Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Erroneous signature inference misses subclassing of inner class #432

Closed
pylint-bot opened this issue Jan 7, 2015 · 1 comment
Closed
Labels

Comments

@pylint-bot
Copy link

Originally reported by: Yann Dirson (BitBucket: ydirson, GitHub: @ydirson?)


This is somewhat linked to #431: here the subclass adding an optional arg is used as an inner class, and pylint seems to take the superclass' methods signatures as constraining the subclass:

#!python

class A(object):
    def method(self, a):
        pass

class B(A):
    def method(self, a, b=None):
        pass

class X(object):
    M = A
    def __init__(self):
        self.m = self.M()
class Y(X):
    M = B
    def blurb(self):
        self.m.method(1, 2)
        self.m.method(1, b=2)

gives:

W:  6, 4: Arguments number differs from overridden method (arguments-differ)
E: 16, 8: Too many positional arguments for method call (too-many-function-args)
E: 17, 8: Unexpected keyword argument 'b' in method call (unexpected-keyword-arg)

Note that I have written a module to help dealing with such a use of inner classes (since standard Python behaviour was more in the way than anything else when one goes further than the above sample), so I'm using this pattern quite much - see https://alioth.debian.org/plugins/scmgit/cgi-bin/gitweb.cgi?p=omaha/omaha.git;a=blob;f=Overlord/innerclass.py;hb=HEAD


@pylint-bot
Copy link
Author

Original comment by Claudiu Popa (BitBucket: PCManticore, GitHub: @PCManticore):


Indeed, it seems that it doesn't take into consideration the fact that M was redefined.

PCManticore added a commit to pylint-dev/astroid that referenced this issue Apr 26, 2016
The inference can handle the case where the attribute is accessed through a subclass
of a base class and the attribute is defined at the base class's level,
by taking in consideration a redefinition in the subclass.

This should fix pylint-dev/pylint#432
brycepg pushed a commit to brycepg/astroid that referenced this issue Feb 27, 2018
The inference can handle the case where the attribute is accessed through a subclass
of a base class and the attribute is defined at the base class's level,
by taking in consideration a redefinition in the subclass.

This should fix pylint-dev/pylint#432
bors bot added a commit to aragilar/DiscSolver that referenced this issue Nov 8, 2019
24: Pin astroid to latest version 2.3.3 r=aragilar a=pyup-bot


This PR pins [astroid](https://pypi.org/project/astroid) to the latest release **2.3.3**.



<details>
  <summary>Changelog</summary>
  
  
   ### 2.3.2
   ```
   ============================
Release Date: TBA

* All type comments have as parent the corresponding `astroid` node

  Until now they had as parent the builtin `ast` node which meant
  we were operating with primitive objects instead of our own.

  Close PyCQA/pylint3174


* Pass an inference context to `metaclass()` when inferring an object type

  This should prevent a bunch of recursion errors happening in pylint.
  Also refactor the inference of `IfExp` nodes to use separate contexts
  for each potential branch.

  Close PyCQA/pylint3152
  Close PyCQA/pylint3159
   ```
   
  
  
   ### 2.3.1
   ```
   ============================
Release Date: 2019-09-30

* A transform for the builtin `dataclasses` module was added.

  This should address various `dataclasses` issues that were surfaced
  even more after the release of pylint 2.4.0.
  In the previous versions of `astroid`, annotated assign nodes were
  allowed to be retrieved via `getattr()` but that no longer happens
  with the latest `astroid` release, as those attribute are not actual
  attributes, but rather virtual ones, thus an operation such as `getattr()`
  does not make sense for them.

* Update attr brain to partly understand annotated attributes

  Close 656
   ```
   
  
  
   ### 2.3.0
   ```
   ============================
Release Date: 2019-09-24

* Add a brain tip for ``subprocess.check_output``

  Close 689

* Remove NodeNG.nearest method because of lack of usage in astroid and pylint.

  Close 691

* Allow importing wheel files. Close 541

* Annotated AST follows PEP8 coding style when converted to string.

* Fix a bug where defining a class using type() could cause a DuplicateBasesError.

  Close 644

* Dropped support for Python 3.4.

* Numpy brain support is improved.

  Numpy&#39;s fundamental type ``numpy.ndarray`` has its own brain : ``brain_numpy_ndarray`` and
  each numpy module that necessitates brain action has now its own numpy brain :

    - ``numpy.core.numeric``
    - ``numpy.core.function_base``
    - ``numpy.core.multiarray``
    - ``numpy.core.numeric``
    - ``numpy.core.numerictypes``
    - ``numpy.core.umath``
    - ``numpy.random.mtrand``

  Close PyCQA/pylint2865
  Close PyCQA/pylint2747
  Close PyCQA/pylint2721
  Close PyCQA/pylint2326
  Close PyCQA/pylint2021

* ``assert`` only functions are properly inferred as returning ``None``

  Close 668

* Add support for Python 3.8&#39;s `NamedExpr` nodes, which is part of assignment expressions.

  Close 674

* Added support for inferring `IfExp` nodes.

* Instances of exceptions are inferred as such when inferring in non-exception context

  This allows special inference support for exception attributes such as `.args`.

  Close PyCQA/pylint2333

* Drop a superfluous and wrong callcontext when inferring the result of a context manager

  Close PyCQA/pylint2859

* ``igetattr`` raises ``InferenceError`` on re-inference of the same object

  This prevents ``StopIteration`` from leaking when we encounter the same
  object in the current context, which could result in various ``RuntimeErrors``
  leaking in other parts of the inference.
  Until we get a global context per inference, the solution is sort of a hack,
  as with the suggested global context improvement, we could theoretically
  reuse the same inference object.

  Close 663

* Variable annotations can no longer be retrieved with `ClassDef.getattr`

  Unless they have an attached value, class variable annotations can no longer
  be retrieved with `ClassDef.getattr.`

* Improved builtin inference for ``tuple``, ``set``, ``frozenset``, ``list`` and ``dict``

  We were properly inferring these callables *only* if they had consts as
  values, but that is not the case most of the time. Instead we try to infer
  the values that their arguments can be and use them instead of assuming
  Const nodes all the time.

  Close PyCQA/pylint2841

* The last except handler wins when inferring variables bound in an except handler.

  Close PyCQA/pylint2777


* ``threading.Lock.locked()`` is properly recognized as a member of ``threading.Lock``

  Close PyCQA/pylint2791


* Fix recursion error involving ``len`` and self referential attributes

  Close PyCQA/pylint2736
  Close PyCQA/pylint2734
  Close PyCQA/pylint2740

* Can access per argument type comments through new ``Arguments.type_comment_args`` attribute.

  Close 665

* Fix being unable to access class attributes on a NamedTuple.

  Close PyCQA/pylint1628

* Fixed being unable to find distutils submodules by name when in a virtualenv.

  Close PyCQA/pylint73
   ```
   
  
  
   ### 2.2.0
   ```
   ============================
Release Date: 2019-02-27


* Fix a bug concerning inference of calls to numpy function that should not return Tuple or List instances.

 Close PyCQA/pylint2436

* Fix a bug where a method, which is a lambda built from a function, is not inferred as ``BoundMethod``

  Close PyCQA/pylint2594

* ``typed_ast`` gets installed for Python 3.7, meaning type comments can now work on 3.7.

* Fix a bug concerning inference of unary operators on numpy types.

  Close PyCQA/pylint2436 (first part)

* Fix a crash with ``typing.NamedTuple`` and empty fields. Close PyCQA/pylint2745

* Add a proper ``strerror`` inference to the ``OSError`` exceptions.

  Close PyCQA/pylint2553

* Support non-const nodes as values of Enum attributes.

  Close 612

* Fix a crash in the ``enum`` brain tip caused by non-assign members in class definitions.

  Close PyCQA/pylint2719

* ``brain_numpy`` returns an undefined type for ``numpy`` methods to avoid ``assignment-from-no-return``

  Close PyCQA/pylint2694

* Fix a bug where a call to a function that has been previously called via
  functools.partial was wrongly inferred

  Close PyCQA/pylint2588

* Fix a recursion error caused by inferring the ``slice`` builtin.

  Close PyCQA/pylint2667

* Remove the restriction that &quot;old style classes&quot; cannot have a MRO.

  This does not make sense any longer given that we run against Python 3
  code.
  Close PyCQA/pylint2701

* Added more builtin exceptions attributes. Close 580

* Add a registry for builtin exception models. Close PyCQA/pylint1432

* Add brain tips for `http.client`. Close PyCQA/pylint2687

* Prevent crashing when processing ``enums`` with mixed single and double quotes.

  Close PyCQA/pylint2676

* ``typing`` types have the `__args__` property. Close PyCQA/pylint2419

* Fix a bug where an Attribute used as a base class was triggering a crash

  Close 626

* Added special support for `enum.IntFlag`

  Close PyCQA/pylint2534

* Extend detection of data classes defined with attr

  Close 628

* Fix typo in description for brain_attrs
   ```
   
  
  
   ### 2.1.0
   ```
   ============================
Release Date: 2018-11-25

   * ``threading.Lock.acquire`` has the ``timeout`` parameter now.

     Close PyCQA/pylint2457

   * Pass parameters by keyword name when inferring sequences.

     Close PyCQA/pylint2526

   * Correct line numbering for f-strings for complex embedded expressions

     When a f-string contained a complex expression, such as an attribute access,
     we weren&#39;t cloning all the subtree of the f-string expression for attaching the correct
     line number. This problem is coming from the builtin AST parser which gives for the f-string
     and for its underlying elements the line number 1, but this is causing all sorts of bugs and
     problems in pylint, which expects correct line numbering.

     Close PyCQA/pylint2449

   * Add support for `argparse.Namespace`

     Close PyCQA/pylint2413

   * `async` functions are now inferred as `AsyncGenerator` when inferring their call result.

   * Filter out ``Uninferable`` when inferring the call result result of a class with an uninferable ``__call__`` method.

     Close PyCQA/pylint2434

   * Make compatible with AST changes in Python 3.8.

   * Subscript inference (e.g. &quot;`a[i]`&quot;) now pays attention to multiple inferred values for value
     (e.g. &quot;`a`&quot;) and slice (e.g. &quot;`i`&quot;)

     Close 614
   ```
   
  
  
   ### 2.0.4
   ```
   ============================
Release Date: 2018-08-10

   * Make sure that assign nodes can find ``yield`` statements in their values

     Close PyCQA/pylint2400
   ```
   
  
  
   ### 2.0.3
   ```
   ============================

Release Date: 2018-08-08

   * The environment markers for PyPy were invalid.
   ```
   
  
  
   ### 2.0.2
   ```
   ============================

Release Date: 2018-08-01

   * Stop repeat inference attempt causing a RuntimeError in Python3.7

     Close PyCQA/pylint2317

   *  infer_call_result can raise InferenceError so make sure to handle that for the call sites
      where it is used

     infer_call_result started recently to raise InferenceError for objects for which it
     could not find any returns. Previously it was silently raising a StopIteration,
     which was especially leaking when calling builtin methods.
     Since it is after all an inference method, it is expected that it
     could raise an InferenceError rather than returning nothing.

     Close PyCQA/pylint2350
   ```
   
  
  
   ### 2.0.1
   ```
   ============================

Release Date: 2018-07-19

   * Released to clear an old wheel package on PyPI
   ```
   
  
  
   ### 2.0
   ```
   ==========================

Release Date: 2018-07-15

   * String representation of nodes takes in account precedence and associativity rules of operators.

   * Fix loading files with `modutils.load_from_module` when
     the path that contains it in `sys.path` is a symlink and
     the file is contained in a symlinked folder.

     Close 583

   * Reworking of the numpy brain dealing with numerictypes
     (use of inspect module to determine the class hierarchy of
      numpy.core.numerictypes module)

     Close PyCQA/pylint2140

   * Added inference support for starred nodes in for loops

     Close 146

   * Support unpacking for dicts in assignments

     Close 268

   * Add support for inferring functools.partial

     Close 125

   * Inference support for `dict.fromkeys`

     Close 110

   * `int()` builtin is inferred as returning integers.

     Close 150

   * `str()` builtin is inferred as returning strings.

     Close 148

   * DescriptorBoundMethod has the correct number of arguments defined.

   * Improvement of the numpy numeric types definition.

     Close PyCQA/pylint1971

   * Subclasses of *property* are now interpreted as properties

     Close PyCQA/pylint1601

   * AsStringRegexpPredicate has been removed.

     Use transform predicates instead of it.

   * Switched to using typed_ast for getting access to type comments

     As a side effect of this change, some nodes gained a new `type_annotation` attribute,
     which, if the type comments were correctly parsed, should contain a node object
     with the corresponding objects from the type comment.

   * typing.X[...] and typing.NewType are inferred as classes instead of instances.

   * Module.__path__ is now a list

     It used to be a string containing the path, but it doesn&#39;t reflect the situation
     on Python, where it is actually a list.

   * Fix a bug with namespace package&#39;s __path__ attribute.

     Close 528

   * Added brain tips for random.sample

     Part of PyCQA/pylint811

   * Add brain tip for `issubclass` builtin

     Close 101.

   * Fix submodule imports from six

     Close PyCQA/pylint1640

   * Fix missing __module__ and __qualname__ from class definition locals

     Close PYCQA/pylint1753

   * Fix a crash when __annotations__ access a parent&#39;s __init__ that does not have arguments

     Close 473

   * Fix multiple objects sharing the same InferenceContext.path causing uninferable results

     Close 483

   * Fix improper modification of col_offset, lineno upon inference of builtin functions

     Close PyCQA/pylint1839

   * Subprocess.Popen brain now knows of the args member

     Close PyCQA/pylint1860

   * add move_to_end method to collections.OrderedDict brain

     Close PyCQA/pylint1872

   * Include new hashlib classes added in python 3.6

   * Fix RecursionError for augmented assign

     Close 437, 447, 313, PyCQA/pylint1642, PyCQA/pylint1805, PyCQA/pylint1854, PyCQA/pylint1452

   * Add missing attrs special attribute

     Close PyCQA/pylint1884

   * Inference now understands the &#39;isinstance&#39; builtin

     Close 98

   * Stop duplicate nodes with the same key values
     from appearing in dictionaries from dictionary unpacking.

     Close PyCQA/pylint1843

   * Fix ``contextlib.contextmanager`` inference for nested context managers

     Close 1699

   * Implement inference for len builtin

     Close 112

   * Add qname method to Super object preventing potential errors in upstream
     pylint

     Close 533

   * Stop astroid from getting stuck in an infinite loop if a function shares
   its name with its decorator

     Close 375

   * Fix issue with inherited __call__ improperly inferencing self

     Close PyCQA/pylint2199

   * Fix __call__ precedence for classes with custom metaclasses

     Close PyCQA/pylint2159

   * Limit the maximum amount of interable result in an NodeNG.infer() call to
    100 by default for performance issues with variables with large amounts of
    possible values.

    The max inferable value can be tuned by setting the `max_inferable_values` flag on
    astroid.MANAGER.
   ```
   
  
  
   ### 1.6.0
   ```
   ============================

Release Date: 2017-12-15


   * When verifying duplicates classes in MRO, ignore on-the-fly generated classes

     Close PyCQA/pylint1706

   * Add brain tip for attrs library to prevent unsupported-assignment-operation false positives

	 Close PYCQA/pylint1698

   * file_stream was removed, since it was deprecated for three releases

     Instead one should use the .stream() method.

   * Vast improvements to numpy support

   * Add brain tips for curses

     Close PyCQA/pylint1703

   * Add brain tips for UUID.int

     Close PyCQA/pylint961

   * The result of using object.__new__ as class decorator is correctly inferred as instance

     Close 172

   * Enums created with functional syntax are now iterable

   * Enums created with functional syntax are now subscriptable

   * Don&#39;t crash when getting the string representation of BadUnaryOperationMessage

     In some cases, when the operand does not have a .name attribute,
     getting the string representation of a BadUnaryOperationMessage leads
     to a crash.

     Close PyCQA/pylint1563

   * Don&#39;t raise DuplicateBaseError when classes at different locations are used

     For instance, one can implement a namedtuple base class, which gets reused
     on a class with the same name later on in the file. Until now, we considered
     these two classes as being the same, because they shared the name, but in fact
     they are different, being created at different locations and through different
     means.

     Close PyCQA/pylint1458

    * The func form of namedtuples with keywords is now understood

      Close PyCQA/pylint1530

    * Fix inference for nested calls

    * Dunder class at method level is now inferred as the class of the method

      Close PyCQA/pylint1328

    * Stop most inference tip overwrites from happening by using
		predicates on existing inference_tip transforms.

      Close 472

    * Fix object.__new__(cls) calls in classmethods by using
        a context which has the proper boundnode for the given
        argument

        Close 404

    * Fix Pathlib type inference

        Close PyCQA/pylint224
        Close PyCQA/pylint1660
   ```
   
  
  
   ### 1.5.3
   ```
   ============================

Release Date: 2017-06-03


    * enum34 dependency is forced to be at least version 1.1.3. Fixes spurious
    bug related to enum classes being falsy in boolean context, which caused
    _Inconsistent Hierarchy_ `RuntimeError` in `singledispatch` module.

    See links below for details:
    - http://bugs.python.org/issue26748
    - https://bitbucket.org/ambv/singledispatch/issues/8/inconsistent-hierarchy-with-enum
    - https://bitbucket.org/stoneleaf/enum34/commits/da50803651ab644e6fce66ebc85562f1117c344b

    * Do not raise an exception when uninferable value is unpacked in ``with`` statement.

    * Lock objects from ``threading`` module are now correctly recognised
      as context managers.
   ```
   
  
  
   ### 1.5.2
   ```
   ============================

Release Date: 2017-04-17


   * Basic support for the class form of typing.NamedTuple

   * mro() can be computed for classes with old style classes in the hierarchy
   ```
   
  
  
   ### 1.5.0
   ```
   ============================

Release Date: 2017-04-13


    * Arguments node gained a new attribute, ``kwonlyargs_annotations``

      This new attribute holds the annotations for the keyword-only
      arguments.

    * `namedtuple` inference now understands `rename` keyword argument

    * Classes can now know their definition-time arguments.

      Classes can support keyword arguments, which are passed when
      a class is constructed using ``__new__``.

    * Add support for inferring typing.NamedTuple.

    * ClassDef now supports __getitem__ inference through the metaclass.

    * getitem() method accepts nodes now, instead of Python objects.

    * Add support for explicit namespace packages, created with pkg_resources.

    * Add brain tips for _io.TextIOWrapper&#39;s buffer and raw attributes.

    * Add `returns` into the proper order in FunctionDef._astroid_fields

      The order is important, since it determines the last child,
      which in turn determines the last line number of a scoped node.

    * Add brain tips for functools.lru_cache.

    * New function, astroid.extract_node, exported out from astroid.test_utils.

    * Stop saving assignment locals in ExceptHandlers, when the context is a store.

      This fixes a tripping case, where the RHS of a ExceptHandler can be redefined
      by the LHS, leading to a local save. For instance, ``except KeyError, exceptions.IndexError``
      could result in a local save for IndexError as KeyError, resulting in potential unexpected
      inferences. Since we don&#39;t lose a lot, this syntax gets prohibited.

    * Fix a crash which occurred when the class of a namedtuple could not be inferred.

    * Add support for implicit namespace packages (PEP 420)

      This change involves a couple of modifications. First, we&#39;re relying on a
      spec finder protocol, inspired by importlib&#39;s ModuleSpec, for finding where
      a file or package is, using importlib&#39;s PathFinder as well, which enable
      us to discover namespace packages as well.
      This discovery is the center piece of the namespace package support,
      the other part being the construction of a dummy Module node whenever
      a namespace package root directory is requested during astroid&#39;s import
      references.

    * Introduce a special attributes model

      Through this model, astroid starts knowing special attributes of certain Python objects,
      such as functions, classes, super objects and so on. This was previously possible before,
      but now the lookup and the attributes themselves are separated into a new module,
      objectmodel.py, which describes, in a more comprehensive way, the data model of each
      object.

    * Exceptions have their own object model

      Some of exceptions&#39;s attributes, such as .args and .message,
      can&#39;t be inferred correctly since they are descriptors that get
      transformed into the proper objects at runtime. This can cause issues
      with the static analysis, since they are inferred as different than
      what&#39;s expected. Now when we&#39;re creating instances of exceptions,
      we&#39;re inferring a special object that knows how to transform those
      runtime attributes into the proper objects via a custom object model.
      Closes issue 81

    * dict.values, dict.keys and dict.items are properly
      inferred to their corresponding type, which also
      includes the proper containers for Python 3.

    * Fix a crash which occurred when a method had a same name as a builtin object,
      decorated at the same time by that builtin object ( a property for instance)

    * The inference can handle the case where the attribute is accessed through a subclass
      of a base class and the attribute is defined at the base class&#39;s level,
      by taking in consideration a redefinition in the subclass.

      This should fix https://github.com/PyCQA/pylint/issues/432

    * Calling lambda methods (defined at class level) can be understood.

    * Don&#39;t take in consideration invalid assignments, especially when __slots__
      declaration forbids them.

      Close issue 332

    * Functional form of enums support accessing values through __call__.

    * Brain tips for the ssl library.

    * decoratornames() does not leak InferenceError anymore.

    * wildcard_imported_names() got replaced by _public_names()

      Our understanding of wildcard imports through __all__ was
      half baked to say at least, since we couldn&#39;t account for
      modifications of the list, which results in tons of false positives.
      Instead, we replaced it with _public_names(), a method which returns
      all the names that are publicly available in a module, that is that
      don&#39;t start with an underscore, even though this means that there
      is a possibility for other names to be leaked out even though
      they are not present in the __all__ variable.

      The method is private in 1.4.X.

    * unpack_infer raises InferenceError if it can&#39;t operate
      with the given sequences of nodes.

    * Support accessing properties with super().

    * Enforce strong updates per frames.

      When looking up a name in a scope, Scope.lookup will return
      only the values which will be reachable after execution, as seen
      in the following code:

           a = 1
           a = 2

      In this case it doesn&#39;t make sense to return two values, but
      only the last one.

    * Add support for inference on threading.Lock

      As a matter of fact, astroid can infer on threading.RLock,
      threading.Semaphore, but can&#39;t do it on threading.Lock (because it comes
      from an extension module).

    * pkg_resources brain tips are a bit more specific,
      by specifying proper returns.

    * The slots() method conflates all the slots from the ancestors
      into a list of current and parent slots.

      We&#39;re doing this because this is the right semantics of slots,
      they get inherited, as long as each parent defines a __slots__
      entry.

    * Some nodes got a new attribute, &#39;ctx&#39;, which tells in which context
      the said node was used.

      The possible values for the contexts are `Load` (&#39;a&#39;), `Del`
      (&#39;del a&#39;), `Store` (&#39;a = 4&#39;) and the nodes that got the new
      attribute are Starred, Subscript, List and Tuple. Closes issue 267.

    * relative_to_absolute_name or methods calling it will now raise
      TooManyLevelsError when a relative import was trying to
      access something beyond the top-level package.

    * AstroidBuildingException is now AstroidBuildingError. The first
      name will exist until astroid 2.0.

    * Add two new exceptions, AstroidImportError and AstroidSyntaxError.
      They are subclasses of AstroidBuildingException and are raised when
      a module can&#39;t be imported from various reasons.
      Also do_import_module lets the errors to bubble up without converting
      them to InferenceError. This particular conversion happens only
      during the inference.

    * Revert to using printf-style formatting in as_string, in order
      to avoid a potential problem with encodings when using .format.
      Closes issue 273. Patch by notsqrt.

    * assigned_stmts methods have the same signature from now on.

      They used to have different signatures and each one made
      assumptions about what could be passed to other implementations,
      leading to various possible crashes when one or more arguments
      weren&#39;t given. Closes issue 277.

    * Fix metaclass detection, when multiple keyword arguments
      are used in class definition.

    * Add support for annotated variable assignments (PEP 526)

    * Starred expressions are now inferred correctly for tuple,
      list, set, and dictionary literals.

    * Support for asynchronous comprehensions introduced in Python 3.6.

      Fixes 399. See PEP530 for details.
   ```
   
  
  
   ### 1.4.1
   ```
   ============================

Release Date: 2015-11-29


    * Add support for handling Uninferable nodes when calling as_string

      Some object, for instance List or Tuple can have, after inference,
      Uninferable as their elements, happening when their components
      weren&#39;t couldn&#39;t be inferred properly. This means that as_string
      needs to cope with expecting Uninferable nodes part of the other
      nodes coming for a string transformation. The patch adds a visit
      method in AsString and ``accept`` on Yes / Uninferable nodes.
      Closes issue 270.
   ```
   
  
  
   ### 1.4.0
   ```
   ============================

Release Date: 2015-11-29


    * Class.getattr(&#39;__mro__&#39;) returns the actual MRO. Closes issue 128.

    * The logilab-common dependency is not needed anymore as the needed code
      was integrated into astroid.

    * Generated enum member stubs now support IntEnum and multiple
      base classes.

    * astroid.builder.AstroidBuilder.string_build and
      astroid.builder.AstroidBuilder.file_build are now raising
      AstroidBuildingException when the parsing of the string raises
      a SyntaxError.

    * Add brain tips for multiprocessing.Manager and
      multiprocessing.managers.SyncManager.

    * Add some fixes which enhances the Jython support.
      The fix mostly includes updates to modutils, which is
      modified in order to properly lookup paths from live objects,
      which ends in $py.class, not pyc as for Python 2,
      Closes issue 83.

    * The Generator objects inferred with `infer_call_result`
      from functions have as parent the function from which they
      are returned.

    * Add brain tips for multiprocessing post Python 3.4+,
      where the module level functions are retrieved with getattr
      from a context object, leading to many no-member errors
      in Pylint.

    * Understand partially the 3-argument form of `type`.
      The only change is that astroid understands members
      passed in as dictionaries as the third argument.

    * .slots() will return an empty list for classes with empty slots.
      Previously it returned None, which is the same value for
      classes without slots at all. This was changed in order
      to better reflect what&#39;s actually happening.

    * Improve the inference of Getattr nodes when dealing with
      abstract properties from the abc module.

      In astroid.bases.Instance._wrap_attr we had a detection
      code for properties, which basically inferred whatever
      a property returned, passing the results up the stack,
      to the igetattr() method. It handled only the builtin property
      but the new patch also handles a couple of other properties,
      such as abc.abstractproperty.

    * UnboundMethod.getattr calls the getattr of its _proxied object
      and doesn&#39;t call super(...) anymore.

      It previously crashed, since the first ancestor in its mro was
      bases.Proxy and bases.Proxy doesn&#39;t implement the .getattr method.
      Closes issue 91.

    * Don&#39;t hard fail when calling .mro() on a class which has
      combined both newstyle and old style classes. The class
      in question is actually newstyle (and the __mro__ can be
      retrieved using Python).

      .mro() fallbacks to using .ancestors() in that case.

    * Class.local_attr and Class.local_attr_ancestors uses internally
      a mro lookup, using .mro() method, if they can.

      That means for newstyle classes, when trying to lookup a member
      using one of these functions, the first one according to the
      mro will be returned. This reflects nicely the reality,
      but it can have as a drawback the fact that it is a behaviour
      change (the previous behaviour was incorrect though). Also,
      having bases which can return multiple values when inferred
      will not work with the new approach, because .mro() only
      retrieves the first value inferred from a base.

    * Expose an implicit_metaclass() method in Class. This will return
      a builtins.type instance for newstyle classes.

    * Add two new exceptions for handling MRO error cases. DuplicateBasesError
      is emitted when duplicate bases are found in a class,
      InconsistentMroError is raised when the method resolution is determined
      to be inconsistent. They share a common class, MroError, which
      is a subclass of ResolveError, meaning that this change is backwards
      compatible.

    * Classes aren&#39;t marked as interfaces anymore, in the `type` attribute.

    * Class.has_dynamic_getattr doesn&#39;t return True for special methods
      which aren&#39;t implemented in pure Python, as it is the case for extension modules.

      Since most likely the methods were coming from a live object, this implies
      that all of them will have __getattr__ and __getattribute__ present and it
      is wrong to consider that those methods were actually implemented.

    * Add basic support for understanding context managers.

      Currently, there&#39;s no way to understand whatever __enter__ returns in a
      context manager and what it is binded using the ``as`` keyword. With these changes,
      we can understand ``bar`` in ``with foo() as bar``, which will be the result of __enter__.

    * Add a new type of node, called *inference objects*. Inference objects are similar with
      AST nodes, but they can be obtained only after inference, so they can&#39;t be found
      inside the original AST tree. Their purpose is to handle at astroid level
      some operations which can&#39;t be handled when using brain transforms.
      For instance, the first object added is FrozenSet, which can be manipulated
      at astroid&#39;s level (inferred, itered etc). Code such as this &#39;frozenset((1,2))&#39;
      will not return an Instance of frozenset, without having access to its
      content, but a new objects.FrozenSet, which can be used just as a nodes.Set.

    * Add a new *inference object* called Super, which also adds support for understanding
      super calls. astroid understands the zero-argument form of super, specific to
      Python 3, where the interpreter fills itself the arguments of the call. Also, we
      are understanding the 2-argument form of super, both for bounded lookups
      (super(X, instance)) as well as for unbounded lookups (super(X, Y)),
      having as well support for validating that the object-or-type is a subtype
      of the first argument. The unbounded form of super (one argument) is not
      understood, since it&#39;s useless in practice and should be removed from
      Python&#39;s specification. Closes issue 89.

    * Add inference support for getattr builtin. Now getattr builtins are
      properly understood. Closes issue 103.

    * Add inference support for hasattr builtin. Closes issue 102.

    * Add &#39;assert_equals&#39; method in nose.tools&#39;s brain plugin.

    * Don&#39;t leak StopIteration when inferring invalid UnaryOps (+[], +None etc.).

    * Improve the inference of UnaryOperands.

      When inferring unary operands, astroid looks up the return value
      of __pos__, __neg__ and __invert__ to determine the inferred value
      of ``~node``, ``+node`` or ``-node``.

    * Improve the inference of six.moves, especially when using `from ... import ...`
      syntax. Also, we added a new fail import hook for six.moves, which fixes the
      import-error false positive from pylint. Closes issue 107.

    * Make the first steps towards detecting type errors for unary and binary
      operations.

      In exceptions, one object was added for holding information about a possible
      UnaryOp TypeError, object called `UnaryOperationError`. Even though the name
      suggests it&#39;s an exception, it&#39;s actually not one. When inferring UnaryOps,
      we use this special object to mark a possible TypeError,
      object which can be interpreted by pylint in order to emit a new warning.
      We are also exposing a new method for UnaryOps, called `type_errors`,
      which returns a list of UnaryOperationsError.

    * A new method was added to the AST nodes, &#39;bool_value&#39;. It is used to deduce
      the value of a node when used in a boolean context, which is useful
      for both inference, as well as for data flow analysis, where we are interested
      in what branches will be followed when the program will be executed.
      `bool_value` returns True, False or YES, if the node&#39;s boolean value can&#39;t
      be deduced. The method is used when inferring the unary operand `not`.
      Thus, `not something` will result in calling `something.bool_value` and
      negating the result, if it is a boolean.

    * Add inference support for boolean operations (`and` and `not`).

    * Add inference support for the builtin `callable`.

    * astroid.inspector was moved to pylint.pyreverse, since
      it is the only known client of this module. No other change
      was made to the exported API.

    * astroid.utils.ASTWalker and astroid.utils.LocalsVisitor
      were moved to pylint.pyreverse.utils.

    * Add inference support for the builtin `bool`.

    * Add `igetattr` method to scoped_nodes.Function.

    * Add support for Python 3.5&#39;s MatMul operation: see PEP 465 for more
      details.

    * NotImplemented is detected properly now as being part of the
      builtins module. Previously trying to infer the Name(NotImplemented)
      returned an YES object.

    * Add astroid.helpers, a module of various useful utilities which don&#39;t
      belong yet into other components. Added *object_type*, a function
      which can be used to obtain the type of almost any astroid object,
      similar to how the builtin *type* works.

    * Understand the one-argument form of the builtin *type*.

      This uses the recently added *astroid.helpers.object_type* in order to
      retrieve the Python type of the first argument of the call.

    * Add helpers.is_supertype and helpers.is_subtype, two functions for
      checking if an object is a super/sub type of another.

    * Improve the inference of binary arithmetic operations (normal
      and augmented).

    * Add support for retrieving TypeErrors for binary arithmetic operations.

      The change is similar to what was added for UnaryOps: a new method
      called *type_errors* for both AugAssign and BinOp, which can be used
      to retrieve type errors occurred during inference. Also, a new
      exception object was added, BinaryOperationError.

    * Lambdas found at class level, which have a `self` argument, are considered
      BoundMethods when accessing them from instances of their class.

    * Add support for multiplication of tuples and lists with instances
      which provides an __index__ returning-int method.

    * Add support for indexing containers with instances which provides
      an __index__ returning-int method.

    * Star unpacking in assignments returns properly a list,
      not the individual components. Closes issue 138.

    * Add annotation support for function.as_string(). Closes issue 37.

    * Add support for indexing bytes on Python 3.

    * Add support for inferring subscript on instances, which will
      use __getitem__. Closes issue 124.

    * Add support for pkg_resources.declare_namespaces.

    * Move pyreverse specific modules and functionality back into pyreverse
      (astroid.manager.Project, astroid.manager.Manager.project_from_files).

    * Understand metaclasses added with six.add_metaclass decorator. Closes issue 129.

    * Add a new convenience API, `astroid.parse`, which can be used to retrieve
      an astroid AST from a source code string, similar to how ast.parse can be
      used to obtain a Python AST from a source string. This is the test_utils.build_module
      promoted to a public API.

    * do_import_module passes the proper relative_only flag if the level is higher
      than 1. This has the side effect that using `from .something import something`
      in a non-package will finally result in an import-error on Pylint&#39;s side.
      Until now relative_only was ignored, leading to the import of `something`,
      if it was globally available.

    * Add get_wrapping_class API to scoped_nodes, which can be used to
      retrieve the class that wraps a node.

    * Class.getattr looks by default in the implicit and the explicit metaclasses,
      which is `type` on Python 3.

      Closes issue 114.

    * There&#39;s a new separate step for transforms.

      Until now, the transforms were applied at the same time the tree was
      being built. This was problematic if the transform functions were
      using inference, since the inference was executed on a partially
      constructed tree, which led to failures when post-building
      information was needed (such as setting the _from_names
      for the From imports).
      Now there&#39;s a separate step for transforms, which are applied
      using transform.TransformVisitor.
      There&#39;s a couple of other related changes:

          * astroid.parse and AstroidBuilder gained a new parameter
            `apply_transforms`, which is a boolean flag, which will
            control if the transforms are applied. We do this because
            there are uses when the vanilla tree is wanted, without
            any implicit modification.

          * the transforms are also applied for builtin modules,
            as a side effect of the fact that transform visiting
            was moved in AstroidBuilder._post_build from
            AstroidBuilder._data_build.

      Closes issue 116.

    * Class._explicit_metaclass is now a public API, in the form of
      Class.declared_metaclass.

      Class.mro remains the de facto method for retrieving the metaclass
      of a class, which will also do an evaluation of what declared_metaclass
      returns.

    * Understand slices of tuples, lists, strings and instances with support
      for slices.

      Closes issue 137.

    * Add proper grammatical names for `inferred` and `ass_type` methods,
      namely `inferred` and `assign_type`.

      The old methods will raise PendingDeprecationWarning, being slated
      for removal in astroid 2.0.

    * Add new AST names in order to be similar to the ones
      from the builtin ast module.

      With this change, Getattr becomes Attributes, Backquote becomes
      Repr, Class is ClassDef, Function is FunctionDef,  Discard is Expr,
      CallFunc is Call, From is ImportFrom, AssName is AssignName
      and AssAttr is AssignAttr. The old names are maintained for backwards
      compatibility and they are interchangeable, in the sense that using
      Discard will use Expr under the hood and the implemented visit_discard
      in checkers will be called with Expr nodes instead. The AST does not
      contain the old nodes, only the interoperability between them hides this
      fact. Recommendations to move to the new nodes are emitted accordingly,
      the old names will be removed in astroid 2.0.

    * Add support for understanding class creation using `type.__new__(mcs, name, bases, attrs)``

      Until now, inferring this kind of calls resulted in Instances, not in classes,
      since astroid didn&#39;t understand that the presence of the metaclass in the call
      leads to a class creating, not to an instance creation.

    * Understand the `slice` builtin. Closes issue 184.

    * Add brain tips for numpy.core, which should fix Pylint&#39;s 453.

    * Add a new node, DictUnpack, which is used to represent the unpacking
      of a dictionary into another dictionary, using PEP 448 specific syntax
      ({1:2, **{2:3})

      This is a different approach than what the builtin ast module does,
      since it just uses None to represent this kind of operation,
      which seems conceptually wrong, due to the fact the AST contains
      non-AST nodes. Closes issue 206.
   ```
   
  
  
   ### 1.3.6
   ```
   ============================

Release Date: 2015-03-14


    * Class.slots raises NotImplementedError for old style classes.
      Closes issue 67.

    * Add a new option to AstroidManager, `optimize_ast`, which
      controls if peephole optimizer should be enabled or not.
      This prevents a regression, where the visit_binop method
      wasn&#39;t called anymore with astroid 1.3.5, due to the differences
      in the resulting AST. Closes issue 82.
   ```
   
  
  
   ### 1.3.5
   ```
   ============================

Release Date: 2015-03-11


    * Add the ability to optimize small ast subtrees,
      with the first use in the optimization of multiple
      BinOp nodes. This removes recursivity in the rebuilder
      when dealing with a lot of small strings joined by the
      addition operator. Closes issue 59.

    * Obtain the methods for the nose brain tip through an
      unittest.TestCase instance. Closes Pylint issue 457.

    * Fix a crash which occurred when a class was the ancestor
      of itself. Closes issue 78.

    * Improve the scope_lookup method for Classes regarding qualified
      objects, with an attribute name exactly as one provided in the
      class itself.

      For example, a class containing an attribute &#39;first&#39;,
      which was also an import and which had, as a base, a qualified name
      or a Gettattr node, in the form &#39;module.first&#39;, then Pylint would
      have inferred the `first` name as the function from the Class,
      not the import. Closes Pylint issue 466.

    * Implement the assigned_stmts operation for Starred nodes,
      which was omitted when support for Python 3 was added in astroid.
      Closes issue 36.
   ```
   
  
  
   ### 1.3.4
   ```
   ============================

Release Date: 2015-01-17


    * Get the first element from the method list when obtaining
      the functions from nose.tools.trivial. Closes Pylint issue 448.
   ```
   
  
  
   ### 1.3.3
   ```
   ============================

Release Date: 2015-01-16


    * Restore file_stream to a property, but deprecate it in favour of
      the newly added method Module.stream. By using a method instead of a
      property, it will be easier to properly close the file right
      after it is used, which will ensure that no file descriptors are
      leaked. Until now, due to the fact that a module was cached,
      it was not possible to close the file_stream anywhere.
      file_stream will start emitting PendingDeprecationWarnings in
      astroid 1.4, DeprecationWarnings in astroid 1.5 and it will
      be finally removed in astroid 1.6.

    * Add inference tips for &#39;tuple&#39;, &#39;list&#39;, &#39;dict&#39; and &#39;set&#39; builtins.

    * Add brain definition for most string and unicode methods

    * Changed the API for Class.slots. It returns None when the class
      doesn&#39;t define any slots. Previously, for both the cases where
      the class didn&#39;t have slots defined and when it had an empty list
      of slots, Class.slots returned an empty list.

    * Add a new method to Class nodes, &#39;mro&#39;, for obtaining the
      the method resolution order of the class.

    * Add brain tips for six.moves. Closes issue 63.

    * Improve the detection for functions decorated with decorators
      which returns static or class methods.

    * .slots() can contain unicode strings on Python 2.

    * Add inference tips for nose.tools.
   ```
   
  
  
   ### 1.3.2
   ```
   ============================

Release Date: 2014-11-22


    * Fixed a crash with invalid subscript index.

    * Implement proper base class semantics for Python 3, where
      every class derives from object.

    * Allow more fine-grained control over C extension loading
      in the manager.
   ```
   
  
  
   ### 1.3.1
   ```
   ============================

Release Date: 2014-11-21


    * Fixed a crash issue with the pytest brain module.
   ```
   
  
  
   ### 1.3.0
   ```
   ============================

Release Date: 2014-11-20


    * Fix a maximum recursion error occurred during the inference,
      where statements with the same name weren&#39;t filtered properly.
      Closes pylint issue 295.

    * Check that EmptyNode has an underlying object in
      EmptyNode.has_underlying_object.

    * Simplify the understanding of enum members.

    * Fix an infinite loop with decorator call chain inference,
      where the decorator returns itself. Closes issue 50.

    * Various speed improvements. Patch by Alex Munroe.

    * Add pytest brain plugin. Patch by Robbie Coomber.

    * Support for Python versions &lt; 2.7 has been dropped, and the
      source has been made compatible with Python 2 and 3. Running
      2to3 on installation for Python 3 is not needed anymore.

    * astroid now depends on six.

    * modutils._module_file opens __init__.py in binary mode.
      Closes issues 51 and 13.

    * Only C extensions from trusted sources (the standard library)
      are loaded into the examining Python process to build an AST
      from the live module.

    * Path names on case-insensitive filesystems are now properly
      handled. This fixes the stdlib detection code on Windows.

    * Metaclass-generating functions like six.with_metaclass
      are now supported via some explicit detection code.

    * astroid.register_module_extender has been added to generalize
      the support for module extenders as used by many brain plugins.

    * brain plugins can now register hooks to handle failed imports,
      as done by the gobject-introspection plugin.

    * The modules have been moved to a separate package directory,
      `setup.py develop` now works correctly.
   ```
   
  
  
   ### 1.2.1
   ```
   ============================

Release Date: 2014-08-24


    * Fix a crash occurred when inferring decorator call chain.
      Closes issue 42.

    * Set the parent of vararg and kwarg nodes when inferring them.
      Closes issue 43.

    * namedtuple inference knows about &#39;_fields&#39; attribute.

    * enum members knows about the methods from the enum class.

    * Name inference will lookup in the parent function
      of the current scope, in case searching in the current scope
      fails.

    * Inference of the functional form of the enums takes into
      consideration the various inputs that enums accepts.

    * The inference engine handles binary operations (add, mul etc.)
      between instances.

    * Fix an infinite loop in the inference, by returning a copy
      of instance attributes, when calling &#39;instance_attr&#39;.
      Closes issue 34 (patch by Emile Anclin).

    * Don&#39;t crash when trying to infer unbound object.__new__ call.
      Closes issue 11.
   ```
   
  
  
   ### 1.2.0
   ```
   ============================

Release Date: 2014-07-25


    * Function nodes can detect decorator call chain and see if they are
      decorated with builtin descriptors (`classmethod` and `staticmethod`).

    * infer_call_result called on a subtype of the builtin type will now
      return a new `Class` rather than an `Instance`.

    * `Class.metaclass()` now handles module-level __metaclass__ declaration
      on python 2, and no longer looks at the __metaclass__ class attribute on
      python 3.

    * Function nodes can detect if they are decorated with subclasses
      of builtin descriptors when determining their type
      (`classmethod` and `staticmethod`).

    * Add `slots` method to `Class` nodes, for retrieving
      the list of valid slots it defines.

    * Expose function annotation to astroid: `Arguments` node
      exposes &#39;varargannotation&#39;, &#39;kwargannotation&#39; and &#39;annotations&#39;
      attributes, while `Function` node has the &#39;returns&#39; attribute.

    * Backported most of the logilab.common.modutils module there, as
      most things there are for pylint/astroid only and we want to be
      able to fix them without requiring a new logilab.common release

    * Fix names grabbed using wildcard import in &quot;absolute import mode&quot;
      (ie with absolute_import activated from the __future__ or with
      python 3). Fix pylint issue 58.

    * Add support in pylint-brain for understanding enum classes.
   ```
   
  
  
   ### 1.1.1
   ```
   ============================

Release Date: 2014-04-30

    * `Class.metaclass()` looks in ancestors when the current class
      does not define explicitly a metaclass.

    * Do not cache modules if a module with the same qname is already
      known, and only return cached modules if both name and filepath
      match. Fixes pylint Bitbucket issue 136.
   ```
   
  
  
   ### 1.1.0
   ```
   ============================

Release Date: 2014-04-18

    * All class nodes are marked as new style classes for Py3k.

    * Add a `metaclass` function to `Class` nodes to
      retrieve their metaclass.

    * Add a new YieldFrom node.

    * Add support for inferring arguments to namedtuple invocations.

    * Make sure that objects returned for namedtuple
      inference have parents.

    * Don&#39;t crash when inferring nodes from `with` clauses
      with multiple context managers. Closes 18.

    * Don&#39;t crash when a class has some __call__ method that is not
      inferable. Closes 17.

    * Unwrap instances found in `.ancestors()`, by using their _proxied
      class.
   ```
   
  
  
   ### 1.0.1
   ```
   ============================

Release Date: 2013-10-18

    * fix py3k/windows installation issue (issue 4)

    * fix bug with namedtuple inference (issue 3)

    * get back gobject introspection from pylint-brain

    * fix some test failures under pypy and py3.3, though there is one remaining
      in each of these platform (2.7 tests are all green)
   ```
   
  
  
   ### 1.0.0
   ```
   =============================

Release Date: 2013-07-29

    * Fix some omissions in py2stdlib&#39;s version of hashlib and
      add a small test for it.

    * Properly recognize methods annotated with abc.abstract{property,method}
      as abstract.

    * Allow transformation functions on any node, providing a
      `register_transform` function on the manager instead of the
     `register_transformer` to make it more flexible wrt node selection

    * Use the new transformation API to provide support for namedtuple
      (actually in pylint-brain, closes 8766)

    * Added the test_utils module for building ASTs and
      extracting deeply nested nodes for easier testing.

    * Add support for py3k&#39;s keyword only arguments (PEP 3102)

    * RENAME THE PROJECT to astroid
   ```
   
  
  
   ### 0.24.3
   ```
   =============================

Release Date: 2013-04-16

    * 124360 [py3.3]: Don&#39;t crash on &#39;yield from&#39; nodes

    * 123062 [pylint-brain]: Use correct names for keywords for urlparse

    * 123056 [pylint-brain]: Add missing methods for hashlib

    * 123068: Fix inference for generator methods to correctly handle yields
      in lambdas.

    * 123068: Make sure .as_string() returns valid code for yields in
      expressions.

    * 47957: Set literals are now correctly treated as inference leaves.

    * 123074: Add support for inference of subscript operations on dict
      literals.
   ```
   
  
  
   ### 0.24.2
   ```
   =============================

Release Date: 2013-02-27

    * pylint-brain: more subprocess.Popen faking (see 46273)

    * 109562 [jython]: java modules have no __doc__, causing crash

    * 120646 [py3]: fix for python3.3 _ast changes which may cause crash

    * 109988 [py3]: test fixes
   ```
   
  
  
   ### 0.24.1
   ```
   =============================

Release Date: 2012-10-05

    * 106191: fix __future__ absolute import w/ From node

    * 50395: fix function fromlineno when some decorator is splited on
      multiple lines (patch by Mark Gius)

    * 92362: fix pyreverse crash on relative import

    * 104041: fix crash &#39;module object has no file_encoding attribute&#39;

    * 4294 (pylint-brain): bad inference on mechanize.Browser.open

    * 46273 (pylint-brain): bad inference subprocess.Popen.communicate
   ```
   
  
  
   ### 0.24.0
   ```
   =============================

Release Date: 2012-07-18

    * include pylint brain extension, describing some stuff not properly understood until then.
      (100013, 53049, 23986, 72355)

    * 99583: fix raw_building.object_build for pypy implementation

    * use `open` rather than `file` in scoped_nodes as 2to3 miss it
   ```
   
  
  
   ### 0.23.1
   ```
   =============================

Release Date: 2011-12-08

    * 62295: avoid &quot;OSError: Too many open files&quot; by moving
      .file_stream as a Module property opening the file only when needed

    * Lambda nodes should have a `name` attribute

    * only call transformers if modname specified
   ```
   
  
  
   ### 0.23.0
   ```
   =============================

Release Date: 2011-10-07

    * 77187: ancestor() only returns the first class when inheriting
      from two classes coming from the same module

    * 76159: putting module&#39;s parent directory on the path causes problems
      linting when file names clash

    * 74746: should return empty module when __main__ is imported (patch by
      google)

    * 74748: getitem protocol return constant value instead of a Const node
      (patch by google)

    * 77188: support lgc.decorators.classproperty

    * 77253: provide a way for user code to register astng &quot;transformers&quot;
      using manager.register_transformer(callable) where callable will be
      called after an astng has been built and given the related module node
      as argument
   ```
   
  
  
   ### 0.22.0
   ```
   =============================

Release Date: 2011-07-18

    * added column offset information on nodes (patch by fawce)

    * 70497: Crash on AttributeError: &#39;NoneType&#39; object has no attribute &#39;_infer_name&#39;

    * 70381: IndentationError in import causes crash

    * 70565: absolute imports treated as relative (patch by Jacek Konieczny)

    * 70494: fix file encoding detection with python2.x

    * py3k: __builtin__ module renamed to builtins, we should consider this to properly
      build ast for builtin objects
   ```
   
  
  
   ### 0.21.1
   ```
   =============================

Release Date: 2011-01-11

    * python3: handle file encoding; fix a lot of tests

    * fix 52006: &quot;True&quot; and &quot;False&quot; can be assigned as variable in Python2x

    * fix 8847: pylint doesn&#39;t understand function attributes at all

    * fix 8774: iterator / generator / next method

    * fix bad building of ast from living object w/ container classes
      (eg dict, set, list, tuple): contained elements should be turned to
      ast as well (not doing it will much probably cause crash later)

    * somewhat fix 57299 and other similar issue: Exception when
      trying to validate file using PyQt&#39;s PyQt4.QtCore module: we can&#39;t
      do much about it but at least catch such exception to avoid crash
   ```
   
  
  
   ### 0.21.0
   ```
   =============================

Release Date: 2010-11-15

    * python3.x: first python3.x release

    * fix 37105: Crash on AttributeError: &#39;NoneType&#39; object has no attribute &#39;_infer_name&#39;

    * python2.4: drop python &lt; 2.5 support
   ```
   
  
  
   ### 0.20.4
   ```
   =============================

Release Date: 2010-10-27

    * fix 37868 37665 33638 37909: import problems with absolute_import_activated

    * fix 8969: false positive when importing from zip-safe eggs

    * fix 46131: minimal class decorator support

    * minimal python2.7 support (dict and set comprehension)

    * important progress on Py3k compatibility
   ```
   
  
  
   ### 0.20.3
   ```
   =============================

Release Date: 2010-09-28

    * restored python 2.3 compatibility

    * fix 45959: AttributeError: &#39;NoneType&#39; object has no attribute &#39;frame&#39;, due
       to handling of __class__ when importing from living object (because of missing
       source code or C-compiled object)
   ```
   
  
  
   ### 0.20.2
   ```
   =============================

Release Date: 2010-09-10

    * fix astng building bug: we&#39;ve to set module.package flag at the node
      creation time otherwise we&#39;ll miss this information when inferring relative
      import during the build process (this should fix for instance some problems
      with numpy)

    * added __subclasses__ to special class attribute

    * fix Class.interfaces so that no InferenceError raised on empty __implements__

    * yield YES on multiplication of tuple/list with non valid operand
   ```
   
  
  
   ### 0.20.1
   ```
   =============================

Release Date: 2010-05-11

    * fix licensing to LGPL

    * add ALL_NODES_CLASSES constant to nodes module

    * nodes redirection cleanup (possible since refactoring)

    * bug fix for python &lt; 2.5: add Delete node on Subscript nodes if we are in a
   del context
   ```
   
  
  
   ### 0.20.0
   ```
   =============================

Release Date: 2010-03-22

    * fix 20464: raises ?TypeError: &#39;_Yes&#39; object is not iterable? on list inference

    * fix 19882: pylint hangs

    * fix 20759: crash on pyreverse UNARY_OP_METHOD KeyError &#39;~&#39;

    * fix 20760: crash on pyreverse : AttributeError: &#39;Subscript&#39;
      object has no attribute &#39;infer_lhs&#39;

    * fix 21980: [Python-modules-team] Bug573229 : Pylint hangs;
      improving the cache yields a speed improvement on big projects

    * major refactoring: rebuild the tree instead of modify / monkey patching

    * fix 19641: &quot;maximum recursion depth exceeded&quot; messages w/ python 2.6
      this was introduced by a refactoring

    * Ned Batchelder patch to properly import eggs with Windows line
      endings.  This fixes a problem with pylint not being able to
      import setuptools.

    * Winfried Plapper patches fixing .op attribute value for AugAssign nodes,
      visit_ifexp in nodes_as_string

    * Edward K. Ream / Tom Fleck patch closes 19641 (maximum recursion depth
      exceeded&quot; messages w/ python 2.6), see https://bugs.launchpad.net/pylint/+bug/456870
   ```
   
  
  
   ### 0.19.3
   ```
   =============================

Release Date: 2009-12-18

    * fix name error making 0.19.2 almost useless
   ```
   
  
  
   ### 0.19.2
   ```
   =============================

Release Date: 2009-12-18

    * fix 18773: inference bug on class member (due to bad handling of instance
      / class nodes &quot;bounded&quot; to method calls)

    * fix 9515: strange message for non-class &quot;Class baz has no egg member&quot; (due to
      bad inference of function call)

    * fix 18953: inference fails with augmented assignment (special case for augmented
      assignement in infer_ass method)

    * fix 13944: false positive for class/instance attributes (Instance.getattr
      should return assign nodes on instance classes as well as instance.

    * include spelling fixes provided by Dotan Barak
   ```
   
  
  
   ### 0.19.1
   ```
   =============================

Release Date: 2009-08-27

    * fix 8771: crash on yield expression

    * fix 10024: line numbering bug with try/except/finally

    * fix 10020: when building from living object, __name__ may be None

    * fix 9891: help(logilab.astng) throws TypeError

    * fix 9588: false positive E1101 for augmented assignment
   ```
   
  
  
   ### 0.19.0
   ```
   =============================

Release Date: 2009-03-25

    * fixed python 2.6 issue (tests ok w/ 2.4, 2.5, 2.6. Anyone using 2.2 / 2.3
      to tell us if it works?)

    * some understanding of the __builtin__.property decorator

    * inference: introduce UnboundMethod / rename InstanceMethod to BoundMethod
   ```
   
  
  
   ### 0.18.0
   ```
   =============================

Release Date: 2009-03-19

    * major api / tree structure changes to make it works with compiler *and*
      python &gt;= 2.5 _ast module

    * cleanup and refactoring on the way
   ```
   
  
  
   ### 0.17.4
   ```
   =============================

Release Date: 2008-11-19

    * fix 6015: filter statements bug triggering W0631 false positive in pylint

    * fix 5571: Function.is_method() should return False on module level
      functions decorated by staticmethod/classmethod (avoid some crash in pylint)

    * fix 5010: understand python 2.5 explicit relative imports
   ```
   
  
  
   ### 0.17.3
   ```
   =============================

Release Date: 2008-09-10

    * fix 5889: astng crash on certain pyreverse projects

    * fix bug w/ loop assignment in .lookup

    * apply Maarten patch fixing a crash on TryFinalaly.block_range and fixing
      &#39;else&#39;/&#39;final&#39; block line detection
   ```
   
  
  
   ### 0.17.2
   ```
   =============================

Release Date: 2008-01-14

    * &quot;with&quot; statement support, patch provided by Brian Hawthorne

    * fixed recursion arguments in nodes_of_class method as notified by
      Dave Borowitz

    * new InstanceMethod node introduced to wrap bound method (e.g. Function
      node), patch provided by Dave Borowitz
   ```
   
  
  
   ### 0.17.1
   ```
   =============================

Release Date: 2007-06-07

    * fix 3651: crash when callable as default arg

    * fix 3670: subscription inference crash in some cases

    * fix 3673: Lambda instance has no attribute &#39;pytype&#39;

    * fix crash with chained &quot;import as&quot;

    * fix crash on numpy

    * fix potential InfiniteRecursion error with builtin objects

    * include patch from Marien Zwart fixing some test / py 2.5

    * be more error resilient when accessing living objects from external
      code in the manager
   ```
   
  
  
   ### 0.17.0
   ```
   =============================

Release Date: 2007-02-22

    * api change to be able to infer using a context (used to infer function call
      result only for now)

    * slightly better inference on astng built from living object by trying to infer
      dummy nodes (able to infer &#39;help&#39; builtin for instance)

    * external attribute definition support

    * basic math operation inference

    * new pytype method on possibly inferred node (e.g. module, classes, const...)

    * fix a living object astng building bug, which was making &quot;open&quot; uninferable

    * fix lookup of name in method bug (3289)

    * fix decorator lookup bug (3261)
   ```
   
  
  
   ### 0.16.3
   ```
   =============================

Release Date: 2006-11-23

    * enhance inference for the subscription notation (motivated by a patch from Amaury)
      and for unary sub/add
   ```
   
  
  
   ### 0.16.2
   ```
   =============================

Release Date: 2006-11-15

    * grrr, fixed python 2.3 incompatibility introduced by generator expression
      scope handling

    * upgrade to avoid warnings with logilab-common 0.21.0 (on which now
      depends so)

    * backported astutils module from logilab-common
   ```
   
  
  
   ### 0.16.1
   ```
   =============================

Release Date: 2006-09-25

    * python 2.5 support, patch provided by Marien Zwart

    * fix [Class|Module].block_range method (this fixes pylint&#39;s inline
      disabling of messages on classes/modules)

    * handle class.__bases__ and class.__mro__ (proper metaclass handling
      still needed though)

    * drop python2.2 support: remove code that was working around python2.2

    * fixed generator expression scope bug

    * patch transformer to extract correct line information
   ```
   
  
  
   ### 0.16.0
   ```
   =============================

Release Date: 2006-04-19

    * fix living object building to consider classes such as property as
      a class instead of a data descriptor

    * fix multiple assignment inference which was discarding some solutions

    * added some line manipulation methods to handle pylint&#39;s block messages
      control feature (Node.last_source_line(), None.block_range(lineno)
   ```
   
  
  
   ### 0.15.1
   ```
   =============================

Release Date: 2006-03-10

    * fix avoiding to load everything from living objects... Thanks Amaury!

    * fix a possible NameError in Instance.infer_call_result
   ```
   
  
  
   ### 0.15.0
   ``…
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant