundamental:
         return typeval.target_fundamental.replace(' ', '_')
@@ -41,6 +43,7 @@ def uscore_from_type(typeval):
     else:
         assert False, typeval
 
+
 class EverythingCodeGenerator(object):
 
     def __init__(self, out_h_filename, out_c_filename):
diff --git a/giscanner/transformer.py b/giscanner/transformer.py
index f1aa165d..f70d756a 100644
--- a/giscanner/transformer.py
+++ b/giscanner/transformer.py
@@ -34,6 +34,7 @@ from .sourcescanner import (
     CSYMBOL_TYPE_MEMBER, CSYMBOL_TYPE_ELLIPSIS, CSYMBOL_TYPE_CONST,
     TYPE_QUALIFIER_CONST, TYPE_QUALIFIER_VOLATILE)
 
+
 class TransformerException(Exception):
     pass
 
@@ -54,7 +55,7 @@ class Transformer(object):
         self._namespace = namespace
         self._pkg_config_packages = set()
         self._typedefs_ns = {}
-        self._parsed_includes = {} # <string namespace -> Namespace>
+        self._parsed_includes = {}  # <string namespace -> Namespace>
         self._includepaths = []
         self._passthrough_mode = False
 
@@ -104,7 +105,7 @@ class Transformer(object):
             if not ns_compound:
                 ns_compound = self._namespace.get('_' + compound.name)
             if (not ns_compound and isinstance(compound, (ast.Record, ast.Union))
-                and len(compound.fields) == 0):
+            and len(compound.fields) == 0):
                 disguised = ast.Record(compound.name, typedef, disguised=True)
                 self._namespace.append(disguised)
             elif not ns_compound:
@@ -126,8 +127,8 @@ class Transformer(object):
     def register_include_uninstalled(self, include_path):
         basename = os.path.basename(include_path)
         if not basename.endswith('.gir'):
-            raise SystemExit(
-"Include path %r must be a filename path ending in .gir" % (include_path, ))
+            raise SystemExit("Include path %r must be a filename path "
+                             "ending in .gir" % (include_path, ))
         girname = basename[:-4]
         include = ast.Include.from_string(girname)
         if include in self._namespace.includes:
@@ -175,8 +176,7 @@ None."""
             path = os.path.join(d, girname)
             if os.path.exists(path):
                 return path
-        sys.stderr.write("Couldn't find include %r (search path: %r)\n"\
-                         % (girname, searchdirs))
+        sys.stderr.write("Couldn't find include %r (search path: %r)\n" % (girname, searchdirs))
         sys.exit(1)
 
     @classmethod
@@ -228,7 +228,7 @@ currently-scanned namespace is first."""
 
     def _split_c_string_for_namespace_matches(self, name, is_identifier=False):
         matches = []  # Namespaces which might contain this name
-        unprefixed_namespaces = [] # Namespaces with no prefix, last resort
+        unprefixed_namespaces = []  # Namespaces with no prefix, last resort
         for ns in self._iter_namespaces():
             if is_identifier:
                 prefixes = ns.identifier_prefixes
@@ -502,8 +502,8 @@ raise ValueError."""
 
     def _create_member(self, symbol, parent_symbol=None):
         source_type = symbol.base_type
-        if (source_type.type == CTYPE_POINTER and
-            symbol.base_type.base_type.type == CTYPE_FUNCTION):
+        if (source_type.type == CTYPE_POINTER
+        and symbol.base_type.base_type.type == CTYPE_FUNCTION):
             node = self._create_callback(symbol, member=True)
         elif source_type.type == CTYPE_STRUCT and source_type.name is None:
             node = self._create_struct(symbol, anonymous=True)
@@ -519,8 +519,8 @@ raise ValueError."""
                 # to be able to properly calculate the size of the compound
                 # type (e.g. GValue) that contains this array, see
                 # <https://bugzilla.gnome.org/show_bug.cgi?id=657040>.
-                if (source_type.base_type.type == CTYPE_UNION and
-                    source_type.base_type.name is None):
+                if (source_type.base_type.type == CTYPE_UNION
+                and source_type.base_type.name is None):
                     synthesized_type = self._synthesize_union_type(symbol, parent_symbol)
                     ftype = ast.Array(None, synthesized_type, complete_ctype=complete_ctype)
                 else:
@@ -558,11 +558,9 @@ raise ValueError."""
 
     def _create_typedef(self, symbol):
         ctype = symbol.base_type.type
-        if (ctype == CTYPE_POINTER and
-            symbol.base_type.base_type.type == CTYPE_FUNCTION):
+        if (ctype == CTYPE_POINTER and symbol.base_type.base_type.type == CTYPE_FUNCTION):
             node = self._create_typedef_callback(symbol)
-        elif (ctype == CTYPE_POINTER and
-            symbol.base_type.base_type.type == CTYPE_STRUCT):
+        elif (ctype == CTYPE_POINTER and symbol.base_type.base_type.type == CTYPE_STRUCT):
             node = self._create_typedef_struct(symbol, disguised=True)
         elif ctype == CTYPE_STRUCT:
             node = self._create_typedef_struct(symbol)
@@ -627,8 +625,7 @@ raise ValueError."""
         derefed_typename = canonical.replace('*', '')
 
         # Preserve "pointerness" of struct/union members
-        if (is_member and canonical.endswith('*') and
-            derefed_typename in ast.basic_type_names):
+        if (is_member and canonical.endswith('*') and derefed_typename in ast.basic_type_names):
             return 'gpointer'
         else:
             return derefed_typename
@@ -714,8 +711,7 @@ raise ValueError."""
 
         # Don't create constants for non-public things
         # http://bugzilla.gnome.org/show_bug.cgi?id=572790
-        if (symbol.source_filename is None or
-            not symbol.source_filename.endswith('.h')):
+        if (symbol.source_filename is None or not symbol.source_filename.endswith('.h')):
             return None
         try:
             name = self._strip_symbol(symbol)
@@ -738,13 +734,13 @@ raise ValueError."""
                 if isinstance(target, ast.Type):
                     unaliased = target
             if unaliased == ast.TYPE_UINT64:
-                value = str(symbol.const_int % 2**64)
+                value = str(symbol.const_int % 2 ** 64)
             elif unaliased == ast.TYPE_UINT32:
-                value = str(symbol.const_int % 2**32)
+                value = str(symbol.const_int % 2 ** 32)
             elif unaliased == ast.TYPE_UINT16:
-                value = str(symbol.const_int % 2**16)
+                value = str(symbol.const_int % 2 ** 16)
             elif unaliased == ast.TYPE_UINT8:
-                value = str(symbol.const_int % 2**16)
+                value = str(symbol.const_int % 2 ** 16)
             else:
                 value = str(symbol.const_int)
         elif symbol.const_double is not None:
@@ -842,8 +838,7 @@ raise ValueError."""
 
         # Mark the 'user_data' arguments
         for i, param in enumerate(parameters):
-            if (param.type.target_fundamental == 'gpointer' and
-                param.argname == 'user_data'):
+            if (param.type.target_fundamental == 'gpointer' and param.argname == 'user_data'):
                 param.closure_name = param.argname
 
         if member:
diff --git a/giscanner/utils.py b/giscanner/utils.py
index 9adf6d93..77d05b9e 100644
--- a/giscanner/utils.py
+++ b/giscanner/utils.py
@@ -22,7 +22,10 @@ import re
 import os
 import subprocess
 
+
 _debugflags = None
+
+
 def have_debug_flag(flag):
     """Check for whether a specific debugging feature is enabled.
 Well-known flags:
@@ -38,6 +41,7 @@ Well-known flags:
             _debugflags.remove('')
     return flag in _debugflags
 
+
 def break_on_debug_flag(flag):
     if have_debug_flag(flag):
         import pdb
@@ -69,8 +73,10 @@ def to_underscores_noprefix(name):
     name = _upperstr_pat2.sub(r'\1_\2', name)
     return name
 
+
 _libtool_pat = re.compile("dlname='([A-z0-9\.\-\+]+)'\n")
 
+
 def _extract_dlname_field(la_file):
     f = open(la_file)
     data = f.read()
@@ -81,6 +87,7 @@ def _extract_dlname_field(la_file):
     else:
         return None
 
+
 # Returns the name that we would pass to dlopen() the library
 # corresponding to this .la file
 def extract_libtool_shlib(la_file):
@@ -92,6 +99,7 @@ def extract_libtool_shlib(la_file):
     # a path rather than the raw dlname
     return os.path.basename(dlname)
 
+
 def extract_libtool(la_file):
     dlname = _extract_dlname_field(la_file)
     if dlname is None:
@@ -104,6 +112,7 @@ def extract_libtool(la_file):
     libname = libname.replace('.libs/.libs', '.libs')
     return libname
 
+
 # Returns arguments for invoking libtool, if applicable, otherwise None
 def get_libtool_command(options):
     libtool_infection = not options.nolibtool
diff --git a/giscanner/xmlwriter.py b/giscanner/xmlwriter.py
index 4a1dc9fb..6efe684d 100755
--- a/giscanner/xmlwriter.py
+++ b/giscanner/xmlwriter.py
@@ -44,8 +44,7 @@ def _calc_attrs_length(attributes, indent, self_indent):
     return attr_length + indent + self_indent
 
 
-def collect_attributes(tag_name, attributes, self_indent,
-                       self_indent_char, indent=-1):
+def collect_attributes(tag_name, attributes, self_indent, self_indent_char, indent=-1):
     if not attributes:
         return ''
     if _calc_attrs_length(attributes, indent, self_indent) > 79:
@@ -68,6 +67,7 @@ def collect_attributes(tag_name, attributes, self_indent,
             first = False
     return attr_value
 
+
 def build_xml_tag(tag_name, attributes=None, data=None, self_indent=0,
                   self_indent_char=' '):
     if attributes is None:
@@ -86,6 +86,7 @@ def build_xml_tag(tag_name, attributes=None, data=None, self_indent=0,
         len(prefix) + len(suffix))
     return prefix + attrs + suffix
 
+
 with LibtoolImporter(None, None):
     if 'UNINSTALLED_INTROSPECTION_SRCDIR' in os.environ:
         from _giscanner import collect_attributes
@@ -108,10 +109,8 @@ class XMLWriter(object):
     def _open_tag(self, tag_name, attributes=None):
         if attributes is None:
             attributes = []
-        attrs = collect_attributes(
-            tag_name, attributes, self._indent,
-            self._indent_char,
-            len(tag_name) + 2)
+        attrs = collect_attributes(tag_name, attributes,
+                                   self._indent, self._indent_char, len(tag_name) + 2)
         self.write_line(u'<%s%s>' % (tag_name, attrs))
 
     def _close_tag(self, tag_name):
@@ -137,10 +136,9 @@ class XMLWriter(object):
         if do_escape:
             line = escape(line)
         if indent:
-            self._data.write('%s%s%s' % (
-                    self._indent_char * self._indent,
-                    line.encode('utf-8'),
-                    self._newline_char))
+            self._data.write('%s%s%s' % (self._indent_char * self._indent,
+                                         line.encode('utf-8'),
+                                         self._newline_char))
         else:
             self._data.write('%s%s' % (line.encode('utf-8'), self._newline_char))
 
diff --git a/misc/pep8.py b/misc/pep8.py
index 38906294..f99ae3a3 100644
--- a/misc/pep8.py
+++ b/misc/pep8.py
@@ -1,6 +1,7 @@
-#!/usr/bin/python
+#!/usr/bin/env python
 # pep8.py - Check Python source code formatting, according to PEP 8
-# Copyright (C) 2006 Johann C. Rocholl <johann@browsershots.org>
+# Copyright (C) 2006-2009 Johann C. Rocholl <johann@rocholl.net>
+# Copyright (C) 2009-2013 Florent Xicluna <florent.xicluna@gmail.com>
 #
 # Permission is hereby granted, free of charge, to any person
 # obtaining a copy of this software and associated documentation files
@@ -22,7 +23,7 @@
 # CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 # SOFTWARE.
 
-"""
+r"""
 Check Python source code formatting, according to PEP 8:
 http://www.python.org/dev/peps/pep-0008/
 
@@ -30,8 +31,7 @@ For usage and a list of options, try this:
 $ python pep8.py -h
 
 This program and its regression test suite live here:
-http://svn.browsershots.org/trunk/devtools/pep8/
-http://trac.browsershots.org/browser/trunk/devtools/pep8/
+http://github.com/jcrocholl/pep8
 
 Groups of errors and warnings:
 E errors
@@ -43,72 +43,72 @@ W warnings
 500 line length
 600 deprecation
 700 statements
-
-You can add checks to this program by writing plugins. Each plugin is
-a simple function that is called for each line of source code, either
-physical or logical.
-
-Physical line:
-- Raw line of text from the input file.
-
-Logical line:
-- Multi-line statements converted to a single line.
-- Stripped left and right.
-- Contents of strings replaced with 'xxx' of same length.
-- Comments removed.
-
-The check function requests physical or logical lines by the name of
-the first argument:
-
-def maximum_line_length(physical_line)
-def extraneous_whitespace(logical_line)
-def blank_lines(logical_line, blank_lines, indent_level, line_number)
-
-The last example above demonstrates how check plugins can request
-additional information with extra arguments. All attributes of the
-Checker object are available. Some examples:
-
-lines: a list of the raw lines from the input file
-tokens: the tokens that contribute to this logical line
-line_number: line number in the input file
-blank_lines: blank lines before this one
-indent_char: first indentation character in this file (' ' or '\t')
-indent_level: indentation (with tabs expanded to multiples of 8)
-previous_indent_level: indentation on previous line
-previous_logical: previous logical line
-
-The docstring of each check function shall be the relevant part of
-text from PEP 8. It is printed if the user enables --show-pep8.
-
+900 syntax error
 """
+__version__ = '1.4.5'
 
 import os
 import sys
 import re
 import time
 import inspect
+import keyword
 import tokenize
 from optparse import OptionParser
-from keyword import iskeyword
 from fnmatch import fnmatch
-
-__version__ = '0.2.0'
-__revision__ = '$Rev$'
-
-default_exclude = '.svn,CVS,*.pyc,*.pyo'
-
-indent_match = re.compile(r'([ \t]*)').match
-raise_comma_match = re.compile(r'raise\s+\w+\s*(,)').match
-
-operators = """
-+  -  *  /  %  ^  &  |  =  <  >  >>  <<
-+= -= *= /= %= ^= &= |= == <= >= >>= <<=
-!= <> :
-in is or not and
-""".split()
-
-options = None
-args = None
+try:
+    from configparser import RawConfigParser
+    from io import TextIOWrapper
+except ImportError:
+    from ConfigParser import RawConfigParser
+
+DEFAULT_EXCLUDE = '.svn,CVS,.bzr,.hg,.git,__pycache__'
+DEFAULT_IGNORE = 'E226,E24'
+if sys.platform == 'win32':
+    DEFAULT_CONFIG = os.path.expanduser(r'~\.pep8')
+else:
+    DEFAULT_CONFIG = os.path.join(os.getenv('XDG_CONFIG_HOME') or
+                                  os.path.expanduser('~/.config'), 'pep8')
+PROJECT_CONFIG = ('.pep8', 'tox.ini', 'setup.cfg')
+TESTSUITE_PATH = os.path.join(os.path.dirname(__file__), 'testsuite')
+MAX_LINE_LENGTH = 79
+REPORT_FORMAT = {
+    'default': '%(path)s:%(row)d:%(col)d: %(code)s %(text)s',
+    'pylint': '%(path)s:%(row)d: [%(code)s] %(text)s',
+}
+
+PyCF_ONLY_AST = 1024
+SINGLETONS = frozenset(['False', 'None', 'True'])
+KEYWORDS = frozenset(keyword.kwlist + ['print']) - SINGLETONS
+UNARY_OPERATORS = frozenset(['>>', '**', '*', '+', '-'])
+ARITHMETIC_OP = frozenset(['**', '*', '/', '//', '+', '-'])
+WS_OPTIONAL_OPERATORS = ARITHMETIC_OP.union(['^', '&', '|', '<<', '>>', '%'])
+WS_NEEDED_OPERATORS = frozenset([
+    '**=', '*=', '/=', '//=', '+=', '-=', '!=', '<>', '<', '>',
+    '%=', '^=', '&=', '|=', '==', '<=', '>=', '<<=', '>>=', '='])
+WHITESPACE = frozenset(' \t')
+SKIP_TOKENS = frozenset([tokenize.COMMENT, tokenize.NL, tokenize.NEWLINE,
+                         tokenize.INDENT, tokenize.DEDENT])
+BENCHMARK_KEYS = ['directories', 'files', 'logical lines', 'physical lines']
+
+INDENT_REGEX = re.compile(r'([ \t]*)')
+RAISE_COMMA_REGEX = re.compile(r'raise\s+\w+\s*,')
+RERAISE_COMMA_REGEX = re.compile(r'raise\s+\w+\s*,\s*\w+\s*,\s*\w+')
+ERRORCODE_REGEX = re.compile(r'\b[A-Z]\d{3}\b')
+DOCSTRING_REGEX = re.compile(r'u?r?["\']')
+EXTRANEOUS_WHITESPACE_REGEX = re.compile(r'[[({] | []}),;:]')
+WHITESPACE_AFTER_COMMA_REGEX = re.compile(r'[,;:]\s*(?:  |\t)')
+COMPARE_SINGLETON_REGEX = re.compile(r'([=!]=)\s*(None|False|True)')
+COMPARE_TYPE_REGEX = re.compile(r'(?:[=!]=|is(?:\s+not)?)\s*type(?:s.\w+Type'
+                                r'|\s*\(\s*([^)]*[^ )])\s*\))')
+KEYWORD_REGEX = re.compile(r'(\s*)\b(?:%s)\b(\s*)' % r'|'.join(KEYWORDS))
+OPERATOR_REGEX = re.compile(r'(?:[^,\s])(\s*)(?:[-+*/|!<=>%&^]+)(\s*)')
+LAMBDA_REGEX = re.compile(r'\blambda\b')
+HUNK_REGEX = re.compile(r'^@@ -\d+(?:,\d+)? \+(\d+)(?:,(\d+))? @@.*$')
+
+# Work around Python < 2.6 behaviour, which does not generate NL after
+# a comment which is on a line by itself.
+COMMENT_WITH_NL = tokenize.generate_tokens(['#\n'].pop).send(None)[1] == '#\n'
 
 
 ##############################################################################
@@ -117,7 +117,7 @@ args = None
 
 
 def tabs_or_spaces(physical_line, indent_char):
-    """
+    r"""
     Never mix tabs and spaces.
 
     The most popular way of indenting Python is with spaces only.  The
@@ -126,52 +126,81 @@ def tabs_or_spaces(physical_line, indent_char):
     invoking the Python command line interpreter with the -t option, it issues
     warnings about code that illegally mixes tabs and spaces.  When using -tt
     these warnings become errors.  These options are highly recommended!
+
+    Okay: if a == 0:\n        a = 1\n        b = 1
+    E101: if a == 0:\n        a = 1\n\tb = 1
     """
-    indent = indent_match(physical_line).group(1)
+    indent = INDENT_REGEX.match(physical_line).group(1)
     for offset, char in enumerate(indent):
         if char != indent_char:
             return offset, "E101 indentation contains mixed spaces and tabs"
 
 
 def tabs_obsolete(physical_line):
-    """
+    r"""
     For new projects, spaces-only are strongly recommended over tabs.  Most
     editors have features that make this easy to do.
+
+    Okay: if True:\n    return
+    W191: if True:\n\treturn
     """
-    indent = indent_match(physical_line).group(1)
-    if indent.count('\t'):
+    indent = INDENT_REGEX.match(physical_line).group(1)
+    if '\t' in indent:
         return indent.index('\t'), "W191 indentation contains tabs"
 
 
 def trailing_whitespace(physical_line):
-    """
+    r"""
     JCR: Trailing whitespace is superfluous.
+    FBM: Except when it occurs as part of a blank line (i.e. the line is
+         nothing but whitespace). According to Python docs[1] a line with only
+         whitespace is considered a blank line, and is to be ignored. However,
+         matching a blank line to its indentation level avoids mistakenly
+         terminating a multi-line statement (e.g. class declaration) when
+         pasting code into the standard Python interpreter.
+
+         [1] http://docs.python.org/reference/lexical_analysis.html#blank-lines
+
+    The warning returned varies on whether the line itself is blank, for easier
+    filtering for those who want to indent their blank lines.
+
+    Okay: spam(1)\n#
+    W291: spam(1) \n#
+    W293: class Foo(object):\n    \n    bang = 12
     """
-    physical_line = physical_line.rstrip('\n') # chr(10), newline
-    physical_line = physical_line.rstrip('\r') # chr(13), carriage return
-    physical_line = physical_line.rstrip('\x0c') # chr(12), form feed, ^L
-    stripped = physical_line.rstrip()
+    physical_line = physical_line.rstrip('\n')    # chr(10), newline
+    physical_line = physical_line.rstrip('\r')    # chr(13), carriage return
+    physical_line = physical_line.rstrip('\x0c')  # chr(12), form feed, ^L
+    stripped = physical_line.rstrip(' \t\v')
     if physical_line != stripped:
-        return len(stripped), "W291 trailing whitespace"
+        if stripped:
+            return len(stripped), "W291 trailing whitespace"
+        else:
+            return 0, "W293 blank line contains whitespace"
 
 
 def trailing_blank_lines(physical_line, lines, line_number):
-    """
+    r"""
     JCR: Trailing blank lines are superfluous.
+
+    Okay: spam(1)
+    W391: spam(1)\n
     """
-    if physical_line.strip() == '' and line_number == len(lines):
+    if not physical_line.rstrip() and line_number == len(lines):
         return 0, "W391 blank line at end of file"
 
 
 def missing_newline(physical_line):
     """
     JCR: The last line should have a newline.
+
+    Reports warning W292.
     """
     if physical_line.rstrip() == physical_line:
         return len(physical_line), "W292 no newline at end of file"
 
 
-def maximum_line_length(physical_line):
+def maximum_line_length(physical_line, max_line_length):
     """
     Limit all lines to a maximum of 79 characters.
 
@@ -181,10 +210,23 @@ def maximum_line_length(physical_line):
     ugly.  Therefore, please limit all lines to a maximum of 79 characters.
     For flowing long blocks of text (docstrings or comments), limiting the
     length to 72 characters is recommended.
+
+    Reports error E501.
     """
-    length = len(physical_line.rstrip())
-    if length > 99:
-        return 79, "E501 line too long (%d characters)" % length
+    line = physical_line.rstrip()
+    length = len(line)
+    if length > max_line_length:
+        if noqa(line):
+            return
+        if hasattr(line, 'decode'):   # Python 2
+            # The line could contain multi-byte characters
+            try:
+                length = len(line.decode('utf-8'))
+            except UnicodeError:
+                pass
+        if length > max_line_length:
+            return (max_line_length, "E501 line too long "
+                    "(%d > %d characters)" % (length, max_line_length))
 
 
 ##############################################################################
@@ -192,6 +234,44 @@ def maximum_line_length(physical_line):
 ##############################################################################
 
 
+def blank_lines(logical_line, blank_lines, indent_level, line_number,
+                previous_logical, previous_indent_level):
+    r"""
+    Separate top-level function and class definitions with two blank lines.
+
+    Method definitions inside a class are separated by a single blank line.
+
+    Extra blank lines may be used (sparingly) to separate groups of related
+    functions.  Blank lines may be omitted between a bunch of related
+    one-liners (e.g. a set of dummy implementations).
+
+    Use blank lines in functions, sparingly, to indicate logical sections.
+
+    Okay: def a():\n    pass\n\n\ndef b():\n    pass
+    Okay: def a():\n    pass\n\n\n# Foo\n# Bar\n\ndef b():\n    pass
+
+    E301: class Foo:\n    b = 0\n    def bar():\n        pass
+    E302: def a():\n    pass\n\ndef b(n):\n    pass
+    E303: def a():\n    pass\n\n\n\ndef b(n):\n    pass
+    E303: def a():\n\n\n\n    pass
+    E304: @decorator\n\ndef a():\n    pass
+    """
+    if line_number < 3 and not previous_logical:
+        return  # Don't expect blank lines before the first line
+    if previous_logical.startswith('@'):
+        if blank_lines:
+            yield 0, "E304 blank lines found after function decorator"
+    elif blank_lines > 2 or (indent_level and blank_lines == 2):
+        yield 0, "E303 too many blank lines (%d)" % blank_lines
+    elif logical_line.startswith(('def ', 'class ', '@')):
+        if indent_level:
+            if not (blank_lines or previous_indent_level < indent_level or
+                    DOCSTRING_REGEX.match(previous_logical)):
+                yield 0, "E301 expected 1 blank line, found 0"
+        elif blank_lines != 2:
+            yield 0, "E302 expected 2 blank lines, found %d" % blank_lines
+
+
 def extraneous_whitespace(logical_line):
     """
     Avoid extraneous whitespace in the following situations:
@@ -199,51 +279,273 @@ def extraneous_whitespace(logical_line):
     - Immediately inside parentheses, brackets or braces.
 
     - Immediately before a comma, semicolon, or colon.
+
+    Okay: spam(ham[1], {eggs: 2})
+    E201: spam( ham[1], {eggs: 2})
+    E201: spam(ham[ 1], {eggs: 2})
+    E201: spam(ham[1], { eggs: 2})
+    E202: spam(ham[1], {eggs: 2} )
+    E202: spam(ham[1 ], {eggs: 2})
+    E202: spam(ham[1], {eggs: 2 })
+
+    E203: if x == 4: print x, y; x, y = y , x
+    E203: if x == 4: print x, y ; x, y = y, x
+    E203: if x == 4 : print x, y; x, y = y, x
     """
     line = logical_line
-    for char in '([{':
-        found = line.find(char + ' ')
-        if found > -1:
-            return found + 1, "E201 whitespace after '%s'" % char
-    for char in '}])':
-        found = line.find(' ' + char)
-        if found > -1 and line[found - 1] != ',':
-            return found, "E202 whitespace before '%s'" % char
-    for char in ',;:':
-        found = line.find(' ' + char)
-        if found > -1:
-            return found, "E203 whitespace before '%s'" % char
+    for match in EXTRANEOUS_WHITESPACE_REGEX.finditer(line):
+        text = match.group()
+        char = text.strip()
+        found = match.start()
+        if text == char + ' ':
+            # assert char in '([{'
+            yield found + 1, "E201 whitespace after '%s'" % char
+        elif line[found - 1] != ',':
+            code = ('E202' if char in '}])' else 'E203')  # if char in ',;:'
+            yield found, "%s whitespace before '%s'" % (code, char)
+
+
+def whitespace_around_keywords(logical_line):
+    r"""
+    Avoid extraneous whitespace around keywords.
+
+    Okay: True and False
+    E271: True and  False
+    E272: True  and False
+    E273: True and\tFalse
+    E274: True\tand False
+    """
+    for match in KEYWORD_REGEX.finditer(logical_line):
+        before, after = match.groups()
+
+        if '\t' in before:
+            yield match.start(1), "E274 tab before keyword"
+        elif len(before) > 1:
+            yield match.start(1), "E272 multiple spaces before keyword"
+
+        if '\t' in after:
+            yield match.start(2), "E273 tab after keyword"
+        elif len(after) > 1:
+            yield match.start(2), "E271 multiple spaces after keyword"
 
 
 def missing_whitespace(logical_line):
     """
     JCR: Each comma, semicolon or colon should be followed by whitespace.
+
+    Okay: [a, b]
+    Okay: (3,)
+    Okay: a[1:4]
+    Okay: a[:4]
+    Okay: a[1:]
+    Okay: a[1:4:2]
+    E231: ['a','b']
+    E231: foo(bar,baz)
+    E231: [{'a':'b'}]
     """
     line = logical_line
     for index in range(len(line) - 1):
         char = line[index]
-        if char in ',;:' and line[index + 1] != ' ':
+        if char in ',;:' and line[index + 1] not in WHITESPACE:
             before = line[:index]
-            if char == ':' and before.count('[') > before.count(']'):
-                continue # Slice syntax, no space required
-            return index, "E231 missing whitespace after '%s'" % char
+            if char == ':' and before.count('[') > before.count(']') and \
+                    before.rfind('{') < before.rfind('['):
+                continue  # Slice syntax, no space required
+            if char == ',' and line[index + 1] == ')':
+                continue  # Allow tuple with only one element: (3,)
+            yield index, "E231 missing whitespace after '%s'" % char
 
 
 def indentation(logical_line, previous_logical, indent_char,
                 indent_level, previous_indent_level):
-    """
+    r"""
     Use 4 spaces per indentation level.
 
     For really old code that you don't want to mess up, you can continue to
     use 8-space tabs.
+
+    Okay: a = 1
+    Okay: if a == 0:\n    a = 1
+    E111:   a = 1
+
+    Okay: for item in items:\n    pass
+    E112: for item in items:\npass
+
+    Okay: a = 1\nb = 2
+    E113: a = 1\n    b = 2
     """
     if indent_char == ' ' and indent_level % 4:
-        return 0, "E111 indentation is not a multiple of four"
+        yield 0, "E111 indentation is not a multiple of four"
     indent_expect = previous_logical.endswith(':')
     if indent_expect and indent_level <= previous_indent_level:
-        return 0, "E112 expected an indented block"
+        yield 0, "E112 expected an indented block"
     if indent_level > previous_indent_level and not indent_expect:
-        return 0, "E113 unexpected indentation"
+        yield 0, "E113 unexpected indentation"
+
+
+def continuation_line_indentation(logical_line, tokens, indent_level, verbose):
+    r"""
+    Continuation lines should align wrapped elements either vertically using
+    Python's implicit line joining inside parentheses, brackets and braces, or
+    using a hanging indent.
+
+    When using a hanging indent the following considerations should be applied:
+
+    - there should be no arguments on the first line, and
+
+    - further indentation should be used to clearly distinguish itself as a
+      continuation line.
+
+    Okay: a = (\n)
+    E123: a = (\n    )
+
+    Okay: a = (\n    42)
+    E121: a = (\n   42)
+    E122: a = (\n42)
+    E123: a = (\n    42\n    )
+    E124: a = (24,\n     42\n)
+    E125: if (a or\n    b):\n    pass
+    E126: a = (\n        42)
+    E127: a = (24,\n      42)
+    E128: a = (24,\n    42)
+    """
+    first_row = tokens[0][2][0]
+    nrows = 1 + tokens[-1][2][0] - first_row
+    if nrows == 1 or noqa(tokens[0][4]):
+        return
+
+    # indent_next tells us whether the next block is indented; assuming
+    # that it is indented by 4 spaces, then we should not allow 4-space
+    # indents on the final continuation line; in turn, some other
+    # indents are allowed to have an extra 4 spaces.
+    indent_next = logical_line.endswith(':')
+
+    row = depth = 0
+    # remember how many brackets were opened on each line
+    parens = [0] * nrows
+    # relative indents of physical lines
+    rel_indent = [0] * nrows
+    # visual indents
+    indent_chances = {}
+    last_indent = tokens[0][2]
+    indent = [last_indent[1]]
+    if verbose >= 3:
+        print(">>> " + tokens[0][4].rstrip())
+
+    for token_type, text, start, end, line in tokens:
+
+        newline = row < start[0] - first_row
+        if newline:
+            row = start[0] - first_row
+            newline = (not last_token_multiline and
+                       token_type not in (tokenize.NL, tokenize.NEWLINE))
+
+        if newline:
+            # this is the beginning of a continuation line.
+            last_indent = start
+            if verbose >= 3:
+                print("... " + line.rstrip())
+
+            # record the initial indent.
+            rel_indent[row] = expand_indent(line) - indent_level
+
+            if depth:
+                # a bracket expression in a continuation line.
+                # find the line that it was opened on
+                for open_row in range(row - 1, -1, -1):
+                    if parens[open_row]:
+                        break
+            else:
+                # an unbracketed continuation line (ie, backslash)
+                open_row = 0
+            hang = rel_indent[row] - rel_indent[open_row]
+            visual_indent = indent_chances.get(start[1])
+
+            if token_type == tokenize.OP and text in ']})':
+                # this line starts with a closing bracket
+                if indent[depth]:
+                    if start[1] != indent[depth]:
+                        yield (start, "E124 closing bracket does not match "
+                               "visual indentation")
+                elif hang:
+                    yield (start, "E123 closing bracket does not match "
+                           "indentation of opening bracket's line")
+            elif visual_indent is True:
+                # visual indent is verified
+                if not indent[depth]:
+                    indent[depth] = start[1]
+            elif visual_indent in (text, str):
+                # ignore token lined up with matching one from a previous line
+                pass
+            elif indent[depth] and start[1] < indent[depth]:
+                # visual indent is broken
+                yield (start, "E128 continuation line "
+                       "under-indented for visual indent")
+            elif hang == 4 or (indent_next and rel_indent[row] == 8):
+                # hanging indent is verified
+                pass
+            else:
+                # indent is broken
+                if hang <= 0:
+                    error = "E122", "missing indentation or outdented"
+                elif indent[depth]:
+                    error = "E127", "over-indented for visual indent"
+                elif hang % 4:
+                    error = "E121", "indentation is not a multiple of four"
+                else:
+                    error = "E126", "over-indented for hanging indent"
+                yield start, "%s continuation line %s" % error
+
+        # look for visual indenting
+        if (parens[row] and token_type not in (tokenize.NL, tokenize.COMMENT)
+                and not indent[depth]):
+            indent[depth] = start[1]
+            indent_chances[start[1]] = True
+            if verbose >= 4:
+                print("bracket depth %s indent to %s" % (depth, start[1]))
+        # deal with implicit string concatenation
+        elif (token_type in (tokenize.STRING, tokenize.COMMENT) or
+              text in ('u', 'ur', 'b', 'br')):
+            indent_chances[start[1]] = str
+        # special case for the "if" statement because len("if (") == 4
+        elif not indent_chances and not row and not depth and text == 'if':
+            indent_chances[end[1] + 1] = True
+
+        # keep track of bracket depth
+        if token_type == tokenize.OP:
+            if text in '([{':
+                depth += 1
+                indent.append(0)
+                parens[row] += 1
+                if verbose >= 4:
+                    print("bracket depth %s seen, col %s, visual min = %s" %
+                          (depth, start[1], indent[depth]))
+            elif text in ')]}' and depth > 0:
+                # parent indents should not be more than this one
+                prev_indent = indent.pop() or last_indent[1]
+                for d in range(depth):
+                    if indent[d] > prev_indent:
+                        indent[d] = 0
+                for ind in list(indent_chances):
+                    if ind >= prev_indent:
+                        del indent_chances[ind]
+                depth -= 1
+                if depth:
+                    indent_chances[indent[depth]] = True
+                for idx in range(row, -1, -1):
+                    if parens[idx]:
+                        parens[idx] -= 1
+                        break
+            assert len(indent) == depth + 1
+            if start[1] not in indent_chances:
+                # allow to line up tokens
+                indent_chances[start[1]] = text
+
+        last_token_multiline = (start[0] != end[0])
+
+    if indent_next and rel_indent[-1] == 4:
+        yield (last_indent, "E125 continuation line does not distinguish "
+               "itself from next logical line")
 
 
 def whitespace_before_parameters(logical_line, tokens):
@@ -255,6 +557,13 @@ def whitespace_before_parameters(logical_line, tokens):
 
     - Immediately before the open parenthesis that starts an indexing or
       slicing.
+
+    Okay: spam(1)
+    E211: spam (1)
+
+    Okay: dict['key'] = list[index]
+    E211: dict ['key'] = list[index]
+    E211: dict['key'] = list [index]
     """
     prev_type = tokens[0][0]
     prev_text = tokens[0][1]
@@ -264,97 +573,391 @@ def whitespace_before_parameters(logical_line, tokens):
         if (token_type == tokenize.OP and
             text in '([' and
             start != prev_end and
-            prev_type == tokenize.NAME and
+            (prev_type == tokenize.NAME or prev_text in '}])') and
+            # Syntax "class A (B):" is allowed, but avoid it
             (index < 2 or tokens[index - 2][1] != 'class') and
-            (not iskeyword(prev_text))):
-            return prev_end, "E211 whitespace before '%s'" % text
+                # Allow "return (a.foo for a in range(5))"
+                not keyword.iskeyword(prev_text)):
+            yield prev_end, "E211 whitespace before '%s'" % text
         prev_type = token_type
         prev_text = text
         prev_end = end
 
 
 def whitespace_around_operator(logical_line):
-    """
+    r"""
     Avoid extraneous whitespace in the following situations:
 
     - More than one space around an assignment (or other) operator to
       align it with another.
+
+    Okay: a = 12 + 3
+    E221: a = 4  + 5
+    E222: a = 4 +  5
+    E223: a = 4\t+ 5
+    E224: a = 4 +\t5
     """
-    line = logical_line
-    for operator in operators:
-        found = line.find('  ' + operator)
-        if found > -1:
-            return found, "E221 multiple spaces before operator"
-        found = line.find(operator + '  ')
-        if found > -1:
-            return found, "E222 multiple spaces after operator"
-        found = line.find('\t' + operator)
-        if found > -1:
-            return found, "E223 tab before operator"
-        found = line.find(operator + '\t')
-        if found > -1:
-            return found, "E224 tab after operator"
+    for match in OPERATOR_REGEX.finditer(logical_line):
+        before, after = match.groups()
+
+        if '\t' in before:
+            yield match.start(1), "E223 tab before operator"
+        elif len(before) > 1:
+            yield match.start(1), "E221 multiple spaces before operator"
+
+        if '\t' in after:
+            yield match.start(2), "E224 tab after operator"
+        elif len(after) > 1:
+            yield match.start(2), "E222 multiple spaces after operator"
+
+
+def missing_whitespace_around_operator(logical_line, tokens):
+    r"""
+    - Always surround these binary operators with a single space on
+      either side: assignment (=), augmented assignment (+=, -= etc.),
+      comparisons (==, <, >, !=, <>, <=, >=, in, not in, is, is not),
+      Booleans (and, or, not).
+
+    - Use spaces around arithmetic operators.
+
+    Okay: i = i + 1
+    Okay: submitted += 1
+    Okay: x = x * 2 - 1
+    Okay: hypot2 = x * x + y * y
+    Okay: c = (a + b) * (a - b)
+    Okay: foo(bar, key='word', *args, **kwargs)
+    Okay: alpha[:-i]
+
+    E225: i=i+1
+    E225: submitted +=1
+    E225: x = x /2 - 1
+    E225: z = x **y
+    E226: c = (a+b) * (a-b)
+    E226: hypot2 = x*x + y*y
+    E227: c = a|b
+    E228: msg = fmt%(errno, errmsg)
+    """
+    parens = 0
+    need_space = False
+    prev_type = tokenize.OP
+    prev_text = prev_end = None
+    for token_type, text, start, end, line in tokens:
+        if token_type in (tokenize.NL, tokenize.NEWLINE, tokenize.ERRORTOKEN):
+            # ERRORTOKEN is triggered by backticks in Python 3
+            continue
+        if text in ('(', 'lambda'):
+            parens += 1
+        elif text == ')':
+            parens -= 1
+        if need_space:
+            if start != prev_end:
+                # Found a (probably) needed space
+                if need_space is not True and not need_space[1]:
+                    yield (need_space[0],
+                           "E225 missing whitespace around operator")
+                need_space = False
+            elif text == '>' and prev_text in ('<', '-'):
+                # Tolerate the "<>" operator, even if running Python 3
+                # Deal with Python 3's annotated return value "->"
+                pass
+            else:
+                if need_space is True or need_space[1]:
+                    # A needed trailing space was not found
+                    yield prev_end, "E225 missing whitespace around operator"
+                else:
+                    code, optype = 'E226', 'arithmetic'
+                    if prev_text == '%':
+                        code, optype = 'E228', 'modulo'
+                    elif prev_text not in ARITHMETIC_OP:
+                        code, optype = 'E227', 'bitwise or shift'
+                    yield (need_space[0], "%s missing whitespace "
+                           "around %s operator" % (code, optype))
+                need_space = False
+        elif token_type == tokenize.OP and prev_end is not None:
+            if text == '=' and parens:
+                # Allow keyword args or defaults: foo(bar=None).
+                pass
+            elif text in WS_NEEDED_OPERATORS:
+                need_space = True
+            elif text in UNARY_OPERATORS:
+                # Check if the operator is being used as a binary operator
+                # Allow unary operators: -123, -x, +1.
+                # Allow argument unpacking: foo(*args, **kwargs).
+                if prev_type == tokenize.OP:
+                    binary_usage = (prev_text in '}])')
+                elif prev_type == tokenize.NAME:
+                    binary_usage = (prev_text not in KEYWORDS)
+                else:
+                    binary_usage = (prev_type not in SKIP_TOKENS)
+
+                if binary_usage:
+                    need_space = None
+            elif text in WS_OPTIONAL_OPERATORS:
+                need_space = None
+
+            if need_space is None:
+                # Surrounding space is optional, but ensure that
+                # trailing space matches opening space
+                need_space = (prev_end, start != prev_end)
+            elif need_space and start == prev_end:
+                # A needed opening space was not found
+                yield prev_end, "E225 missing whitespace around operator"
+                need_space = False
+        prev_type = token_type
+        prev_text = text
+        prev_end = end
 
 
 def whitespace_around_comma(logical_line):
-    """
+    r"""
     Avoid extraneous whitespace in the following situations:
 
     - More than one space around an assignment (or other) operator to
       align it with another.
 
-    JCR: This should also be applied around comma etc.
+    Note: these checks are disabled by default
+
+    Okay: a = (1, 2)
+    E241: a = (1,  2)
+    E242: a = (1,\t2)
     """
     line = logical_line
-    for separator in ',;:':
-        found = line.find(separator + '  ')
-        if found > -1:
-            return found + 1, "E241 multiple spaces after '%s'" % separator
-        found = line.find(separator + '\t')
-        if found > -1:
-            return found + 1, "E242 tab after '%s'" % separator
+    for m in WHITESPACE_AFTER_COMMA_REGEX.finditer(line):
+        found = m.start() + 1
+        if '\t' in m.group():
+            yield found, "E242 tab after '%s'" % m.group()[0]
+        else:
+            yield found, "E241 multiple spaces after '%s'" % m.group()[0]
 
 
-def imports_on_separate_lines(logical_line):
+def whitespace_around_named_parameter_equals(logical_line, tokens):
+    """
+    Don't use spaces around the '=' sign when used to indicate a
+    keyword argument or a default parameter value.
+
+    Okay: def complex(real, imag=0.0):
+    Okay: return magic(r=real, i=imag)
+    Okay: boolean(a == b)
+    Okay: boolean(a != b)
+    Okay: boolean(a <= b)
+    Okay: boolean(a >= b)
+
+    E251: def complex(real, imag = 0.0):
+    E251: return magic(r = real, i = imag)
+    """
+    parens = 0
+    no_space = False
+    prev_end = None
+    message = "E251 unexpected spaces around keyword / parameter equals"
+    for token_type, text, start, end, line in tokens:
+        if no_space:
+            no_space = False
+            if start != prev_end:
+                yield (prev_end, message)
+        elif token_type == tokenize.OP:
+            if text == '(':
+                parens += 1
+            elif text == ')':
+                parens -= 1
+            elif parens and text == '=':
+                no_space = True
+                if start != prev_end:
+                    yield (prev_end, message)
+        prev_end = end
+
+
+def whitespace_before_inline_comment(logical_line, tokens):
+    """
+    Separate inline comments by at least two spaces.
+
+    An inline comment is a comment on the same line as a statement.  Inline
+    comments should be separated by at least two spaces from the statement.
+    They should start with a # and a single space.
+
+    Okay: x = x + 1  # Increment x
+    Okay: x = x + 1    # Increment x
+    E261: x = x + 1 # Increment x
+    E262: x = x + 1  #Increment x
+    E262: x = x + 1  #  Increment x
     """
+    prev_end = (0, 0)
+    for token_type, text, start, end, line in tokens:
+        if token_type == tokenize.COMMENT:
+            if not line[:start[1]].strip():
+                continue
+            if prev_end[0] == start[0] and start[1] < prev_end[1] + 2:
+                yield (prev_end,
+                       "E261 at least two spaces before inline comment")
+            symbol, sp, comment = text.partition(' ')
+            if symbol not in ('#', '#:') or comment[:1].isspace():
+                yield start, "E262 inline comment should start with '# '"
+        elif token_type != tokenize.NL:
+            prev_end = end
+
+
+def imports_on_separate_lines(logical_line):
+    r"""
     Imports should usually be on separate lines.
+
+    Okay: import os\nimport sys
+    E401: import sys, os
+
+    Okay: from subprocess import Popen, PIPE
+    Okay: from myclas import MyClass
+    Okay: from foo.bar.yourclass import YourClass
+    Okay: import myclass
+    Okay: import foo.bar.yourclass
     """
     line = logical_line
     if line.startswith('import '):
         found = line.find(',')
-        if found > -1:
-            return found, "E401 multiple imports on one line"
+        if -1 < found and ';' not in line[:found]:
+            yield found, "E401 multiple imports on one line"
 
 
 def compound_statements(logical_line):
-    """
+    r"""
     Compound statements (multiple statements on the same line) are
     generally discouraged.
+
+    While sometimes it's okay to put an if/for/while with a small body
+    on the same line, never do this for multi-clause statements. Also
+    avoid folding such long lines!
+
+    Okay: if foo == 'blah':\n    do_blah_thing()
+    Okay: do_one()
+    Okay: do_two()
+    Okay: do_three()
+
+    E701: if foo == 'blah': do_blah_thing()
+    E701: for x in lst: total += x
+    E701: while t < 10: t = delay()
+    E701: if foo == 'blah': do_blah_thing()
+    E701: else: do_non_blah_thing()
+    E701: try: something()
+    E701: finally: cleanup()
+    E701: if foo == 'blah': one(); two(); three()
+
+    E702: do_one(); do_two(); do_three()
+    E703: do_four();  # useless semicolon
     """
     line = logical_line
+    last_char = len(line) - 1
     found = line.find(':')
-    if -1 < found < len(line) - 1:
+    if -1 < found < last_char:
         before = line[:found]
-        if (before.count('{') <= before.count('}') and # {'a': 1} (dict)
-            before.count('[') <= before.count(']') and # [1:2] (slice)
-            not re.search(r'\blambda\b', before)):     # lambda x: x
-            return found, "E701 multiple statements on one line (colon)"
+        if (before.count('{') <= before.count('}') and  # {'a': 1} (dict)
+            before.count('[') <= before.count(']') and  # [1:2] (slice)
+            before.count('(') <= before.count(')') and  # (Python 3 annotation)
+                not LAMBDA_REGEX.search(before)):       # lambda x: x
+            yield found, "E701 multiple statements on one line (colon)"
     found = line.find(';')
     if -1 < found:
-        return found, "E702 multiple statements on one line (semicolon)"
+        if found < last_char:
+            yield found, "E702 multiple statements on one line (semicolon)"
+        else:
+            yield found, "E703 statement ends with a semicolon"
 
 
-def python_3000_has_key(logical_line):
+def explicit_line_join(logical_line, tokens):
+    r"""
+    Avoid explicit line join between brackets.
+
+    The preferred way of wrapping long lines is by using Python's implied line
+    continuation inside parentheses, brackets and braces.  Long lines can be
+    broken over multiple lines by wrapping expressions in parentheses.  These
+    should be used in preference to using a backslash for line continuation.
+
+    E502: aaa = [123, \\n       123]
+    E502: aaa = ("bbb " \\n       "ccc")
+
+    Okay: aaa = [123,\n       123]
+    Okay: aaa = ("bbb "\n       "ccc")
+    Okay: aaa = "bbb " \\n    "ccc"
+    """
+    prev_start = prev_end = parens = 0
+    for token_type, text, start, end, line in tokens:
+        if start[0] != prev_start and parens and backslash:
+            yield backslash, "E502 the backslash is redundant between brackets"
+        if end[0] != prev_end:
+            if line.rstrip('\r\n').endswith('\\'):
+                backslash = (end[0], len(line.splitlines()[-1]) - 1)
+            else:
+                backslash = None
+            prev_start = prev_end = end[0]
+        else:
+            prev_start = start[0]
+        if token_type == tokenize.OP:
+            if text in '([{':
+                parens += 1
+            elif text in ')]}':
+                parens -= 1
+
+
+def comparison_to_singleton(logical_line):
+    """
+    Comparisons to singletons like None should always be done
+    with "is" or "is not", never the equality operators.
+
+    Okay: if arg is not None:
+    E711: if arg != None:
+    E712: if arg == True:
+
+    Also, beware of writing if x when you really mean if x is not None --
+    e.g. when testing whether a variable or argument that defaults to None was
+    set to some other value.  The other value might have a type (such as a
+    container) that could be false in a boolean context!
+    """
+    match = COMPARE_SINGLETON_REGEX.search(logical_line)
+    if match:
+        same = (match.group(1) == '==')
+        singleton = match.group(2)
+        msg = "'if cond is %s:'" % (('' if same else 'not ') + singleton)
+        if singleton in ('None',):
+            code = 'E711'
+        else:
+            code = 'E712'
+            nonzero = ((singleton == 'True' and same) or
+                       (singleton == 'False' and not same))
+            msg += " or 'if %scond:'" % ('' if nonzero else 'not ')
+        yield match.start(1), ("%s comparison to %s should be %s" %
+                               (code, singleton, msg))
+
+
+def comparison_type(logical_line):
     """
-    The {}.has_key() method will be removed in the future version of
-    Python. Use the 'in' operation instead, like:
-    d = {"a": 1, "b": 2}
-    if "b" in d:
-        print d["b"]
+    Object type comparisons should always use isinstance() instead of
+    comparing types directly.
+
+    Okay: if isinstance(obj, int):
+    E721: if type(obj) is type(1):
+
+    When checking if an object is a string, keep in mind that it might be a
+    unicode string too! In Python 2.3, str and unicode have a common base
+    class, basestring, so you can do:
+
+    Okay: if isinstance(obj, basestring):
+    Okay: if type(a1) is type(b1):
+    """
+    match = COMPARE_TYPE_REGEX.search(logical_line)
+    if match:
+        inst = match.group(1)
+        if inst and isidentifier(inst) and inst not in SINGLETONS:
+            return  # Allow comparison for types which are not obvious
+        yield match.start(), "E721 do not compare types, use 'isinstance()'"
+
+
+def python_3000_has_key(logical_line):
+    r"""
+    The {}.has_key() method is removed in the Python 3.
+    Use the 'in' operation instead.
+
+    Okay: if "alph" in d:\n    print d["alph"]
+    W601: assert d.has_key('alph')
     """
     pos = logical_line.find('.has_key(')
     if pos > -1:
-        return pos, "W601 .has_key() is deprecated, use 'in'"
+        yield pos, "W601 .has_key() is deprecated, use 'in'"
 
 
 def python_3000_raise_comma(logical_line):
@@ -365,11 +968,41 @@ def python_3000_raise_comma(logical_line):
     The paren-using form is preferred because when the exception arguments
     are long or include string formatting, you don't need to use line
     continuation characters thanks to the containing parentheses.  The older
-    form will be removed in Python 3000.
+    form is removed in Python 3.
+
+    Okay: raise DummyError("Message")
+    W602: raise DummyError, "Message"
     """
-    match = raise_comma_match(logical_line)
-    if match:
-        return match.start(1), "W602 deprecated form of raising exception"
+    match = RAISE_COMMA_REGEX.match(logical_line)
+    if match and not RERAISE_COMMA_REGEX.match(logical_line):
+        yield match.end() - 1, "W602 deprecated form of raising exception"
+
+
+def python_3000_not_equal(logical_line):
+    """
+    != can also be written <>, but this is an obsolete usage kept for
+    backwards compatibility only. New code should always use !=.
+    The older syntax is removed in Python 3.
+
+    Okay: if a != 'no':
+    W603: if a <> 'no':
+    """
+    pos = logical_line.find('<>')
+    if pos > -1:
+        yield pos, "W603 '<>' is deprecated, use '!='"
+
+
+def python_3000_backticks(logical_line):
+    """
+    Backticks are removed in Python 3.
+    Use repr() instead.
+
+    Okay: val = repr(1 + 2)
+    W604: val = `1 + 2`
+    """
+    pos = logical_line.find('`')
+    if pos > -1:
+        yield pos, "W604 backticks are deprecated, use 'repr()'"
 
 
 ##############################################################################
@@ -377,26 +1010,63 @@ def python_3000_raise_comma(logical_line):
 ##############################################################################
 
 
+if '' == ''.encode():
+    # Python 2: implicit encoding.
+    def readlines(filename):
+        f = open(filename)
+        try:
+            return f.readlines()
+        finally:
+            f.close()
+
+    isidentifier = re.compile(r'[a-zA-Z_]\w*').match
+    stdin_get_value = sys.stdin.read
+else:
+    # Python 3
+    def readlines(filename):
+        f = open(filename, 'rb')
+        try:
+            coding, lines = tokenize.detect_encoding(f.readline)
+            f = TextIOWrapper(f, coding, line_buffering=True)
+            return [l.decode(coding) for l in lines] + f.readlines()
+        except (LookupError, SyntaxError, UnicodeError):
+            f.close()
+            # Fall back if files are improperly declared
+            f = open(filename, encoding='latin-1')
+            return f.readlines()
+        finally:
+            f.close()
+
+    isidentifier = str.isidentifier
+
+    def stdin_get_value():
+        return TextIOWrapper(sys.stdin.buffer, errors='ignore').read()
+readlines.__doc__ = "    Read the source code."
+noqa = re.compile(r'# no(?:qa|pep8)\b', re.I).search
+
+
 def expand_indent(line):
-    """
+    r"""
     Return the amount of indentation.
     Tabs are expanded to the next multiple of 8.
 
     >>> expand_indent('    ')
     4
-    >>> expand_indent('\\t')
+    >>> expand_indent('\t')
     8
-    >>> expand_indent('    \\t')
+    >>> expand_indent('    \t')
     8
-    >>> expand_indent('       \\t')
+    >>> expand_indent('       \t')
     8
-    >>> expand_indent('        \\t')
+    >>> expand_indent('        \t')
     16
     """
+    if '\t' not in line:
+        return len(line) - len(line.lstrip())
     result = 0
     for char in line:
         if char == '\t':
-            result = result / 8 * 8 + 8
+            result = result // 8 * 8 + 8
         elif char == ' ':
             result += 1
         else:
@@ -404,34 +1074,6 @@ def expand_indent(line):
     return result
 
 
-##############################################################################
-# Framework to run all checks
-##############################################################################
-
-
-def message(text):
-    """Print a message."""
-    # print >> sys.stderr, options.prog + ': ' + text
-    # print >> sys.stderr, text
-    print text
-
-
-def find_checks(argument_name):
-    """
-    Find all globally visible functions where the first argument name
-    starts with argument_name.
-    """
-    checks = []
-    function_type = type(find_checks)
-    for name, function in globals().iteritems():
-        if type(function) is function_type:
-            args = inspect.getargspec(function)[0]
-            if len(args) >= 1 and args[0].startswith(argument_name):
-                checks.append((name, function, args))
-    checks.sort()
-    return checks
-
-
 def mute_string(text):
     """
     Replace contents with 'xxx' to prevent syntax matching.
@@ -443,32 +1085,135 @@ def mute_string(text):
     >>> mute_string("r'abc'")
     "r'xxx'"
     """
-    start = 1
-    end = len(text) - 1
     # String modifiers (e.g. u or r)
-    if text.endswith('"'):
-        start += text.index('"')
-    elif text.endswith("'"):
-        start += text.index("'")
+    start = text.index(text[-1]) + 1
+    end = len(text) - 1
     # Triple quotes
-    if text.endswith('"""') or text.endswith("'''"):
+    if text[-3:] in ('"""', "'''"):
         start += 2
         end -= 2
     return text[:start] + 'x' * (end - start) + text[end:]
 
 
-class Checker:
+def parse_udiff(diff, patterns=None, parent='.'):
+    """Return a dictionary of matching lines."""
+    # For each file of the diff, the entry key is the filename,
+    # and the value is a set of row numbers to consider.
+    rv = {}
+    path = nrows = None
+    for line in diff.splitlines():
+        if nrows:
+            if line[:1] != '-':
+                nrows -= 1
+            continue
+        if line[:3] == '@@ ':
+            hunk_match = HUNK_REGEX.match(line)
+            row, nrows = [int(g or '1') for g in hunk_match.groups()]
+            rv[path].update(range(row, row + nrows))
+        elif line[:3] == '+++':
+            path = line[4:].split('\t', 1)[0]
+            if path[:2] == 'b/':
+                path = path[2:]
+            rv[path] = set()
+    return dict([(os.path.join(parent, path), rows)
+                 for (path, rows) in rv.items()
+                 if rows and filename_match(path, patterns)])
+
+
+def filename_match(filename, patterns, default=True):
+    """
+    Check if patterns contains a pattern that matches filename.
+    If patterns is unspecified, this always returns True.
+    """
+    if not patterns:
+        return default
+    return any(fnmatch(filename, pattern) for pattern in patterns)
+
+
+##############################################################################
+# Framework to run all checks
+##############################################################################
+
+
+_checks = {'physical_line': {}, 'logical_line': {}, 'tree': {}}
+
+
+def register_check(check, codes=None):
+    """
+    Register a new check object.
+    """
+    def _add_check(check, kind, codes, args):
+        if check in _checks[kind]:
+            _checks[kind][check][0].extend(codes or [])
+        else:
+            _checks[kind][check] = (codes or [''], args)
+    if inspect.isfunction(check):
+        args = inspect.getargspec(check)[0]
+        if args and args[0] in ('physical_line', 'logical_line'):
+            if codes is None:
+                codes = ERRORCODE_REGEX.findall(check.__doc__ or '')
+            _add_check(check, args[0], codes, args)
+    elif inspect.isclass(check):
+        if inspect.getargspec(check.__init__)[0][:2] == ['self', 'tree']:
+            _add_check(check, 'tree', codes, None)
+
+
+def init_checks_registry():
+    """
+    Register all globally visible functions where the first argument name
+    is 'physical_line' or 'logical_line'.
+    """
+    mod = inspect.getmodule(register_check)
+    for (name, function) in inspect.getmembers(mod, inspect.isfunction):
+        register_check(function)
+init_checks_registry()
+
+
+class Checker(object):
     """
     Load a Python source file, tokenize it, check coding style.
     """
 
-    def __init__(self, filename):
+    def __init__(self, filename=None, lines=None,
+                 options=None, report=None, **kwargs):
+        if options is None:
+            options = StyleGuide(kwargs).options
+        else:
+            assert not kwargs
+        self._io_error = None
+        self._physical_checks = options.physical_checks
+        self._logical_checks = options.logical_checks
+        self._ast_checks = options.ast_checks
+        self.max_line_length = options.max_line_length
+        self.verbose = options.verbose
         self.filename = filename
-        self.lines = file(filename).readlines()
-        self.physical_checks = find_checks('physical_line')
-        self.logical_checks = find_checks('logical_line')
-        options.counters['physical lines'] = \
-            options.counters.get('physical lines', 0) + len(self.lines)
+        if filename is None:
+            self.filename = 'stdin'
+            self.lines = lines or []
+        elif filename == '-':
+            self.filename = 'stdin'
+            self.lines = stdin_get_value().splitlines(True)
+        elif lines is None:
+            try:
+                self.lines = readlines(filename)
+            except IOError:
+                exc_type, exc = sys.exc_info()[:2]
+                self._io_error = '%s: %s' % (exc_type.__name__, exc)
+                self.lines = []
+        else:
+            self.lines = lines
+        self.report = report or options.report
+        self.report_error = self.report.error
+
+    def report_invalid_syntax(self):
+        exc_type, exc = sys.exc_info()[:2]
+        offset = exc.args[1]
+        if len(offset) > 2:
+            offset = offset[1:3]
+        self.report_error(offset[0], offset[1] or 0,
+                          'E901 %s: %s' % (exc_type.__name__, exc.args[0]),
+                          self.report_invalid_syntax)
+    report_invalid_syntax.__doc__ = "    Check if the syntax is valid."
 
     def readline(self):
         """
@@ -503,9 +1248,9 @@ class Checker:
         Run all physical checks on a raw input line.
         """
         self.physical_line = line
-        if self.indent_char is None and len(line) and line[0] in ' \t':
+        if self.indent_char is None and line[:1] in WHITESPACE:
             self.indent_char = line[0]
-        for name, check, argument_names in self.physical_checks:
+        for name, check, argument_names in self._physical_checks:
             result = self.run_check(check, argument_names)
             if result is not None:
                 offset, text = result
@@ -521,21 +1266,21 @@ class Checker:
         previous = None
         for token in self.tokens:
             token_type, text = token[0:2]
-            if token_type in (tokenize.COMMENT, tokenize.NL,
-                              tokenize.INDENT, tokenize.DEDENT,
-                              tokenize.NEWLINE):
+            if token_type in SKIP_TOKENS:
                 continue
             if token_type == tokenize.STRING:
                 text = mute_string(text)
             if previous:
-                end_line, end = previous[3]
-                start_line, start = token[2]
-                if end_line != start_line: # different row
-                    if self.lines[end_line - 1][end - 1] not in '{[(':
+                end_row, end = previous[3]
+                start_row, start = token[2]
+                if end_row != start_row:    # different row
+                    prev_text = self.lines[end_row - 1][end - 1]
+                    if prev_text == ',' or (prev_text not in '{[('
+                                            and text not in '}])'):
                         logical.append(' ')
                         length += 1
-                elif end != start: # different column
-                    fill = self.lines[end_line - 1][end:start]
+                elif end != start:  # different column
+                    fill = self.lines[end_row - 1][end:start]
                     logical.append(fill)
                     length += len(fill)
             self.mapping.append((length, token))
@@ -543,300 +1288,573 @@ class Checker:
             length += len(text)
             previous = token
         self.logical_line = ''.join(logical)
-        assert self.logical_line.lstrip() == self.logical_line
-        assert self.logical_line.rstrip() == self.logical_line
+        # With Python 2, if the line ends with '\r\r\n' the assertion fails
+        # assert self.logical_line.strip() == self.logical_line
 
     def check_logical(self):
         """
         Build a line from tokens and run all logical checks on it.
         """
-        options.counters['logical lines'] = \
-            options.counters.get('logical lines', 0) + 1
         self.build_tokens_line()
+        self.report.increment_logical_line()
         first_line = self.lines[self.mapping[0][1][2][0] - 1]
         indent = first_line[:self.mapping[0][1][2][1]]
         self.previous_indent_level = self.indent_level
         self.indent_level = expand_indent(indent)
-        if options.verbose >= 2:
-            print self.logical_line[:80].rstrip()
-        for name, check, argument_names in self.logical_checks:
-            if options.verbose >= 3:
-                print '   ', name
-            result = self.run_check(check, argument_names)
-            if result is not None:
+        if self.verbose >= 2:
+            print(self.logical_line[:80].rstrip())
+        for name, check, argument_names in self._logical_checks:
+            if self.verbose >= 4:
+                print('   ' + name)
+            for result in self.run_check(check, argument_names):
                 offset, text = result
-                if type(offset) is tuple:
-                    original_number, original_offset = offset
+                if isinstance(offset, tuple):
+                    orig_number, orig_offset = offset
                 else:
                     for token_offset, token in self.mapping:
                         if offset >= token_offset:
-                            original_number = token[2][0]
-                            original_offset = (token[2][1]
-                                               + offset - token_offset)
-                self.report_error(original_number, original_offset,
-                                  text, check)
+                            orig_number = token[2][0]
+                            orig_offset = (token[2][1] + offset - token_offset)
+                self.report_error(orig_number, orig_offset, text, check)
         self.previous_logical = self.logical_line
 
-    def check_all(self):
+    def check_ast(self):
+        try:
+            tree = compile(''.join(self.lines), '', 'exec', PyCF_ONLY_AST)
+        except SyntaxError:
+            return self.report_invalid_syntax()
+        for name, cls, _ in self._ast_checks:
+            checker = cls(tree, self.filename)
+            for lineno, offset, text, check in checker.run():
+                if not noqa(self.lines[lineno - 1]):
+                    self.report_error(lineno, offset, text, check)
+
+    def generate_tokens(self):
+        if self._io_error:
+            self.report_error(1, 0, 'E902 %s' % self._io_error, readlines)
+        tokengen = tokenize.generate_tokens(self.readline_check_physical)
+        try:
+            for token in tokengen:
+                yield token
+        except (SyntaxError, tokenize.TokenError):
+            self.report_invalid_syntax()
+
+    def check_all(self, expected=None, line_offset=0):
         """
         Run all checks on the input file.
         """
-        self.file_errors = 0
+        self.report.init_file(self.filename, self.lines, expected, line_offset)
+        if self._ast_checks:
+            self.check_ast()
         self.line_number = 0
         self.indent_char = None
         self.indent_level = 0
         self.previous_logical = ''
-        self.blank_lines = 0
         self.tokens = []
+        self.blank_lines = blank_lines_before_comment = 0
         parens = 0
-        for token in tokenize.generate_tokens(self.readline_check_physical):
-            # print tokenize.tok_name[token[0]], repr(token)
+        for token in self.generate_tokens():
             self.tokens.append(token)
             token_type, text = token[0:2]
-            if token_type == tokenize.OP and text in '([{':
-                parens += 1
-            if token_type == tokenize.OP and text in '}])':
-                parens -= 1
-            if token_type == tokenize.NEWLINE and not parens:
-                self.check_logical()
-                self.blank_lines = 0
-                self.tokens = []
-            if token_type == tokenize.NL and not parens:
-                self.blank_lines += 1
-                self.tokens = []
-            if token_type == tokenize.COMMENT:
-                source_line = token[4]
-                token_start = token[2][1]
-                if source_line[:token_start].strip() == '':
+            if self.verbose >= 3:
+                if token[2][0] == token[3][0]:
+                    pos = '[%s:%s]' % (token[2][1] or '', token[3][1])
+                else:
+                    pos = 'l.%s' % token[3][0]
+                print('l.%s\t%s\t%s\t%r' %
+                      (token[2][0], pos, tokenize.tok_name[token[0]], text))
+            if token_type == tokenize.OP:
+                if text in '([{':
+                    parens += 1
+                elif text in '}])':
+                    parens -= 1
+            elif not parens:
+                if token_type == tokenize.NEWLINE:
+                    if self.blank_lines < blank_lines_before_comment:
+                        self.blank_lines = blank_lines_before_comment
+                    self.check_logical()
+                    self.tokens = []
+                    self.blank_lines = blank_lines_before_comment = 0
+                elif token_type == tokenize.NL:
+                    if len(self.tokens) == 1:
+                        # The physical line contains only this token.
+                        self.blank_lines += 1
+                    self.tokens = []
+                elif token_type == tokenize.COMMENT and len(self.tokens) == 1:
+                    if blank_lines_before_comment < self.blank_lines:
+                        blank_lines_before_comment = self.blank_lines
                     self.blank_lines = 0
-        return self.file_errors
+                    if COMMENT_WITH_NL:
+                        # The comment also ends a physical line
+                        self.tokens = []
+        return self.report.get_file_results()
+
+
+class BaseReport(object):
+    """Collect the results of the checks."""
+    print_filename = False
+
+    def __init__(self, options):
+        self._benchmark_keys = options.benchmark_keys
+        self._ignore_code = options.ignore_code
+        # Results
+        self.elapsed = 0
+        self.total_errors = 0
+        self.counters = dict.fromkeys(self._benchmark_keys, 0)
+        self.messages = {}
+
+    def start(self):
+        """Start the timer."""
+        self._start_time = time.time()
+
+    def stop(self):
+        """Stop the timer."""
+        self.elapsed = time.time() - self._start_time
+
+    def init_file(self, filename, lines, expected, line_offset):
+        """Signal a new file."""
+        self.filename = filename
+        self.lines = lines
+        self.expected = expected or ()
+        self.line_offset = line_offset
+        self.file_errors = 0
+        self.counters['files'] += 1
+        self.counters['physical lines'] += len(lines)
 
-    def report_error(self, line_number, offset, text, check):
-        """
-        Report an error, according to options.
-        """
-        if options.quiet == 1 and not self.file_errors:
-            message(self.filename)
-        self.file_errors += 1
+    def increment_logical_line(self):
+        """Signal a new logical line."""
+        self.counters['logical lines'] += 1
+
+    def error(self, line_number, offset, text, check):
+        """Report an error, according to options."""
         code = text[:4]
-        options.counters[code] = options.counters.get(code, 0) + 1
-        options.messages[code] = text[5:]
-        if options.quiet:
+        if self._ignore_code(code):
             return
-        if options.testsuite:
-            base = os.path.basename(self.filename)[:4]
-            if base == code:
-                return
-            if base[0] == 'E' and code[0] == 'W':
-                return
-        if ignore_code(code):
+        if code in self.counters:
+            self.counters[code] += 1
+        else:
+            self.counters[code] = 1
+            self.messages[code] = text[5:]
+        # Don't care about expected errors or warnings
+        if code in self.expected:
             return
-        if options.counters[code] == 1 or options.repeat:
-            message("%s:%s:%d: %s" %
-                    (self.filename, line_number, offset + 1, text))
-            if options.show_source:
-                line = self.lines[line_number - 1]
-                message(line.rstrip())
-                message(' ' * offset + '^')
-            if options.show_pep8:
-                message(check.__doc__.lstrip('\n').rstrip())
-
-
-def input_file(filename):
-    """
-    Run all checks on a Python source file.
-    """
-    if excluded(filename) or not filename_match(filename):
-        return 0
-    if options.verbose:
-        message('checking ' + filename)
-    options.counters['files'] = options.counters.get('files', 0) + 1
-    errors = Checker(filename).check_all()
-    if options.testsuite and not errors:
-        message("%s: %s" % (filename, "no errors found"))
-    return errors
-
-
-def input_dir(dirname):
-    """
-    Check all Python source files in this directory and all subdirectories.
-    """
-    dirname = dirname.rstrip('/')
-    if excluded(dirname):
-        return 0
-    errors = 0
-    for root, dirs, files in os.walk(dirname):
-        if options.verbose:
-            message('directory ' + root)
-        options.counters['directories'] = \
-            options.counters.get('directories', 0) + 1
-        dirs.sort()
-        for subdir in dirs:
-            if excluded(subdir):
-                dirs.remove(subdir)
-        files.sort()
-        for filename in files:
-            errors += input_file(os.path.join(root, filename))
-    return errors
-
-
-def excluded(filename):
-    """
-    Check if options.exclude contains a pattern that matches filename.
-    """
-    basename = os.path.basename(filename)
-    for pattern in options.exclude:
-        if fnmatch(basename, pattern):
-            # print basename, 'excluded because it matches', pattern
-            return True
-
-
-def filename_match(filename):
-    """
-    Check if options.filename contains a pattern that matches filename.
-    If options.filename is unspecified, this always returns True.
-    """
-    if not options.filename:
-        return True
-    for pattern in options.filename:
-        if fnmatch(filename, pattern):
-            return True
-
-
-def ignore_code(code):
-    """
-    Check if options.ignore contains a prefix of the error code.
-    """
-    for ignore in options.ignore:
-        if code.startswith(ignore):
-            return True
-
-
-def get_error_statistics():
-    """Get error statistics."""
-    return get_statistics("E")
+        if self.print_filename and not self.file_errors:
+            print(self.filename)
+        self.file_errors += 1
+        self.total_errors += 1
+        return code
 
+    def get_file_results(self):
+        """Return the count of errors and warnings for this file."""
+        return self.file_errors
 
-def get_warning_statistics():
-    """Get warning statistics."""
-    return get_statistics("W")
+    def get_count(self, prefix=''):
+        """Return the total count of errors and warnings."""
+        return sum([self.counters[key]
+                    for key in self.messages if key.startswith(prefix)])
 
+    def get_statistics(self, prefix=''):
+        """
+        Get statistics for message codes that start with the prefix.
 
-def get_statistics(prefix=''):
-    """
-    Get statistics for message codes that start with the prefix.
+        prefix='' matches all errors and warnings
+        prefix='E' matches all errors
+        prefix='W' matches all warnings
+        prefix='E4' matches all errors that have to do with imports
+        """
+        return ['%-7s %s %s' % (self.counters[key], key, self.messages[key])
+                for key in sorted(self.messages) if key.startswith(prefix)]
+
+    def print_statistics(self, prefix=''):
+        """Print overall statistics (number of errors and warnings)."""
+        for line in self.get_statistics(prefix):
+            print(line)
+
+    def print_benchmark(self):
+        """Print benchmark numbers."""
+        print('%-7.2f %s' % (self.elapsed, 'seconds elapsed'))
+        if self.elapsed:
+            for key in self._benchmark_keys:
+                print('%-7d %s per second (%d total)' %
+                      (self.counters[key] / self.elapsed, key,
+                       self.counters[key]))
+
+
+class FileReport(BaseReport):
+    """Collect the results of the checks and print only the filenames."""
+    print_filename = True
+
+
+class StandardReport(BaseReport):
+    """Collect and print the results of the checks."""
+
+    def __init__(self, options):
+        super(StandardReport, self).__init__(options)
+        self._fmt = REPORT_FORMAT.get(options.format.lower(),
+                                      options.format)
+        self._repeat = options.repeat
+        self._show_source = options.show_source
+        self._show_pep8 = options.show_pep8
+
+    def init_file(self, filename, lines, expected, line_offset):
+        """Signal a new file."""
+        self._deferred_print = []
+        return super(StandardReport, self).init_file(
+            filename, lines, expected, line_offset)
+
+    def error(self, line_number, offset, text, check):
+        """Report an error, according to options."""
+        code = super(StandardReport, self).error(line_number, offset,
+                                                 text, check)
+        if code and (self.counters[code] == 1 or self._repeat):
+            self._deferred_print.append(
+                (line_number, offset, code, text[5:], check.__doc__))
+        return code
+
+    def get_file_results(self):
+        """Print the result and return the overall count for this file."""
+        self._deferred_print.sort()
+        for line_number, offset, code, text, doc in self._deferred_print:
+            print(self._fmt % {
+                'path': self.filename,
+                'row': self.line_offset + line_number, 'col': offset + 1,
+                'code': code, 'text': text,
+            })
+            if self._show_source:
+                if line_number > len(self.lines):
+                    line = ''
+                else:
+                    line = self.lines[line_number - 1]
+                print(line.rstrip())
+                print(' ' * offset + '^')
+            if self._show_pep8 and doc:
+                print(doc.lstrip('\n').rstrip())
+        return self.file_errors
 
-    prefix='' matches all errors and warnings
-    prefix='E' matches all errors
-    prefix='W' matches all warnings
-    prefix='E4' matches all errors that have to do with imports
-    """
-    stats = []
-    keys = options.messages.keys()
-    keys.sort()
-    for key in keys:
-        if key.startswith(prefix):
-            stats.append('%-7s %s %s' %
-                         (options.counters[key], key, options.messages[key]))
-    return stats
 
+class DiffReport(StandardReport):
+    """Collect and print the results for the changed lines only."""
 
-def print_statistics(prefix=''):
-    """Print overall statistics (number of errors and warnings)."""
-    for line in get_statistics(prefix):
-        print line
+    def __init__(self, options):
+        super(DiffReport, self).__init__(options)
+        self._selected = options.selected_lines
 
+    def error(self, line_number, offset, text, check):
+        if line_number not in self._selected[self.filename]:
+            return
+        return super(DiffReport, self).error(line_number, offset, text, check)
+
+
+class StyleGuide(object):
+    """Initialize a PEP-8 instance with few options."""
+
+    def __init__(self, *args, **kwargs):
+        # build options from the command line
+        self.checker_class = kwargs.pop('checker_class', Checker)
+        parse_argv = kwargs.pop('parse_argv', False)
+        config_file = kwargs.pop('config_file', None)
+        parser = kwargs.pop('parser', None)
+        options, self.paths = process_options(
+            parse_argv=parse_argv, config_file=config_file, parser=parser)
+        if args or kwargs:
+            # build options from dict
+            options_dict = dict(*args, **kwargs)
+            options.__dict__.update(options_dict)
+            if 'paths' in options_dict:
+                self.paths = options_dict['paths']
+
+        self.runner = self.input_file
+        self.options = options
+
+        if not options.reporter:
+            options.reporter = BaseReport if options.quiet else StandardReport
+
+        for index, value in enumerate(options.exclude):
+            options.exclude[index] = value.rstrip('/')
+        options.select = tuple(options.select or ())
+        if not (options.select or options.ignore or
+                options.testsuite or options.doctest) and DEFAULT_IGNORE:
+            # The default choice: ignore controversial checks
+            options.ignore = tuple(DEFAULT_IGNORE.split(','))
+        else:
+            # Ignore all checks which are not explicitly selected
+            options.ignore = tuple(options.ignore or options.select and ('',))
+        options.benchmark_keys = BENCHMARK_KEYS[:]
+        options.ignore_code = self.ignore_code
+        options.physical_checks = self.get_checks('physical_line')
+        options.logical_checks = self.get_checks('logical_line')
+        options.ast_checks = self.get_checks('tree')
+        self.init_report()
+
+    def init_report(self, reporter=None):
+        """Initialize the report instance."""
+        self.options.report = (reporter or self.options.reporter)(self.options)
+        return self.options.report
+
+    def check_files(self, paths=None):
+        """Run all checks on the paths."""
+        if paths is None:
+            paths = self.paths
+        report = self.options.report
+        runner = self.runner
+        report.start()
+        try:
+            for path in paths:
+                if os.path.isdir(path):
+                    self.input_dir(path)
+                elif not self.excluded(path):
+                    runner(path)
+        except KeyboardInterrupt:
+            print('... stopped')
+        report.stop()
+        return report
+
+    def input_file(self, filename, lines=None, expected=None, line_offset=0):
+        """Run all checks on a Python source file."""
+        if self.options.verbose:
+            print('checking %s' % filename)
+        fchecker = self.checker_class(
+            filename, lines=lines, options=self.options)
+        return fchecker.check_all(expected=expected, line_offset=line_offset)
+
+    def input_dir(self, dirname):
+        """Check all files in this directory and all subdirectories."""
+        dirname = dirname.rstrip('/')
+        if self.excluded(dirname):
+            return 0
+        counters = self.options.report.counters
+        verbose = self.options.verbose
+        filepatterns = self.options.filename
+        runner = self.runner
+        for root, dirs, files in os.walk(dirname):
+            if verbose:
+                print('directory ' + root)
+            counters['directories'] += 1
+            for subdir in sorted(dirs):
+                if self.excluded(os.path.join(root, subdir)):
+                    dirs.remove(subdir)
+            for filename in sorted(files):
+                # contain a pattern that matches?
+                if ((filename_match(filename, filepatterns) and
+                     not self.excluded(filename))):
+                    runner(os.path.join(root, filename))
+
+    def excluded(self, filename):
+        """
+        Check if options.exclude contains a pattern that matches filename.
+        """
+        basename = os.path.basename(filename)
+        return any((filename_match(filename, self.options.exclude,
+                                   default=False),
+                    filename_match(basename, self.options.exclude,
+                                   default=False)))
 
-def print_benchmark(elapsed):
-    """
-    Print benchmark numbers.
-    """
-    print '%-7.2f %s' % (elapsed, 'seconds elapsed')
-    keys = ['directories', 'files',
-            'logical lines', 'physical lines']
-    for key in keys:
-        if key in options.counters:
-            print '%-7d %s per second (%d total)' % (
-                options.counters[key] / elapsed, key,
-                options.counters[key])
+    def ignore_code(self, code):
+        """
+        Check if the error code should be ignored.
 
+        If 'options.select' contains a prefix of the error code,
+        return False.  Else, if 'options.ignore' contains a prefix of
+        the error code, return True.
+        """
+        return (code.startswith(self.options.ignore) and
+                not code.startswith(self.options.select))
 
-def process_options(arglist=None):
-    """
-    Process options passed either via arglist or via command line args.
-    """
-    global options, args
-    usage = "%prog [options] input ..."
-    parser = OptionParser(usage)
+    def get_checks(self, argument_name):
+        """
+        Find all globally visible functions where the first argument name
+        starts with argument_name and which contain selected tests.
+        """
+        checks = []
+        for check, attrs in _checks[argument_name].items():
+            (codes, args) = attrs
+            if any(not (code and self.ignore_code(code)) for code in codes):
+                checks.append((check.__name__, check, args))
+        return sorted(checks)
+
+
+def get_parser(prog='pep8', version=__version__):
+    parser = OptionParser(prog=prog, version=version,
+                          usage="%prog [options] input ...")
+    parser.config_options = [
+        'exclude', 'filename', 'select', 'ignore', 'max-line-length', 'count',
+        'format', 'quiet', 'show-pep8', 'show-source', 'statistics', 'verbose']
     parser.add_option('-v', '--verbose', default=0, action='count',
                       help="print status messages, or debug with -vv")
     parser.add_option('-q', '--quiet', default=0, action='count',
                       help="report only file names, or nothing with -qq")
-    parser.add_option('--exclude', metavar='patterns', default=default_exclude,
-                      help="skip matches (default %s)" % default_exclude)
-    parser.add_option('--filename', metavar='patterns',
-                      help="only check matching files (e.g. *.py)")
+    parser.add_option('-r', '--repeat', default=True, action='store_true',
+                      help="(obsolete) show all occurrences of the same error")
+    parser.add_option('--first', action='store_false', dest='repeat',
+                      help="show first occurrence of each error")
+    parser.add_option('--exclude', metavar='patterns', default=DEFAULT_EXCLUDE,
+                      help="exclude files or directories which match these "
+                           "comma separated patterns (default: %default)")
+    parser.add_option('--filename', metavar='patterns', default='*.py',
+                      help="when parsing directories, only check filenames "
+                           "matching these comma separated patterns "
+                           "(default: %default)")
+    parser.add_option('--select', metavar='errors', default='',
+                      help="select errors and warnings (e.g. E,W6)")
     parser.add_option('--ignore', metavar='errors', default='',
                       help="skip errors and warnings (e.g. E4,W)")
-    parser.add_option('--repeat', action='store_true',
-                      help="show all occurrences of the same error")
     parser.add_option('--show-source', action='store_true',
                       help="show source code for each error")
     parser.add_option('--show-pep8', action='store_true',
-                      help="show text of PEP 8 for each error")
+                      help="show text of PEP 8 for each error "
+                           "(implies --first)")
     parser.add_option('--statistics', action='store_true',
                       help="count errors and warnings")
-    parser.add_option('--benchmark', action='store_true',
-                      help="measure processing speed")
-    parser.add_option('--testsuite', metavar='dir',
-                      help="run regression tests from dir")
-    parser.add_option('--doctest', action='store_true',
-                      help="run doctest on myself")
+    parser.add_option('--count', action='store_true',
+                      help="print total number of errors and warnings "
+                           "to standard error and set exit code to 1 if "
+                           "total is not null")
+    parser.add_option('--max-line-length', type='int', metavar='n',
+                      default=MAX_LINE_LENGTH,
+                      help="set maximum allowed line length "
+                           "(default: %default)")
+    parser.add_option('--format', metavar='format', default='default',
+                      help="set the error format [default|pylint|<custom>]")
+    parser.add_option('--diff', action='store_true',
+                      help="report only lines changed according to the "
+                           "unified diff received on STDIN")
+    group = parser.add_option_group("Testing Options")
+    if os.path.exists(TESTSUITE_PATH):
+        group.add_option('--testsuite', metavar='dir',
+                         help="run regression tests from dir")
+        group.add_option('--doctest', action='store_true',
+                         help="run doctest on myself")
+    group.add_option('--benchmark', action='store_true',
+                     help="measure processing speed")
+    return parser
+
+
+def read_config(options, args, arglist, parser):
+    """Read both user configuration and local configuration."""
+    config = RawConfigParser()
+
+    user_conf = options.config
+    if user_conf and os.path.isfile(user_conf):
+        if options.verbose:
+            print('user configuration: %s' % user_conf)
+        config.read(user_conf)
+
+    parent = tail = args and os.path.abspath(os.path.commonprefix(args))
+    while tail:
+        for name in PROJECT_CONFIG:
+            local_conf = os.path.join(parent, name)
+            if os.path.isfile(local_conf):
+                break
+        else:
+            parent, tail = os.path.split(parent)
+            continue
+        if options.verbose:
+            print('local configuration: %s' % local_conf)
+        config.read(local_conf)
+        break
+
+    pep8_section = parser.prog
+    if config.has_section(pep8_section):
+        option_list = dict([(o.dest, o.type or o.action)
+                            for o in parser.option_list])
+
+        # First, read the default values
+        new_options, _ = parser.parse_args([])
+
+        # Second, parse the configuration
+        for opt in config.options(pep8_section):
+            if options.verbose > 1:
+                print("  %s = %s" % (opt, config.get(pep8_section, opt)))
+            if opt.replace('_', '-') not in parser.config_options:
+                print("Unknown option: '%s'\n  not in [%s]" %
+                      (opt, ' '.join(parser.config_options)))
+                sys.exit(1)
+            normalized_opt = opt.replace('-', '_')
+            opt_type = option_list[normalized_opt]
+            if opt_type in ('int', 'count'):
+                value = config.getint(pep8_section, opt)
+            elif opt_type == 'string':
+                value = config.get(pep8_section, opt)
+            else:
+                assert opt_type in ('store_true', 'store_false')
+                value = config.getboolean(pep8_section, opt)
+            setattr(new_options, normalized_opt, value)
+
+        # Third, overwrite with the command-line options
+        options, _ = parser.parse_args(arglist, values=new_options)
+    options.doctest = options.testsuite = False
+    return options
+
+
+def process_options(arglist=None, parse_argv=False, config_file=None,
+                    parser=None):
+    """Process options passed either via arglist or via command line args."""
+    if not arglist and not parse_argv:
+        # Don't read the command line if the module is used as a library.
+        arglist = []
+    if not parser:
+        parser = get_parser()
+    if not parser.has_option('--config'):
+        if config_file is True:
+            config_file = DEFAULT_CONFIG
+        group = parser.add_option_group("Configuration", description=(
+            "The project options are read from the [%s] section of the "
+            "tox.ini file or the setup.cfg file located in any parent folder "
+            "of the path(s) being processed.  Allowed options are: %s." %
+            (parser.prog, ', '.join(parser.config_options))))
+        group.add_option('--config', metavar='path', default=config_file,
+                         help="user config file location (default: %default)")
     options, args = parser.parse_args(arglist)
-    if options.testsuite:
+    options.reporter = None
+
+    if options.ensure_value('testsuite', False):
         args.append(options.testsuite)
-    if len(args) == 0:
-        parser.error('input not specified')
-    options.prog = os.path.basename(sys.argv[0])
-    options.exclude = options.exclude.split(',')
-    for index in range(len(options.exclude)):
-        options.exclude[index] = options.exclude[index].rstrip('/')
+    elif not options.ensure_value('doctest', False):
+        if parse_argv and not args:
+            if options.diff or any(os.path.exists(name)
+                                   for name in PROJECT_CONFIG):
+                args = ['.']
+            else:
+                parser.error('input not specified')
+        options = read_config(options, args, arglist, parser)
+        options.reporter = parse_argv and options.quiet == 1 and FileReport
+
     if options.filename:
         options.filename = options.filename.split(',')
+    options.exclude = options.exclude.split(',')
+    if options.select:
+        options.select = options.select.split(',')
     if options.ignore:
         options.ignore = options.ignore.split(',')
-    else:
-        options.ignore = []
-    options.counters = {}
-    options.messages = {}
+
+    if options.diff:
+        options.reporter = DiffReport
+        stdin = stdin_get_value()
+        options.selected_lines = parse_udiff(stdin, options.filename, args[0])
+        args = sorted(options.selected_lines)
 
     return options, args
 
 
 def _main():
-    """
-    Parse options and run checks on Python source.
-    """
-    options, args = process_options()
-    if options.doctest:
-        import doctest
-        return doctest.testmod()
-    start_time = time.time()
-    errors = 0
-    for path in args:
-        # skip emacs backups
-        if path.startswith(".#"):
-            continue
-        if os.path.isdir(path):
-            errors += input_dir(path)
-        else:
-            errors += input_file(path)
-    elapsed = time.time() - start_time
+    """Parse options and run checks on Python source."""
+    pep8style = StyleGuide(parse_argv=True, config_file=True)
+    options = pep8style.options
+    if options.doctest or options.testsuite:
+        from testsuite.support import run_tests
+        report = run_tests(pep8style)
+    else:
+        report = pep8style.check_files()
     if options.statistics:
-        print_statistics()
+        report.print_statistics()
     if options.benchmark:
-        print_benchmark(elapsed)
-    return errors > 0
+        report.print_benchmark()
+    if options.testsuite and not options.quiet:
+        report.print_results()
+    if report.total_errors:
+        if options.count:
+            sys.stderr.write(str(report.total_errors) + '\n')
+        sys.exit(1)
 
 if __name__ == '__main__':
-    sys.exit(_main())
+    _main()
diff --git a/tests/scanner/annotationparser/test_parser.py b/tests/scanner/annotationparser/test_parser.py
index a1c5866c..ce3ccd22 100644
--- a/tests/scanner/annotationparser/test_parser.py
+++ b/tests/scanner/annotationparser/test_parser.py
@@ -108,7 +108,7 @@ def parsed2tree(docblock):
                     parsed += '      <annotations>\n'
                     for key, value in tag.options.values:
                         parsed += '        <annotation>\n'
-                        parsed += '          <name>%s</name>\n' %(key, )
+                        parsed += '          <name>%s</name>\n' % (key, )
                         if value is not None:
                             options = value.all()
                             parsed += '          <options>\n'
@@ -130,6 +130,7 @@ def parsed2tree(docblock):
 
     return parsed
 
+
 def expected2tree(docblock):
     # Note: this sucks, but we can't rely on etree.tostring() to generate useable output :(
 
diff --git a/tests/warn/warningtester.py b/tests/warn/warningtester.py
index 27e0a5e6..fc3c3d2d 100644
--- a/tests/warn/warningtester.py
+++ b/tests/warn/warningtester.py
@@ -61,23 +61,24 @@ def _diff(a, b):
                 for line in a[i1:i2]:
                     for l in line.split('\n'):
                         if l != '':
-                            retval +=  ' ' + l + '\n'
+                            retval += ' ' + l + '\n'
                 continue
 
             if tag in ('replace', 'delete'):
                 for line in a[i1:i2]:
                     for l in line.split('\n'):
                         if l != '':
-                            retval +=  '-' + l + '\n'
+                            retval += '-' + l + '\n'
 
             if tag in ('replace', 'insert'):
                 for line in b[j1:j2]:
                     for l in line.split('\n'):
                         if l != '':
-                            retval +=  '+' + l + '\n'
+                            retval += '+' + l + '\n'
 
     return retval
 
+
 def _extract_expected(filename):
     fd = open(filename, 'rU')
     data = fd.read()
@@ -92,6 +93,7 @@ def _extract_expected(filename):
 
     return retval
 
+
 def check(args):
     filename = args[0]
 
@@ -133,10 +135,10 @@ def check(args):
 
     if len(expected_warnings) != len(emitted_warnings):
         raise SystemExit('ERROR in %r: %d warnings were emitted, '
-                         'expected %d:\n%s' %(os.path.basename(filename),
-                                              len(emitted_warnings),
-                                              len(expected_warnings),
-                                              _diff(expected_warnings, emitted_warnings)))
+                         'expected %d:\n%s' % (os.path.basename(filename),
+                                               len(emitted_warnings),
+                                               len(expected_warnings),
+                                               _diff(expected_warnings, emitted_warnings)))
 
     for emitted_warning, expected_warning in zip(emitted_warnings, expected_warnings):
         if expected_warning != emitted_warning: