Edit File - File Manager

Edit File: newtypes.html

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
 <head>
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 
 <title>2. Defining New Types &mdash; Python v2.6.6 documentation</title>
 <link rel="stylesheet" href="../_static/default.css" type="text/css" />
 <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
 <script type="text/javascript">
 var DOCUMENTATION_OPTIONS = {
 URL_ROOT: '../',
 VERSION: '2.6.6',
 COLLAPSE_MODINDEX: false,
 FILE_SUFFIX: '.html',
 HAS_SOURCE: true
 };
 </script>
 <script type="text/javascript" src="../_static/jquery.js"></script>
 <script type="text/javascript" src="../_static/doctools.js"></script>
 <link rel="search" type="application/opensearchdescription+xml"
 title="Search within Python v2.6.6 documentation"
 href="../_static/opensearch.xml"/>
 <link rel="author" title="About these documents" href="../about.html" />
 <link rel="copyright" title="Copyright" href="../copyright.html" />
 <link rel="top" title="Python v2.6.6 documentation" href="../index.html" />
 <link rel="up" title="Extending and Embedding the Python Interpreter" href="index.html" />
 <link rel="next" title="3. Building C and C++ Extensions with distutils" href="building.html" />
 <link rel="prev" title="1. Extending Python with C or C++" href="extending.html" />
 <link rel="shortcut icon" type="image/png" href="../_static/py.png" />

</head>
 <body>
 <div class="related">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="../genindex.html" title="General Index"
 accesskey="I">index</a></li>
 <li class="right" >
 <a href="../modindex.html" title="Global Module Index"
 accesskey="M">modules</a> |</li>
 <li class="right" >
 <a href="building.html" title="3. Building C and C++ Extensions with distutils"
 accesskey="N">next</a> |</li>
 <li class="right" >
 <a href="extending.html" title="1. Extending Python with C or C++"
 accesskey="P">previous</a> |</li>
 <li><img src="../_static/py.png" alt=""
 style="vertical-align: middle; margin-top: -1px"/></li>
 <li><a href="../index.html">Python v2.6.6 documentation</a> &raquo;</li>

<li><a href="index.html" accesskey="U">Extending and Embedding the Python Interpreter</a> &raquo;</li> 
 </ul>
 </div>

<div class="document">
 <div class="documentwrapper">
 <div class="bodywrapper">
 <div class="body">
 
 <div class="section" id="defining-new-types">
<h1>2. Defining New Types<a class="headerlink" href="#defining-new-types" title="Permalink to this headline">¶</a></h1>
As mentioned in the last chapter, Python allows the writer of an extension
module to define new types that can be manipulated from Python code, much like
strings and lists in core Python.
This is not hard; the code for all extension types follows a pattern, but there
are some details that you need to understand before you can get started.
<div class="admonition note">
Note
The way new types are defined changed dramatically (and for the better) in
Python 2.2. This document documents how to define new types for Python 2.2 and
later. If you need to support older versions of Python, you will need to refer
to <a class="reference external" href="http://www.python.org/doc/versions/">older versions of this documentation</a>.
</div>
<div class="section" id="the-basics">
<h2>2.1. The Basics<a class="headerlink" href="#the-basics" title="Permalink to this headline">¶</a></h2>
The Python runtime sees all Python objects as variables of type
<a title="PyObject" class="reference external" href="../c-api/structures.html#PyObject"><tt class="xref docutils literal">PyObject*</tt></a>. A <a title="PyObject" class="reference external" href="../c-api/structures.html#PyObject"><tt class="xref docutils literal">PyObject</tt></a> is not a very magnificent object - it
just contains the refcount and a pointer to the object&#8217;s &#8220;type object&#8221;. This is
where the action is; the type object determines which (C) functions get called
when, for instance, an attribute gets looked up on an object or it is multiplied
by another object. These C functions are called &#8220;type methods&#8221; to distinguish
them from things like <tt class="docutils literal">[].append</tt> (which we call &#8220;object methods&#8221;).
So, if you want to define a new object type, you need to create a new type
object.
This sort of thing can only be explained by example, so here&#8217;s a minimal, but
complete, module that defines a new type:
<div class="highlight-c"><div class="highlight"><pre>#include &lt;Python.h&gt;

typedef struct {
 PyObject_HEAD
 /* Type-specific fields go here. */
} noddy_NoddyObject;

static PyTypeObject noddy_NoddyType = {
 PyObject_HEAD_INIT(NULL)
 0, /*ob_size*/
 &quot;noddy.Noddy&quot;, /*tp_name*/
 sizeof(noddy_NoddyObject), /*tp_basicsize*/
 0, /*tp_itemsize*/
 0, /*tp_dealloc*/
 0, /*tp_print*/
 0, /*tp_getattr*/
 0, /*tp_setattr*/
 0, /*tp_compare*/
 0, /*tp_repr*/
 0, /*tp_as_number*/
 0, /*tp_as_sequence*/
 0, /*tp_as_mapping*/
 0, /*tp_hash */
 0, /*tp_call*/
 0, /*tp_str*/
 0, /*tp_getattro*/
 0, /*tp_setattro*/
 0, /*tp_as_buffer*/
 Py_TPFLAGS_DEFAULT, /*tp_flags*/
 &quot;Noddy objects&quot;, /* tp_doc */
};

static PyMethodDef noddy_methods[] = {
 {NULL} /* Sentinel */
};

#ifndef PyMODINIT_FUNC	/* declarations for DLL import/export */
#define PyMODINIT_FUNC void
#endif
PyMODINIT_FUNC
initnoddy(void) 
{
 PyObject* m;

noddy_NoddyType.tp_new = PyType_GenericNew;
 if (PyType_Ready(&amp;noddy_NoddyType) &lt; 0)
 return;

m = Py_InitModule3(&quot;noddy&quot;, noddy_methods,
 &quot;Example module that creates an extension type.&quot;);

Py_INCREF(&amp;noddy_NoddyType);
 PyModule_AddObject(m, &quot;Noddy&quot;, (PyObject *)&amp;noddy_NoddyType);
}
</pre></div>
</div>
Now that&#8217;s quite a bit to take in at once, but hopefully bits will seem familiar
from the last chapter.
The first bit that will be new is:
<div class="highlight-c"><div class="highlight"><pre>typedef struct {
 PyObject_HEAD
} noddy_NoddyObject;
</pre></div>
</div>
This is what a Noddy object will contain&#8212;in this case, nothing more than every
Python object contains, namely a refcount and a pointer to a type object. These
are the fields the <tt class="docutils literal">PyObject_HEAD</tt> macro brings in. The reason for the macro
is to standardize the layout and to enable special debugging fields in debug
builds. Note that there is no semicolon after the <tt class="docutils literal">PyObject_HEAD</tt> macro; one
is included in the macro definition. Be wary of adding one by accident; it&#8217;s
easy to do from habit, and your compiler might not complain, but someone else&#8217;s
probably will! (On Windows, MSVC is known to call this an error and refuse to
compile the code.)
For contrast, let&#8217;s take a look at the corresponding definition for standard
Python integers:
<div class="highlight-c"><div class="highlight"><pre>typedef struct {
 PyObject_HEAD
 long ob_ival;
} PyIntObject;
</pre></div>
</div>
Moving on, we come to the crunch &#8212; the type object.
<div class="highlight-c"><div class="highlight"><pre>static PyTypeObject noddy_NoddyType = {
 PyObject_HEAD_INIT(NULL)
 0, /*ob_size*/
 &quot;noddy.Noddy&quot;, /*tp_name*/
 sizeof(noddy_NoddyObject), /*tp_basicsize*/
 0, /*tp_itemsize*/
 0, /*tp_dealloc*/
 0, /*tp_print*/
 0, /*tp_getattr*/
 0, /*tp_setattr*/
 0, /*tp_compare*/
 0, /*tp_repr*/
 0, /*tp_as_number*/
 0, /*tp_as_sequence*/
 0, /*tp_as_mapping*/
 0, /*tp_hash */
 0, /*tp_call*/
 0, /*tp_str*/
 0, /*tp_getattro*/
 0, /*tp_setattro*/
 0, /*tp_as_buffer*/
 Py_TPFLAGS_DEFAULT, /*tp_flags*/
 &quot;Noddy objects&quot;, /* tp_doc */
};
</pre></div>
</div>
Now if you go and look up the definition of <a title="PyTypeObject" class="reference external" href="../c-api/type.html#PyTypeObject"><tt class="xref docutils literal">PyTypeObject</tt></a> in
<tt class="docutils literal">object.h</tt> you&#8217;ll see that it has many more fields that the definition
above. The remaining fields will be filled with zeros by the C compiler, and
it&#8217;s common practice to not specify them explicitly unless you need them.
This is so important that we&#8217;re going to pick the top of it apart still
further:
<div class="highlight-c"><div class="highlight"><pre>PyObject_HEAD_INIT(NULL)
</pre></div>
</div>
This line is a bit of a wart; what we&#8217;d like to write is:
<div class="highlight-c"><div class="highlight"><pre>PyObject_HEAD_INIT(&amp;PyType_Type)
</pre></div>
</div>
as the type of a type object is &#8220;type&#8221;, but this isn&#8217;t strictly conforming C and
some compilers complain. Fortunately, this member will be filled in for us by
<a title="PyType_Ready" class="reference external" href="../c-api/type.html#PyType_Ready"><tt class="xref docutils literal">PyType_Ready()</tt></a>.
<div class="highlight-c"><div class="highlight"><pre>0, /* ob_size */
</pre></div>
</div>
The <a title="ob_size" class="reference external" href="../c-api/typeobj.html#ob_size"><tt class="xref docutils literal">ob_size</tt></a> field of the header is not used; its presence in the type
structure is a historical artifact that is maintained for binary compatibility
with extension modules compiled for older versions of Python. Always set this
field to zero.
<div class="highlight-c"><div class="highlight"><pre>&quot;noddy.Noddy&quot;, /* tp_name */
</pre></div>
</div>
The name of our type. This will appear in the default textual representation of
our objects and in some error messages, for example:
<div class="highlight-c"><div class="highlight"><pre>&gt;&gt;&gt; &quot;&quot; + noddy.new_noddy()
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in ?
TypeError: cannot add type &quot;noddy.Noddy&quot; to string
</pre></div>
</div>
Note that the name is a dotted name that includes both the module name and the
name of the type within the module. The module in this case is <tt class="xref docutils literal">noddy</tt> and
the type is <tt class="xref docutils literal">Noddy</tt>, so we set the type name to <tt class="xref docutils literal">noddy.Noddy</tt>.
<div class="highlight-c"><div class="highlight"><pre>sizeof(noddy_NoddyObject), /* tp_basicsize */
</pre></div>
</div>
This is so that Python knows how much memory to allocate when you call
<a title="PyObject_New" class="reference external" href="../c-api/allocation.html#PyObject_New"><tt class="xref docutils literal">PyObject_New()</tt></a>.
<div class="admonition note">
Note
If you want your type to be subclassable from Python, and your type has the same
<a title="tp_basicsize" class="reference external" href="../c-api/typeobj.html#tp_basicsize"><tt class="xref docutils literal">tp_basicsize</tt></a> as its base type, you may have problems with multiple
inheritance. A Python subclass of your type will have to list your type first
in its <tt class="xref docutils literal">__bases__</tt>, or else it will not be able to call your type&#8217;s
<a title="object.__new__" class="reference external" href="../reference/datamodel.html#object.__new__"><tt class="xref docutils literal">__new__()</tt></a> method without getting an error. You can avoid this problem by
ensuring that your type has a larger value for <a title="tp_basicsize" class="reference external" href="../c-api/typeobj.html#tp_basicsize"><tt class="xref docutils literal">tp_basicsize</tt></a> than its
base type does. Most of the time, this will be true anyway, because either your
base type will be <a title="object" class="reference external" href="../library/functions.html#object"><tt class="xref docutils literal">object</tt></a>, or else you will be adding data members to
your base type, and therefore increasing its size.
</div>
<div class="highlight-c"><div class="highlight"><pre>0, /* tp_itemsize */
</pre></div>
</div>
This has to do with variable length objects like lists and strings. Ignore this
for now.
Skipping a number of type methods that we don&#8217;t provide, we set the class flags
to <a title="Py_TPFLAGS_DEFAULT" class="reference external" href="../c-api/typeobj.html#Py_TPFLAGS_DEFAULT"><tt class="xref docutils literal">Py_TPFLAGS_DEFAULT</tt></a>.
<div class="highlight-c"><div class="highlight"><pre>Py_TPFLAGS_DEFAULT, /*tp_flags*/
</pre></div>
</div>
All types should include this constant in their flags. It enables all of the
members defined by the current version of Python.
We provide a doc string for the type in <a title="tp_doc" class="reference external" href="../c-api/typeobj.html#tp_doc"><tt class="xref docutils literal">tp_doc</tt></a>.
<div class="highlight-c"><div class="highlight"><pre>&quot;Noddy objects&quot;, /* tp_doc */
</pre></div>
</div>
Now we get into the type methods, the things that make your objects different
from the others. We aren&#8217;t going to implement any of these in this version of
the module. We&#8217;ll expand this example later to have more interesting behavior.
For now, all we want to be able to do is to create new <tt class="xref docutils literal">Noddy</tt> objects.
To enable object creation, we have to provide a <a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> implementation.
In this case, we can just use the default implementation provided by the API
function <a title="PyType_GenericNew" class="reference external" href="../c-api/type.html#PyType_GenericNew"><tt class="xref docutils literal">PyType_GenericNew()</tt></a>. We&#8217;d like to just assign this to the
<a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> slot, but we can&#8217;t, for portability sake, On some platforms or
compilers, we can&#8217;t statically initialize a structure member with a function
defined in another C module, so, instead, we&#8217;ll assign the <a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> slot
in the module initialization function just before calling
<a title="PyType_Ready" class="reference external" href="../c-api/type.html#PyType_Ready"><tt class="xref docutils literal">PyType_Ready()</tt></a>:
<div class="highlight-c"><div class="highlight"><pre>noddy_NoddyType.tp_new = PyType_GenericNew;
if (PyType_Ready(&amp;noddy_NoddyType) &lt; 0)
 return;
</pre></div>
</div>
All the other type methods are NULL, so we&#8217;ll go over them later &#8212; that&#8217;s
for a later section!
Everything else in the file should be familiar, except for some code in
<tt class="xref docutils literal">initnoddy()</tt>:
<div class="highlight-c"><div class="highlight"><pre>if (PyType_Ready(&amp;noddy_NoddyType) &lt; 0)
 return;
</pre></div>
</div>
This initializes the <tt class="xref docutils literal">Noddy</tt> type, filing in a number of members,
including <a title="ob_type" class="reference external" href="../c-api/typeobj.html#ob_type"><tt class="xref docutils literal">ob_type</tt></a> that we initially set to NULL.
<div class="highlight-c"><div class="highlight"><pre>PyModule_AddObject(m, &quot;Noddy&quot;, (PyObject *)&amp;noddy_NoddyType);
</pre></div>
</div>
This adds the type to the module dictionary. This allows us to create
<tt class="xref docutils literal">Noddy</tt> instances by calling the <tt class="xref docutils literal">Noddy</tt> class:
<div class="highlight-c"><div class="highlight"><pre>&gt;&gt;&gt; import noddy
&gt;&gt;&gt; mynoddy = noddy.Noddy()
</pre></div>
</div>
That&#8217;s it! All that remains is to build it; put the above code in a file called
<tt class="docutils literal">noddy.c</tt> and
<div class="highlight-c"><div class="highlight"><pre>from distutils.core import setup, Extension
setup(name=&quot;noddy&quot;, version=&quot;1.0&quot;,
 ext_modules=[Extension(&quot;noddy&quot;, [&quot;noddy.c&quot;])])
</pre></div>
</div>
in a file called <tt class="docutils literal">setup.py</tt>; then typing
<div class="highlight-c"><pre>$ python setup.py build</pre>
</div>
at a shell should produce a file <tt class="docutils literal">noddy.so</tt> in a subdirectory; move to
that directory and fire up Python &#8212; you should be able to <tt class="docutils literal">import noddy</tt> and
play around with Noddy objects.
That wasn&#8217;t so hard, was it?
Of course, the current Noddy type is pretty uninteresting. It has no data and
doesn&#8217;t do anything. It can&#8217;t even be subclassed.
<div class="section" id="adding-data-and-methods-to-the-basic-example">
<h3>2.1.1. Adding data and methods to the Basic example<a class="headerlink" href="#adding-data-and-methods-to-the-basic-example" title="Permalink to this headline">¶</a></h3>
Let&#8217;s expend the basic example to add some data and methods. Let&#8217;s also make
the type usable as a base class. We&#8217;ll create a new module, <tt class="xref docutils literal">noddy2</tt> that
adds these capabilities:
<div class="highlight-c"><div class="highlight"><pre>#include &lt;Python.h&gt;
#include &quot;structmember.h&quot;

typedef struct {
 PyObject_HEAD
 PyObject *first; /* first name */
 PyObject *last; /* last name */
 int number;
} Noddy;

static PyObject *
Noddy_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
{
 Noddy *self;

self = (Noddy *)type-&gt;tp_alloc(type, 0);
 if (self != NULL) {
 self-&gt;first = PyString_FromString(&quot;&quot;);
 if (self-&gt;first == NULL)
 {
 Py_DECREF(self);
 return NULL;
 }
 
 self-&gt;last = PyString_FromString(&quot;&quot;);
 if (self-&gt;last == NULL)
 {
 Py_DECREF(self);
 return NULL;
 }

self-&gt;number = 0;
 }

static int
Noddy_init(Noddy *self, PyObject *args, PyObject *kwds)
{
 PyObject *first=NULL, *last=NULL, *tmp;

static char *kwlist[] = {&quot;first&quot;, &quot;last&quot;, &quot;number&quot;, NULL};

if (! PyArg_ParseTupleAndKeywords(args, kwds, &quot;|OOi&quot;, kwlist, 
 &amp;first, &amp;last, 
 &amp;self-&gt;number))
 return -1;

if (first) {
 tmp = self-&gt;first;
 Py_INCREF(first);
 self-&gt;first = first;
 Py_XDECREF(tmp);
 }

if (last) {
 tmp = self-&gt;last;
 Py_INCREF(last);
 self-&gt;last = last;
 Py_XDECREF(tmp);
 }

return 0;
}

static PyMemberDef Noddy_members[] = {
 {&quot;first&quot;, T_OBJECT_EX, offsetof(Noddy, first), 0,
 &quot;first name&quot;},
 {&quot;last&quot;, T_OBJECT_EX, offsetof(Noddy, last), 0,
 &quot;last name&quot;},
 {&quot;number&quot;, T_INT, offsetof(Noddy, number), 0,
 &quot;noddy number&quot;},
 {NULL} /* Sentinel */
};

static PyObject *
Noddy_name(Noddy* self)
{
 static PyObject *format = NULL;
 PyObject *args, *result;

if (format == NULL) {
 format = PyString_FromString(&quot;%s %s&quot;);
 if (format == NULL)
 return NULL;
 }

if (self-&gt;first == NULL) {
 PyErr_SetString(PyExc_AttributeError, &quot;first&quot;);
 return NULL;
 }

if (self-&gt;last == NULL) {
 PyErr_SetString(PyExc_AttributeError, &quot;last&quot;);
 return NULL;
 }

args = Py_BuildValue(&quot;OO&quot;, self-&gt;first, self-&gt;last);
 if (args == NULL)
 return NULL;

result = PyString_Format(format, args);
 Py_DECREF(args);
 
 return result;
}

static PyMethodDef Noddy_methods[] = {
 {&quot;name&quot;, (PyCFunction)Noddy_name, METH_NOARGS,
 &quot;Return the name, combining the first and last name&quot;
 },
 {NULL} /* Sentinel */
};

static PyTypeObject NoddyType = {
 PyObject_HEAD_INIT(NULL)
 0, /*ob_size*/
 &quot;noddy.Noddy&quot;, /*tp_name*/
 sizeof(Noddy), /*tp_basicsize*/
 0, /*tp_itemsize*/
 (destructor)Noddy_dealloc, /*tp_dealloc*/
 0, /*tp_print*/
 0, /*tp_getattr*/
 0, /*tp_setattr*/
 0, /*tp_compare*/
 0, /*tp_repr*/
 0, /*tp_as_number*/
 0, /*tp_as_sequence*/
 0, /*tp_as_mapping*/
 0, /*tp_hash */
 0, /*tp_call*/
 0, /*tp_str*/
 0, /*tp_getattro*/
 0, /*tp_setattro*/
 0, /*tp_as_buffer*/
 Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/
 &quot;Noddy objects&quot;, /* tp_doc */
 0,		 /* tp_traverse */
 0,		 /* tp_clear */
 0,		 /* tp_richcompare */
 0,		 /* tp_weaklistoffset */
 0,		 /* tp_iter */
 0,		 /* tp_iternext */
 Noddy_methods, /* tp_methods */
 Noddy_members, /* tp_members */
 0, /* tp_getset */
 0, /* tp_base */
 0, /* tp_dict */
 0, /* tp_descr_get */
 0, /* tp_descr_set */
 0, /* tp_dictoffset */
 (initproc)Noddy_init, /* tp_init */
 0, /* tp_alloc */
 Noddy_new, /* tp_new */
};

static PyMethodDef module_methods[] = {
 {NULL} /* Sentinel */
};

#ifndef PyMODINIT_FUNC	/* declarations for DLL import/export */
#define PyMODINIT_FUNC void
#endif
PyMODINIT_FUNC
initnoddy2(void) 
{
 PyObject* m;

if (PyType_Ready(&amp;NoddyType) &lt; 0)
 return;

m = Py_InitModule3(&quot;noddy2&quot;, module_methods,
 &quot;Example module that creates an extension type.&quot;);

if (m == NULL)
 return;

Py_INCREF(&amp;NoddyType);
 PyModule_AddObject(m, &quot;Noddy&quot;, (PyObject *)&amp;NoddyType);
}
</pre></div>
</div>
This version of the module has a number of changes.
We&#8217;ve added an extra include:
<div class="highlight-c"><div class="highlight"><pre>#include &lt;structmember.h&gt;
</pre></div>
</div>
This include provides declarations that we use to handle attributes, as
described a bit later.
The name of the <tt class="xref docutils literal">Noddy</tt> object structure has been shortened to
<tt class="xref docutils literal">Noddy</tt>. The type object name has been shortened to <tt class="xref docutils literal">NoddyType</tt>.
The <tt class="xref docutils literal">Noddy</tt> type now has three data attributes, first, last, and
number. The first and last variables are Python strings containing first
and last names. The number attribute is an integer.
The object structure is updated accordingly:
<div class="highlight-c"><div class="highlight"><pre>typedef struct {
 PyObject_HEAD
 PyObject *first;
 PyObject *last;
 int number;
} Noddy;
</pre></div>
</div>
Because we now have data to manage, we have to be more careful about object
allocation and deallocation. At a minimum, we need a deallocation method:
<div class="highlight-c"><div class="highlight"><pre>static void
Noddy_dealloc(Noddy* self)
{
 Py_XDECREF(self-&gt;first);
 Py_XDECREF(self-&gt;last);
 self-&gt;ob_type-&gt;tp_free((PyObject*)self);
}
</pre></div>
</div>
which is assigned to the <a title="tp_dealloc" class="reference external" href="../c-api/typeobj.html#tp_dealloc"><tt class="xref docutils literal">tp_dealloc</tt></a> member:
<div class="highlight-c"><div class="highlight"><pre>(destructor)Noddy_dealloc, /*tp_dealloc*/
</pre></div>
</div>
This method decrements the reference counts of the two Python attributes. We use
<a title="Py_XDECREF" class="reference external" href="../c-api/refcounting.html#Py_XDECREF"><tt class="xref docutils literal">Py_XDECREF()</tt></a> here because the <tt class="xref docutils literal">first</tt> and <tt class="xref docutils literal">last</tt> members
could be NULL. It then calls the <a title="tp_free" class="reference external" href="../c-api/typeobj.html#tp_free"><tt class="xref docutils literal">tp_free</tt></a> member of the object&#8217;s type
to free the object&#8217;s memory. Note that the object&#8217;s type might not be
<tt class="xref docutils literal">NoddyType</tt>, because the object may be an instance of a subclass.
We want to make sure that the first and last names are initialized to empty
strings, so we provide a new method:
<div class="highlight-c"><div class="highlight"><pre>static PyObject *
Noddy_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
{
 Noddy *self;

self-&gt;last = PyString_FromString(&quot;&quot;);
 if (self-&gt;last == NULL)
 {
 Py_DECREF(self);
 return NULL;
 }

self-&gt;number = 0;
 }

return (PyObject *)self;
}
</pre></div>
</div>
and install it in the <a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> member:
<div class="highlight-c"><div class="highlight"><pre>Noddy_new, /* tp_new */
</pre></div>
</div>
The new member is responsible for creating (as opposed to initializing) objects
of the type. It is exposed in Python as the <a title="object.__new__" class="reference external" href="../reference/datamodel.html#object.__new__"><tt class="xref docutils literal">__new__()</tt></a> method. See the
paper titled &#8220;Unifying types and classes in Python&#8221; for a detailed discussion of
the <a title="object.__new__" class="reference external" href="../reference/datamodel.html#object.__new__"><tt class="xref docutils literal">__new__()</tt></a> method. One reason to implement a new method is to assure
the initial values of instance variables. In this case, we use the new method
to make sure that the initial values of the members <tt class="xref docutils literal">first</tt> and
<tt class="xref docutils literal">last</tt> are not NULL. If we didn&#8217;t care whether the initial values were
NULL, we could have used <a title="PyType_GenericNew" class="reference external" href="../c-api/type.html#PyType_GenericNew"><tt class="xref docutils literal">PyType_GenericNew()</tt></a> as our new method, as we
did before. <a title="PyType_GenericNew" class="reference external" href="../c-api/type.html#PyType_GenericNew"><tt class="xref docutils literal">PyType_GenericNew()</tt></a> initializes all of the instance variable
members to NULL.
The new method is a static method that is passed the type being instantiated and
any arguments passed when the type was called, and that returns the new object
created. New methods always accept positional and keyword arguments, but they
often ignore the arguments, leaving the argument handling to initializer
methods. Note that if the type supports subclassing, the type passed may not be
the type being defined. The new method calls the tp_alloc slot to allocate
memory. We don&#8217;t fill the <a title="tp_alloc" class="reference external" href="../c-api/typeobj.html#tp_alloc"><tt class="xref docutils literal">tp_alloc</tt></a> slot ourselves. Rather
<a title="PyType_Ready" class="reference external" href="../c-api/type.html#PyType_Ready"><tt class="xref docutils literal">PyType_Ready()</tt></a> fills it for us by inheriting it from our base class,
which is <a title="object" class="reference external" href="../library/functions.html#object"><tt class="xref docutils literal">object</tt></a> by default. Most types use the default allocation.
<div class="admonition note">
Note
If you are creating a co-operative <a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> (one that calls a base type&#8217;s
<a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> or <a title="object.__new__" class="reference external" href="../reference/datamodel.html#object.__new__"><tt class="xref docutils literal">__new__()</tt></a>), you must not try to determine what method
to call using method resolution order at runtime. Always statically determine
what type you are going to call, and call its <a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a> directly, or via
<tt class="docutils literal">type-&gt;tp_base-&gt;tp_new</tt>. If you do not do this, Python subclasses of your
type that also inherit from other Python-defined classes may not work correctly.
(Specifically, you may not be able to create instances of such subclasses
without getting a <a title="exceptions.TypeError" class="reference external" href="../library/exceptions.html#exceptions.TypeError"><tt class="xref docutils literal">TypeError</tt></a>.)
</div>
We provide an initialization function:
<div class="highlight-c"><div class="highlight"><pre>static int
Noddy_init(Noddy *self, PyObject *args, PyObject *kwds)
{
 PyObject *first=NULL, *last=NULL, *tmp;

if (! PyArg_ParseTupleAndKeywords(args, kwds, &quot;|OOi&quot;, kwlist,
 &amp;first, &amp;last,
 &amp;self-&gt;number))
 return -1;

return 0;
}
</pre></div>
</div>
by filling the <a title="tp_init" class="reference external" href="../c-api/typeobj.html#tp_init"><tt class="xref docutils literal">tp_init</tt></a> slot.
<div class="highlight-c"><div class="highlight"><pre>(initproc)Noddy_init, /* tp_init */
</pre></div>
</div>
The <a title="tp_init" class="reference external" href="../c-api/typeobj.html#tp_init"><tt class="xref docutils literal">tp_init</tt></a> slot is exposed in Python as the <a title="object.__init__" class="reference external" href="../reference/datamodel.html#object.__init__"><tt class="xref docutils literal">__init__()</tt></a> method. It
is used to initialize an object after it&#8217;s created. Unlike the new method, we
can&#8217;t guarantee that the initializer is called. The initializer isn&#8217;t called
when unpickling objects and it can be overridden. Our initializer accepts
arguments to provide initial values for our instance. Initializers always accept
positional and keyword arguments.
Initializers can be called multiple times. Anyone can call the <a title="object.__init__" class="reference external" href="../reference/datamodel.html#object.__init__"><tt class="xref docutils literal">__init__()</tt></a>
method on our objects. For this reason, we have to be extra careful when
assigning the new values. We might be tempted, for example to assign the
<tt class="xref docutils literal">first</tt> member like this:
<div class="highlight-c"><div class="highlight"><pre>if (first) {
 Py_XDECREF(self-&gt;first);
 Py_INCREF(first);
 self-&gt;first = first;
}
</pre></div>
</div>
But this would be risky. Our type doesn&#8217;t restrict the type of the
<tt class="xref docutils literal">first</tt> member, so it could be any kind of object. It could have a
destructor that causes code to be executed that tries to access the
<tt class="xref docutils literal">first</tt> member. To be paranoid and protect ourselves against this
possibility, we almost always reassign members before decrementing their
reference counts. When don&#8217;t we have to do this?
<ul class="simple">
<li>when we absolutely know that the reference count is greater than 1</li>
<li>when we know that deallocation of the object <a class="footnote-reference" href="#id7" id="id2">[1]</a> will not cause any calls
back into our type&#8217;s code</li>
<li>when decrementing a reference count in a <a title="tp_dealloc" class="reference external" href="../c-api/typeobj.html#tp_dealloc"><tt class="xref docutils literal">tp_dealloc</tt></a> handler when
garbage-collections is not supported <a class="footnote-reference" href="#id8" id="id3">[2]</a></li>
</ul>
We want to expose our instance variables as attributes. There are a
number of ways to do that. The simplest way is to define member definitions:
<div class="highlight-c"><div class="highlight"><pre>static PyMemberDef Noddy_members[] = {
 {&quot;first&quot;, T_OBJECT_EX, offsetof(Noddy, first), 0,
 &quot;first name&quot;},
 {&quot;last&quot;, T_OBJECT_EX, offsetof(Noddy, last), 0,
 &quot;last name&quot;},
 {&quot;number&quot;, T_INT, offsetof(Noddy, number), 0,
 &quot;noddy number&quot;},
 {NULL} /* Sentinel */
};
</pre></div>
</div>
and put the definitions in the <a title="tp_members" class="reference external" href="../c-api/typeobj.html#tp_members"><tt class="xref docutils literal">tp_members</tt></a> slot:
<div class="highlight-c"><div class="highlight"><pre>Noddy_members, /* tp_members */
</pre></div>
</div>
Each member definition has a member name, type, offset, access flags and
documentation string. See the <a class="reference internal" href="#generic-attribute-management">Generic Attribute Management</a> section below for
details.
A disadvantage of this approach is that it doesn&#8217;t provide a way to restrict the
types of objects that can be assigned to the Python attributes. We expect the
first and last names to be strings, but any Python objects can be assigned.
Further, the attributes can be deleted, setting the C pointers to NULL. Even
though we can make sure the members are initialized to non-NULL values, the
members can be set to NULL if the attributes are deleted.
We define a single method, <tt class="xref docutils literal">name()</tt>, that outputs the objects name as the
concatenation of the first and last names.
<div class="highlight-c"><div class="highlight"><pre>static PyObject *
Noddy_name(Noddy* self)
{
 static PyObject *format = NULL;
 PyObject *args, *result;

return result;
}
</pre></div>
</div>
The method is implemented as a C function that takes a <tt class="xref docutils literal">Noddy</tt> (or
<tt class="xref docutils literal">Noddy</tt> subclass) instance as the first argument. Methods always take an
instance as the first argument. Methods often take positional and keyword
arguments as well, but in this cased we don&#8217;t take any and don&#8217;t need to accept
a positional argument tuple or keyword argument dictionary. This method is
equivalent to the Python method:
<div class="highlight-c"><div class="highlight"><pre>def name(self):
 return &quot;%s %s&quot; % (self.first, self.last)
</pre></div>
</div>
Note that we have to check for the possibility that our <tt class="xref docutils literal">first</tt> and
<tt class="xref docutils literal">last</tt> members are NULL. This is because they can be deleted, in which
case they are set to NULL. It would be better to prevent deletion of these
attributes and to restrict the attribute values to be strings. We&#8217;ll see how to
do that in the next section.
Now that we&#8217;ve defined the method, we need to create an array of method
definitions:
<div class="highlight-c"><div class="highlight"><pre>static PyMethodDef Noddy_methods[] = {
 {&quot;name&quot;, (PyCFunction)Noddy_name, METH_NOARGS,
 &quot;Return the name, combining the first and last name&quot;
 },
 {NULL} /* Sentinel */
};
</pre></div>
</div>
and assign them to the <a title="tp_methods" class="reference external" href="../c-api/typeobj.html#tp_methods"><tt class="xref docutils literal">tp_methods</tt></a> slot:
<div class="highlight-c"><div class="highlight"><pre>Noddy_methods, /* tp_methods */
</pre></div>
</div>
Note that we used the <a title="METH_NOARGS" class="reference external" href="../c-api/structures.html#METH_NOARGS"><tt class="xref docutils literal">METH_NOARGS</tt></a> flag to indicate that the method is
passed no arguments.
Finally, we&#8217;ll make our type usable as a base class. We&#8217;ve written our methods
carefully so far so that they don&#8217;t make any assumptions about the type of the
object being created or used, so all we need to do is to add the
<a title="Py_TPFLAGS_BASETYPE" class="reference external" href="../c-api/typeobj.html#Py_TPFLAGS_BASETYPE"><tt class="xref docutils literal">Py_TPFLAGS_BASETYPE</tt></a> to our class flag definition:
<div class="highlight-c"><div class="highlight"><pre>Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/
</pre></div>
</div>
We rename <tt class="xref docutils literal">initnoddy()</tt> to <tt class="xref docutils literal">initnoddy2()</tt> and update the module name
passed to <a title="Py_InitModule3" class="reference external" href="../c-api/allocation.html#Py_InitModule3"><tt class="xref docutils literal">Py_InitModule3()</tt></a>.
Finally, we update our <tt class="docutils literal">setup.py</tt> file to build the new module:
<div class="highlight-c"><div class="highlight"><pre>from distutils.core import setup, Extension
setup(name=&quot;noddy&quot;, version=&quot;1.0&quot;,
 ext_modules=[
 Extension(&quot;noddy&quot;, [&quot;noddy.c&quot;]),
 Extension(&quot;noddy2&quot;, [&quot;noddy2.c&quot;]),
 ])
</pre></div>
</div>
</div>
<div class="section" id="providing-finer-control-over-data-attributes">
<h3>2.1.2. Providing finer control over data attributes<a class="headerlink" href="#providing-finer-control-over-data-attributes" title="Permalink to this headline">¶</a></h3>
In this section, we&#8217;ll provide finer control over how the <tt class="xref docutils literal">first</tt> and
<tt class="xref docutils literal">last</tt> attributes are set in the <tt class="xref docutils literal">Noddy</tt> example. In the previous
version of our module, the instance variables <tt class="xref docutils literal">first</tt> and <tt class="xref docutils literal">last</tt>
could be set to non-string values or even deleted. We want to make sure that
these attributes always contain strings.
<div class="highlight-c"><div class="highlight"><pre>#include &lt;Python.h&gt;
#include &quot;structmember.h&quot;

typedef struct {
 PyObject_HEAD
 PyObject *first;
 PyObject *last;
 int number;
} Noddy;

self-&gt;number = 0;
 }

if (! PyArg_ParseTupleAndKeywords(args, kwds, &quot;|SSi&quot;, kwlist, 
 &amp;first, &amp;last, 
 &amp;self-&gt;number))
 return -1;

if (first) {
 tmp = self-&gt;first;
 Py_INCREF(first);
 self-&gt;first = first;
 Py_DECREF(tmp);
 }

if (last) {
 tmp = self-&gt;last;
 Py_INCREF(last);
 self-&gt;last = last;
 Py_DECREF(tmp);
 }

return 0;
}

static PyObject *
Noddy_getfirst(Noddy *self, void *closure)
{
 Py_INCREF(self-&gt;first);
 return self-&gt;first;
}

static int
Noddy_setfirst(Noddy *self, PyObject *value, void *closure)
{
 if (value == NULL) {
 PyErr_SetString(PyExc_TypeError, &quot;Cannot delete the first attribute&quot;);
 return -1;
 }
 
 if (! PyString_Check(value)) {
 PyErr_SetString(PyExc_TypeError, 
 &quot;The first attribute value must be a string&quot;);
 return -1;
 }
 
 Py_DECREF(self-&gt;first);
 Py_INCREF(value);
 self-&gt;first = value;

return 0;
}

static PyObject *
Noddy_getlast(Noddy *self, void *closure)
{
 Py_INCREF(self-&gt;last);
 return self-&gt;last;
}

static int
Noddy_setlast(Noddy *self, PyObject *value, void *closure)
{
 if (value == NULL) {
 PyErr_SetString(PyExc_TypeError, &quot;Cannot delete the last attribute&quot;);
 return -1;
 }
 
 if (! PyString_Check(value)) {
 PyErr_SetString(PyExc_TypeError, 
 &quot;The last attribute value must be a string&quot;);
 return -1;
 }
 
 Py_DECREF(self-&gt;last);
 Py_INCREF(value);
 self-&gt;last = value;

return 0;
}

static PyGetSetDef Noddy_getseters[] = {
 {&quot;first&quot;, 
 (getter)Noddy_getfirst, (setter)Noddy_setfirst,
 &quot;first name&quot;,
 NULL},
 {&quot;last&quot;, 
 (getter)Noddy_getlast, (setter)Noddy_setlast,
 &quot;last name&quot;,
 NULL},
 {NULL} /* Sentinel */
};

static PyTypeObject NoddyType = {
 PyObject_HEAD_INIT(NULL)
 0, /*ob_size*/
 &quot;noddy.Noddy&quot;, /*tp_name*/
 sizeof(Noddy), /*tp_basicsize*/
 0, /*tp_itemsize*/
 (destructor)Noddy_dealloc, /*tp_dealloc*/
 0, /*tp_print*/
 0, /*tp_getattr*/
 0, /*tp_setattr*/
 0, /*tp_compare*/
 0, /*tp_repr*/
 0, /*tp_as_number*/
 0, /*tp_as_sequence*/
 0, /*tp_as_mapping*/
 0, /*tp_hash */
 0, /*tp_call*/
 0, /*tp_str*/
 0, /*tp_getattro*/
 0, /*tp_setattro*/
 0, /*tp_as_buffer*/
 Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/
 &quot;Noddy objects&quot;, /* tp_doc */
 0,		 /* tp_traverse */
 0,		 /* tp_clear */
 0,		 /* tp_richcompare */
 0,		 /* tp_weaklistoffset */
 0,		 /* tp_iter */
 0,		 /* tp_iternext */
 Noddy_methods, /* tp_methods */
 Noddy_members, /* tp_members */
 Noddy_getseters, /* tp_getset */
 0, /* tp_base */
 0, /* tp_dict */
 0, /* tp_descr_get */
 0, /* tp_descr_set */
 0, /* tp_dictoffset */
 (initproc)Noddy_init, /* tp_init */
 0, /* tp_alloc */
 Noddy_new, /* tp_new */
};

#ifndef PyMODINIT_FUNC	/* declarations for DLL import/export */
#define PyMODINIT_FUNC void
#endif
PyMODINIT_FUNC
initnoddy3(void) 
{
 PyObject* m;

m = Py_InitModule3(&quot;noddy3&quot;, module_methods,
 &quot;Example module that creates an extension type.&quot;);

static int
Noddy_setfirst(Noddy *self, PyObject *value, void *closure)
{
 if (value == NULL) {
 PyErr_SetString(PyExc_TypeError, &quot;Cannot delete the first attribute&quot;);
 return -1;
 }

if (! PyString_Check(value)) {
 PyErr_SetString(PyExc_TypeError,
 &quot;The first attribute value must be a string&quot;);
 return -1;
 }

Py_DECREF(self-&gt;first);
 Py_INCREF(value);
 self-&gt;first = value;

return 0;
}
</pre></div>
</div>
The getter function is passed a <tt class="xref docutils literal">Noddy</tt> object and a &#8220;closure&#8221;, which is
void pointer. In this case, the closure is ignored. (The closure supports an
advanced usage in which definition data is passed to the getter and setter. This
could, for example, be used to allow a single set of getter and setter functions
that decide the attribute to get or set based on data in the closure.)
The setter function is passed the <tt class="xref docutils literal">Noddy</tt> object, the new value, and the
closure. The new value may be NULL, in which case the attribute is being
deleted. In our setter, we raise an error if the attribute is deleted or if the
attribute value is not a string.
We create an array of <tt class="xref docutils literal">PyGetSetDef</tt> structures:
<div class="highlight-c"><div class="highlight"><pre>static PyGetSetDef Noddy_getseters[] = {
 {&quot;first&quot;,
 (getter)Noddy_getfirst, (setter)Noddy_setfirst,
 &quot;first name&quot;,
 NULL},
 {&quot;last&quot;,
 (getter)Noddy_getlast, (setter)Noddy_setlast,
 &quot;last name&quot;,
 NULL},
 {NULL} /* Sentinel */
};
</pre></div>
</div>
and register it in the <a title="tp_getset" class="reference external" href="../c-api/typeobj.html#tp_getset"><tt class="xref docutils literal">tp_getset</tt></a> slot:
<div class="highlight-c"><div class="highlight"><pre>Noddy_getseters, /* tp_getset */
</pre></div>
</div>
to register our attribute getters and setters.
The last item in a <tt class="xref docutils literal">PyGetSetDef</tt> structure is the closure mentioned
above. In this case, we aren&#8217;t using the closure, so we just pass NULL.
We also remove the member definitions for these attributes:
<div class="highlight-c"><div class="highlight"><pre>static PyMemberDef Noddy_members[] = {
 {&quot;number&quot;, T_INT, offsetof(Noddy, number), 0,
 &quot;noddy number&quot;},
 {NULL} /* Sentinel */
};
</pre></div>
</div>
We also need to update the <a title="tp_init" class="reference external" href="../c-api/typeobj.html#tp_init"><tt class="xref docutils literal">tp_init</tt></a> handler to only allow strings <a class="footnote-reference" href="#id9" id="id4">[3]</a> to
be passed:
<div class="highlight-c"><div class="highlight"><pre>static int
Noddy_init(Noddy *self, PyObject *args, PyObject *kwds)
{
 PyObject *first=NULL, *last=NULL, *tmp;

if (! PyArg_ParseTupleAndKeywords(args, kwds, &quot;|SSi&quot;, kwlist,
 &amp;first, &amp;last,
 &amp;self-&gt;number))
 return -1;

if (first) {
 tmp = self-&gt;first;
 Py_INCREF(first);
 self-&gt;first = first;
 Py_DECREF(tmp);
 }

if (last) {
 tmp = self-&gt;last;
 Py_INCREF(last);
 self-&gt;last = last;
 Py_DECREF(tmp);
 }

return 0;
}
</pre></div>
</div>
With these changes, we can assure that the <tt class="xref docutils literal">first</tt> and <tt class="xref docutils literal">last</tt>
members are never NULL so we can remove checks for NULL values in almost all
cases. This means that most of the <a title="Py_XDECREF" class="reference external" href="../c-api/refcounting.html#Py_XDECREF"><tt class="xref docutils literal">Py_XDECREF()</tt></a> calls can be converted to
<a title="Py_DECREF" class="reference external" href="../c-api/refcounting.html#Py_DECREF"><tt class="xref docutils literal">Py_DECREF()</tt></a> calls. The only place we can&#8217;t change these calls is in the
deallocator, where there is the possibility that the initialization of these
members failed in the constructor.
We also rename the module initialization function and module name in the
initialization function, as we did before, and we add an extra definition to the
<tt class="docutils literal">setup.py</tt> file.
</div>
<div class="section" id="supporting-cyclic-garbage-collection">
<h3>2.1.3. Supporting cyclic garbage collection<a class="headerlink" href="#supporting-cyclic-garbage-collection" title="Permalink to this headline">¶</a></h3>
Python has a cyclic-garbage collector that can identify unneeded objects even
when their reference counts are not zero. This can happen when objects are
involved in cycles. For example, consider:
<div class="highlight-c"><div class="highlight"><pre>&gt;&gt;&gt; l = []
&gt;&gt;&gt; l.append(l)
&gt;&gt;&gt; del l
</pre></div>
</div>
In this example, we create a list that contains itself. When we delete it, it
still has a reference from itself. Its reference count doesn&#8217;t drop to zero.
Fortunately, Python&#8217;s cyclic-garbage collector will eventually figure out that
the list is garbage and free it.
In the second version of the <tt class="xref docutils literal">Noddy</tt> example, we allowed any kind of
object to be stored in the <tt class="xref docutils literal">first</tt> or <tt class="xref docutils literal">last</tt> attributes. <a class="footnote-reference" href="#id10" id="id5">[4]</a> This
means that <tt class="xref docutils literal">Noddy</tt> objects can participate in cycles:
<div class="highlight-c"><div class="highlight"><pre>&gt;&gt;&gt; import noddy2
&gt;&gt;&gt; n = noddy2.Noddy()
&gt;&gt;&gt; l = [n]
&gt;&gt;&gt; n.first = l
</pre></div>
</div>
This is pretty silly, but it gives us an excuse to add support for the
cyclic-garbage collector to the <tt class="xref docutils literal">Noddy</tt> example. To support cyclic
garbage collection, types need to fill two slots and set a class flag that
enables these slots:
<div class="highlight-c"><div class="highlight"><pre>#include &lt;Python.h&gt;
#include &quot;structmember.h&quot;

typedef struct {
 PyObject_HEAD
 PyObject *first;
 PyObject *last;
 int number;
} Noddy;

static int
Noddy_traverse(Noddy *self, visitproc visit, void *arg)
{
 int vret;

if (self-&gt;first) {
 vret = visit(self-&gt;first, arg);
 if (vret != 0)
 return vret;
 }
 if (self-&gt;last) {
 vret = visit(self-&gt;last, arg);
 if (vret != 0)
 return vret;
 }

return 0;
}

static int 
Noddy_clear(Noddy *self)
{
 PyObject *tmp;

tmp = self-&gt;first;
 self-&gt;first = NULL;
 Py_XDECREF(tmp);

tmp = self-&gt;last;
 self-&gt;last = NULL;
 Py_XDECREF(tmp);

return 0;
}

self-&gt;number = 0;
 }

return 0;
}

static PyTypeObject NoddyType = {
 PyObject_HEAD_INIT(NULL)
 0, /*ob_size*/
 &quot;noddy.Noddy&quot;, /*tp_name*/
 sizeof(Noddy), /*tp_basicsize*/
 0, /*tp_itemsize*/
 (destructor)Noddy_dealloc, /*tp_dealloc*/
 0, /*tp_print*/
 0, /*tp_getattr*/
 0, /*tp_setattr*/
 0, /*tp_compare*/
 0, /*tp_repr*/
 0, /*tp_as_number*/
 0, /*tp_as_sequence*/
 0, /*tp_as_mapping*/
 0, /*tp_hash */
 0, /*tp_call*/
 0, /*tp_str*/
 0, /*tp_getattro*/
 0, /*tp_setattro*/
 0, /*tp_as_buffer*/
 Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE | Py_TPFLAGS_HAVE_GC, /*tp_flags*/
 &quot;Noddy objects&quot;, /* tp_doc */
 (traverseproc)Noddy_traverse, /* tp_traverse */
 (inquiry)Noddy_clear, /* tp_clear */
 0,		 /* tp_richcompare */
 0,		 /* tp_weaklistoffset */
 0,		 /* tp_iter */
 0,		 /* tp_iternext */
 Noddy_methods, /* tp_methods */
 Noddy_members, /* tp_members */
 0, /* tp_getset */
 0, /* tp_base */
 0, /* tp_dict */
 0, /* tp_descr_get */
 0, /* tp_descr_set */
 0, /* tp_dictoffset */
 (initproc)Noddy_init, /* tp_init */
 0, /* tp_alloc */
 Noddy_new, /* tp_new */
};

#ifndef PyMODINIT_FUNC	/* declarations for DLL import/export */
#define PyMODINIT_FUNC void
#endif
PyMODINIT_FUNC
initnoddy4(void) 
{
 PyObject* m;

m = Py_InitModule3(&quot;noddy4&quot;, module_methods,
 &quot;Example module that creates an extension type.&quot;);

return 0;
}
</pre></div>
</div>
For each subobject that can participate in cycles, we need to call the
<tt class="xref docutils literal">visit()</tt> function, which is passed to the traversal method. The
<tt class="xref docutils literal">visit()</tt> function takes as arguments the subobject and the extra argument
arg passed to the traversal method. It returns an integer value that must be
returned if it is non-zero.
Python 2.4 and higher provide a <a title="Py_VISIT" class="reference external" href="../c-api/gcsupport.html#Py_VISIT"><tt class="xref docutils literal">Py_VISIT()</tt></a> macro that automates calling
visit functions. With <a title="Py_VISIT" class="reference external" href="../c-api/gcsupport.html#Py_VISIT"><tt class="xref docutils literal">Py_VISIT()</tt></a>, <tt class="xref docutils literal">Noddy_traverse()</tt> can be
simplified:
<div class="highlight-c"><div class="highlight"><pre>static int
Noddy_traverse(Noddy *self, visitproc visit, void *arg)
{
 Py_VISIT(self-&gt;first);
 Py_VISIT(self-&gt;last);
 return 0;
}
</pre></div>
</div>
<div class="admonition note">
Note
Note that the <a title="tp_traverse" class="reference external" href="../c-api/typeobj.html#tp_traverse"><tt class="xref docutils literal">tp_traverse</tt></a> implementation must name its arguments exactly
visit and arg in order to use <a title="Py_VISIT" class="reference external" href="../c-api/gcsupport.html#Py_VISIT"><tt class="xref docutils literal">Py_VISIT()</tt></a>. This is to encourage
uniformity across these boring implementations.
</div>
We also need to provide a method for clearing any subobjects that can
participate in cycles. We implement the method and reimplement the deallocator
to use it:
<div class="highlight-c"><div class="highlight"><pre>static int
Noddy_clear(Noddy *self)
{
 PyObject *tmp;

return 0;
}

static void
Noddy_dealloc(Noddy* self)
{
 Noddy_clear(self);
 self-&gt;ob_type-&gt;tp_free((PyObject*)self);
}
</pre></div>
</div>
Notice the use of a temporary variable in <tt class="xref docutils literal">Noddy_clear()</tt>. We use the
temporary variable so that we can set each member to NULL before decrementing
its reference count. We do this because, as was discussed earlier, if the
reference count drops to zero, we might cause code to run that calls back into
the object. In addition, because we now support garbage collection, we also
have to worry about code being run that triggers garbage collection. If garbage
collection is run, our <a title="tp_traverse" class="reference external" href="../c-api/typeobj.html#tp_traverse"><tt class="xref docutils literal">tp_traverse</tt></a> handler could get called. We can&#8217;t
take a chance of having <tt class="xref docutils literal">Noddy_traverse()</tt> called when a member&#8217;s reference
count has dropped to zero and its value hasn&#8217;t been set to NULL.
Python 2.4 and higher provide a <a title="Py_CLEAR" class="reference external" href="../c-api/refcounting.html#Py_CLEAR"><tt class="xref docutils literal">Py_CLEAR()</tt></a> that automates the careful
decrementing of reference counts. With <a title="Py_CLEAR" class="reference external" href="../c-api/refcounting.html#Py_CLEAR"><tt class="xref docutils literal">Py_CLEAR()</tt></a>, the
<tt class="xref docutils literal">Noddy_clear()</tt> function can be simplified:
<div class="highlight-c"><div class="highlight"><pre>static int
Noddy_clear(Noddy *self)
{
 Py_CLEAR(self-&gt;first);
 Py_CLEAR(self-&gt;last);
 return 0;
}
</pre></div>
</div>
Finally, we add the <a title="Py_TPFLAGS_HAVE_GC" class="reference external" href="../c-api/typeobj.html#Py_TPFLAGS_HAVE_GC"><tt class="xref docutils literal">Py_TPFLAGS_HAVE_GC</tt></a> flag to the class flags:
<div class="highlight-c"><div class="highlight"><pre>Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE | Py_TPFLAGS_HAVE_GC, /*tp_flags*/
</pre></div>
</div>
That&#8217;s pretty much it. If we had written custom <a title="tp_alloc" class="reference external" href="../c-api/typeobj.html#tp_alloc"><tt class="xref docutils literal">tp_alloc</tt></a> or
<a title="tp_free" class="reference external" href="../c-api/typeobj.html#tp_free"><tt class="xref docutils literal">tp_free</tt></a> slots, we&#8217;d need to modify them for cyclic-garbage collection.
Most extensions will use the versions automatically provided.
</div>
<div class="section" id="subclassing-other-types">
<h3>2.1.4. Subclassing other types<a class="headerlink" href="#subclassing-other-types" title="Permalink to this headline">¶</a></h3>
It is possible to create new extension types that are derived from existing
types. It is easiest to inherit from the built in types, since an extension can
easily use the <a title="PyTypeObject" class="reference external" href="../c-api/type.html#PyTypeObject"><tt class="xref docutils literal">PyTypeObject</tt></a> it needs. It can be difficult to share
these <a title="PyTypeObject" class="reference external" href="../c-api/type.html#PyTypeObject"><tt class="xref docutils literal">PyTypeObject</tt></a> structures between extension modules.
In this example we will create a <tt class="xref docutils literal">Shoddy</tt> type that inherits from the
built-in <a title="list" class="reference external" href="../library/functions.html#list"><tt class="xref docutils literal">list</tt></a> type. The new type will be completely compatible with
regular lists, but will have an additional <tt class="xref docutils literal">increment()</tt> method that
increases an internal counter.
<div class="highlight-c"><div class="highlight"><pre>&gt;&gt;&gt; import shoddy
&gt;&gt;&gt; s = shoddy.Shoddy(range(3))
&gt;&gt;&gt; s.extend(s)
&gt;&gt;&gt; print len(s)
6
&gt;&gt;&gt; print s.increment()
1
&gt;&gt;&gt; print s.increment()
2
</pre></div>
</div>
<div class="highlight-c"><div class="highlight"><pre>#include &lt;Python.h&gt;

typedef struct {
 PyListObject list;
 int state;
} Shoddy;

static PyObject *
Shoddy_increment(Shoddy *self, PyObject *unused)
{
 self-&gt;state++;
 return PyInt_FromLong(self-&gt;state);
}

static PyMethodDef Shoddy_methods[] = {
 {&quot;increment&quot;, (PyCFunction)Shoddy_increment, METH_NOARGS,
 PyDoc_STR(&quot;increment state counter&quot;)},
 {NULL,	NULL},
};

static int
Shoddy_init(Shoddy *self, PyObject *args, PyObject *kwds)
{
 if (PyList_Type.tp_init((PyObject *)self, args, kwds) &lt; 0)
 return -1;
 self-&gt;state = 0;
 return 0;
}

static PyTypeObject ShoddyType = {
 PyObject_HEAD_INIT(NULL)
 0, /* ob_size */
 &quot;shoddy.Shoddy&quot;, /* tp_name */
 sizeof(Shoddy), /* tp_basicsize */
 0, /* tp_itemsize */
 0, /* tp_dealloc */
 0, /* tp_print */
 0, /* tp_getattr */
 0, /* tp_setattr */
 0, /* tp_compare */
 0, /* tp_repr */
 0, /* tp_as_number */
 0, /* tp_as_sequence */
 0, /* tp_as_mapping */
 0, /* tp_hash */
 0, /* tp_call */
 0, /* tp_str */
 0, /* tp_getattro */
 0, /* tp_setattro */
 0, /* tp_as_buffer */
 Py_TPFLAGS_DEFAULT |
 Py_TPFLAGS_BASETYPE, /* tp_flags */
 0, /* tp_doc */
 0, /* tp_traverse */
 0, /* tp_clear */
 0, /* tp_richcompare */
 0, /* tp_weaklistoffset */
 0, /* tp_iter */
 0, /* tp_iternext */
 Shoddy_methods, /* tp_methods */
 0, /* tp_members */
 0, /* tp_getset */
 0, /* tp_base */
 0, /* tp_dict */
 0, /* tp_descr_get */
 0, /* tp_descr_set */
 0, /* tp_dictoffset */
 (initproc)Shoddy_init, /* tp_init */
 0, /* tp_alloc */
 0, /* tp_new */
};

PyMODINIT_FUNC
initshoddy(void)
{
 PyObject *m;

ShoddyType.tp_base = &amp;PyList_Type;
 if (PyType_Ready(&amp;ShoddyType) &lt; 0)
 return;

m = Py_InitModule3(&quot;shoddy&quot;, NULL, &quot;Shoddy module&quot;);
 if (m == NULL)
 return;

Py_INCREF(&amp;ShoddyType);
 PyModule_AddObject(m, &quot;Shoddy&quot;, (PyObject *) &amp;ShoddyType);
}
</pre></div>
</div>
As you can see, the source code closely resembles the <tt class="xref docutils literal">Noddy</tt> examples in
previous sections. We will break down the main differences between them.
<div class="highlight-c"><div class="highlight"><pre>typedef struct {
 PyListObject list;
 int state;
} Shoddy;
</pre></div>
</div>
The primary difference for derived type objects is that the base type&#8217;s object
structure must be the first value. The base type will already include the
<a title="PyObject_HEAD" class="reference external" href="../c-api/structures.html#PyObject_HEAD"><tt class="xref docutils literal">PyObject_HEAD()</tt></a> at the beginning of its structure.
When a Python object is a <tt class="xref docutils literal">Shoddy</tt> instance, its PyObject* pointer can
be safely cast to both PyListObject* and Shoddy*.
<div class="highlight-c"><div class="highlight"><pre>static int
Shoddy_init(Shoddy *self, PyObject *args, PyObject *kwds)
{
 if (PyList_Type.tp_init((PyObject *)self, args, kwds) &lt; 0)
 return -1;
 self-&gt;state = 0;
 return 0;
}
</pre></div>
</div>
In the <tt class="xref docutils literal">__init__</tt> method for our type, we can see how to call through to
the <tt class="xref docutils literal">__init__</tt> method of the base type.
This pattern is important when writing a type with custom <tt class="xref docutils literal">new</tt> and
<tt class="xref docutils literal">dealloc</tt> methods. The <tt class="xref docutils literal">new</tt> method should not actually create the
memory for the object with <a title="tp_alloc" class="reference external" href="../c-api/typeobj.html#tp_alloc"><tt class="xref docutils literal">tp_alloc</tt></a>, that will be handled by the base
class when calling its <a title="tp_new" class="reference external" href="../c-api/typeobj.html#tp_new"><tt class="xref docutils literal">tp_new</tt></a>.
When filling out the <a title="PyTypeObject" class="reference external" href="../c-api/type.html#PyTypeObject"><tt class="xref docutils literal">PyTypeObject()</tt></a> for the <tt class="xref docutils literal">Shoddy</tt> type, you see
a slot for <a title="tp_base" class="reference external" href="../c-api/typeobj.html#tp_base"><tt class="xref docutils literal">tp_base()</tt></a>. Due to cross platform compiler issues, you can&#8217;t
fill that field directly with the <a title="PyList_Type" class="reference external" href="../c-api/list.html#PyList_Type"><tt class="xref docutils literal">PyList_Type()</tt></a>; it can be done later in
the module&#8217;s <tt class="xref docutils literal">init()</tt> function.
<div class="highlight-c"><div class="highlight"><pre>PyMODINIT_FUNC
initshoddy(void)
{
 PyObject *m;

/* Methods to implement standard operations */

destructor tp_dealloc;
 printfunc tp_print;
 getattrfunc tp_getattr;
 setattrfunc tp_setattr;
 cmpfunc tp_compare;
 reprfunc tp_repr;

/* Method suites for standard classes */

PyNumberMethods *tp_as_number;
 PySequenceMethods *tp_as_sequence;
 PyMappingMethods *tp_as_mapping;

/* More standard operations (here for binary compatibility) */

hashfunc tp_hash;
 ternaryfunc tp_call;
 reprfunc tp_str;
 getattrofunc tp_getattro;
 setattrofunc tp_setattro;

/* Functions to access object as input/output buffer */
 PyBufferProcs *tp_as_buffer;

/* Flags to define presence of optional/expanded features */
 long tp_flags;

char *tp_doc; /* Documentation string */

/* Assigned meaning in release 2.0 */
 /* call function for all accessible objects */
 traverseproc tp_traverse;

/* delete references to contained objects */
 inquiry tp_clear;

/* Assigned meaning in release 2.1 */
 /* rich comparisons */
 richcmpfunc tp_richcompare;

/* weak reference enabler */
 long tp_weaklistoffset;

/* Added in release 2.2 */
 /* Iterators */
 getiterfunc tp_iter;
 iternextfunc tp_iternext;

/* Attribute descriptor and subclassing stuff */
 struct PyMethodDef *tp_methods;
 struct PyMemberDef *tp_members;
 struct PyGetSetDef *tp_getset;
 struct _typeobject *tp_base;
 PyObject *tp_dict;
 descrgetfunc tp_descr_get;
 descrsetfunc tp_descr_set;
 long tp_dictoffset;
 initproc tp_init;
 allocfunc tp_alloc;
 newfunc tp_new;
 freefunc tp_free; /* Low-level free-memory routine */
 inquiry tp_is_gc; /* For PyObject_IS_GC */
 PyObject *tp_bases;
 PyObject *tp_mro; /* method resolution order */
 PyObject *tp_cache;
 PyObject *tp_subclasses;
 PyObject *tp_weaklist;

} PyTypeObject;
</pre></div>
</div>
Now that&#8217;s a lot of methods. Don&#8217;t worry too much though - if you have a type
you want to define, the chances are very good that you will only implement a
handful of these.
As you probably expect by now, we&#8217;re going to go over this and give more
information about the various handlers. We won&#8217;t go in the order they are
defined in the structure, because there is a lot of historical baggage that
impacts the ordering of the fields; be sure your type initialization keeps the
fields in the right order! It&#8217;s often easiest to find an example that includes
all the fields you need (even if they&#8217;re initialized to <tt class="docutils literal">0</tt>) and then change
the values to suit your new type.
<div class="highlight-c"><div class="highlight"><pre>char *tp_name; /* For printing */
</pre></div>
</div>
The name of the type - as mentioned in the last section, this will appear in
various places, almost entirely for diagnostic purposes. Try to choose something
that will be helpful in such a situation!
<div class="highlight-c"><div class="highlight"><pre>int tp_basicsize, tp_itemsize; /* For allocation */
</pre></div>
</div>
These fields tell the runtime how much memory to allocate when new objects of
this type are created. Python has some built-in support for variable length
structures (think: strings, lists) which is where the <a title="tp_itemsize" class="reference external" href="../c-api/typeobj.html#tp_itemsize"><tt class="xref docutils literal">tp_itemsize</tt></a> field
comes in. This will be dealt with later.
<div class="highlight-c"><div class="highlight"><pre>char *tp_doc;
</pre></div>
</div>
Here you can put a string (or its address) that you want returned when the
Python script references <tt class="docutils literal">obj.__doc__</tt> to retrieve the doc string.
Now we come to the basic type methods&#8212;the ones most extension types will
implement.
<div class="section" id="finalization-and-de-allocation">
<h3>2.2.1. Finalization and De-allocation<a class="headerlink" href="#finalization-and-de-allocation" title="Permalink to this headline">¶</a></h3>
<div class="highlight-c" id="index-170"><div class="highlight"><pre>destructor tp_dealloc;
</pre></div>
</div>
This function is called when the reference count of the instance of your type is
reduced to zero and the Python interpreter wants to reclaim it. If your type
has memory to free or other clean-up to perform, put it here. The object itself
needs to be freed here as well. Here is an example of this function:
<div class="highlight-c"><div class="highlight"><pre>static void
newdatatype_dealloc(newdatatypeobject * obj)
{
 free(obj-&gt;obj_UnderlyingDatatypePtr);
 obj-&gt;ob_type-&gt;tp_free(obj);
}
</pre></div>
</div>
One important requirement of the deallocator function is that it leaves any
pending exceptions alone. This is important since deallocators are frequently
called as the interpreter unwinds the Python stack; when the stack is unwound
due to an exception (rather than normal returns), nothing is done to protect the
deallocators from seeing that an exception has already been set. Any actions
which a deallocator performs which may cause additional Python code to be
executed may detect that an exception has been set. This can lead to misleading
errors from the interpreter. The proper way to protect against this is to save
a pending exception before performing the unsafe action, and restoring it when
done. This can be done using the <a title="PyErr_Fetch" class="reference external" href="../c-api/exceptions.html#PyErr_Fetch"><tt class="xref docutils literal">PyErr_Fetch()</tt></a> and
<a title="PyErr_Restore" class="reference external" href="../c-api/exceptions.html#PyErr_Restore"><tt class="xref docutils literal">PyErr_Restore()</tt></a> functions:
<div class="highlight-c"><div class="highlight"><pre>static void
my_dealloc(PyObject *obj)
{
 MyObject *self = (MyObject *) obj;
 PyObject *cbresult;

if (self-&gt;my_callback != NULL) {
 PyObject *err_type, *err_value, *err_traceback;
 int have_error = PyErr_Occurred() ? 1 : 0;

if (have_error)
 PyErr_Fetch(&amp;err_type, &amp;err_value, &amp;err_traceback);

cbresult = PyObject_CallObject(self-&gt;my_callback, NULL);
 if (cbresult == NULL)
 PyErr_WriteUnraisable(self-&gt;my_callback);
 else
 Py_DECREF(cbresult);

if (have_error)
 PyErr_Restore(err_type, err_value, err_traceback);

Py_DECREF(self-&gt;my_callback);
 }
 obj-&gt;ob_type-&gt;tp_free((PyObject*)self);
}
</pre></div>
</div>
</div>
<div class="section" id="object-presentation">
<h3>2.2.2. Object Presentation<a class="headerlink" href="#object-presentation" title="Permalink to this headline">¶</a></h3>
In Python, there are three ways to generate a textual representation of an
object: the <a title="repr" class="reference external" href="../library/functions.html#repr"><tt class="xref docutils literal">repr()</tt></a> function (or equivalent back-tick syntax), the
<a title="str" class="reference external" href="../library/functions.html#str"><tt class="xref docutils literal">str()</tt></a> function, and the <a class="reference external" href="../reference/simple_stmts.html#print"><tt class="xref docutils literal">print</tt></a> statement. For most objects, the
<a class="reference external" href="../reference/simple_stmts.html#print"><tt class="xref docutils literal">print</tt></a> statement is equivalent to the <a title="str" class="reference external" href="../library/functions.html#str"><tt class="xref docutils literal">str()</tt></a> function, but it is
possible to special-case printing to a <tt class="xref docutils literal">FILE*</tt> if necessary; this should
only be done if efficiency is identified as a problem and profiling suggests
that creating a temporary string object to be written to a file is too
expensive.
These handlers are all optional, and most types at most need to implement the
<a title="tp_str" class="reference external" href="../c-api/typeobj.html#tp_str"><tt class="xref docutils literal">tp_str</tt></a> and <a title="tp_repr" class="reference external" href="../c-api/typeobj.html#tp_repr"><tt class="xref docutils literal">tp_repr</tt></a> handlers.
<div class="highlight-c"><div class="highlight"><pre>reprfunc tp_repr;
reprfunc tp_str;
printfunc tp_print;
</pre></div>
</div>
The <a title="tp_repr" class="reference external" href="../c-api/typeobj.html#tp_repr"><tt class="xref docutils literal">tp_repr</tt></a> handler should return a string object containing a
representation of the instance for which it is called. Here is a simple
example:
<div class="highlight-c"><div class="highlight"><pre>static PyObject *
newdatatype_repr(newdatatypeobject * obj)
{
 return PyString_FromFormat(&quot;Repr-ified_newdatatype{{size:\%d}}&quot;,
 obj-&gt;obj_UnderlyingDatatypePtr-&gt;size);
}
</pre></div>
</div>
If no <a title="tp_repr" class="reference external" href="../c-api/typeobj.html#tp_repr"><tt class="xref docutils literal">tp_repr</tt></a> handler is specified, the interpreter will supply a
representation that uses the type&#8217;s <a title="tp_name" class="reference external" href="../c-api/typeobj.html#tp_name"><tt class="xref docutils literal">tp_name</tt></a> and a uniquely-identifying
value for the object.
The <a title="tp_str" class="reference external" href="../c-api/typeobj.html#tp_str"><tt class="xref docutils literal">tp_str</tt></a> handler is to <a title="str" class="reference external" href="../library/functions.html#str"><tt class="xref docutils literal">str()</tt></a> what the <a title="tp_repr" class="reference external" href="../c-api/typeobj.html#tp_repr"><tt class="xref docutils literal">tp_repr</tt></a> handler
described above is to <a title="repr" class="reference external" href="../library/functions.html#repr"><tt class="xref docutils literal">repr()</tt></a>; that is, it is called when Python code calls
<a title="str" class="reference external" href="../library/functions.html#str"><tt class="xref docutils literal">str()</tt></a> on an instance of your object. Its implementation is very similar
to the <a title="tp_repr" class="reference external" href="../c-api/typeobj.html#tp_repr"><tt class="xref docutils literal">tp_repr</tt></a> function, but the resulting string is intended for human
consumption. If <a title="tp_str" class="reference external" href="../c-api/typeobj.html#tp_str"><tt class="xref docutils literal">tp_str</tt></a> is not specified, the <a title="tp_repr" class="reference external" href="../c-api/typeobj.html#tp_repr"><tt class="xref docutils literal">tp_repr</tt></a> handler is
used instead.
Here is a simple example:
<div class="highlight-c"><div class="highlight"><pre>static PyObject *
newdatatype_str(newdatatypeobject * obj)
{
 return PyString_FromFormat(&quot;Stringified_newdatatype{{size:\%d}}&quot;,
 obj-&gt;obj_UnderlyingDatatypePtr-&gt;size);
}
</pre></div>
</div>
The print function will be called whenever Python needs to &#8220;print&#8221; an instance
of the type. For example, if &#8216;node&#8217; is an instance of type TreeNode, then the
print function is called when Python code calls:
<div class="highlight-c"><div class="highlight"><pre>print node
</pre></div>
</div>
There is a flags argument and one flag, <tt class="xref docutils literal">Py_PRINT_RAW</tt>, and it suggests
that you print without string quotes and possibly without interpreting escape
sequences.
The print function receives a file object as an argument. You will likely want
to write to that file object.
Here is a sample print function:
<div class="highlight-c"><div class="highlight"><pre>static int
newdatatype_print(newdatatypeobject *obj, FILE *fp, int flags)
{
 if (flags &amp; Py_PRINT_RAW) {
 fprintf(fp, &quot;&lt;{newdatatype object--size: %d}&gt;&quot;,
 obj-&gt;obj_UnderlyingDatatypePtr-&gt;size);
 }
 else {
 fprintf(fp, &quot;\&quot;&lt;{newdatatype object--size: %d}&gt;\&quot;&quot;,
 obj-&gt;obj_UnderlyingDatatypePtr-&gt;size);
 }
 return 0;
}
</pre></div>
</div>
</div>
<div class="section" id="attribute-management">
<h3>2.2.3. Attribute Management<a class="headerlink" href="#attribute-management" title="Permalink to this headline">¶</a></h3>
For every object which can support attributes, the corresponding type must
provide the functions that control how the attributes are resolved. There needs
to be a function which can retrieve attributes (if any are defined), and another
to set attributes (if setting attributes is allowed). Removing an attribute is
a special case, for which the new value passed to the handler is NULL.
Python supports two pairs of attribute handlers; a type that supports attributes
only needs to implement the functions for one pair. The difference is that one
pair takes the name of the attribute as a <tt class="xref docutils literal">char*</tt>, while the other
accepts a <a title="PyObject" class="reference external" href="../c-api/structures.html#PyObject"><tt class="xref docutils literal">PyObject*</tt></a>. Each type can use whichever pair makes more
sense for the implementation&#8217;s convenience.
<div class="highlight-c"><div class="highlight"><pre>getattrfunc tp_getattr; /* char * version */
setattrfunc tp_setattr;
/* ... */
getattrofunc tp_getattrofunc; /* PyObject * version */
setattrofunc tp_setattrofunc;
</pre></div>
</div>
If accessing attributes of an object is always a simple operation (this will be
explained shortly), there are generic implementations which can be used to
provide the <a title="PyObject" class="reference external" href="../c-api/structures.html#PyObject"><tt class="xref docutils literal">PyObject*</tt></a> version of the attribute management functions.
The actual need for type-specific attribute handlers almost completely
disappeared starting with Python 2.2, though there are many examples which have
not been updated to use some of the new generic mechanism that is available.
<div class="section" id="generic-attribute-management">
<h4>2.2.3.1. Generic Attribute Management<a class="headerlink" href="#generic-attribute-management" title="Permalink to this headline">¶</a></h4>

New in version 2.2.
Most extension types only use simple attributes. So, what makes the
attributes simple? There are only a couple of conditions that must be met:
<ol class="arabic simple">
<li>The name of the attributes must be known when <a title="PyType_Ready" class="reference external" href="../c-api/type.html#PyType_Ready"><tt class="xref docutils literal">PyType_Ready()</tt></a> is
called.</li>
<li>No special processing is needed to record that an attribute was looked up or
set, nor do actions need to be taken based on the value.</li>
</ol>
Note that this list does not place any restrictions on the values of the
attributes, when the values are computed, or how relevant data is stored.
When <a title="PyType_Ready" class="reference external" href="../c-api/type.html#PyType_Ready"><tt class="xref docutils literal">PyType_Ready()</tt></a> is called, it uses three tables referenced by the
type object to create <a class="reference external" href="../glossary.html#term-descriptor">descriptor</a>s which are placed in the dictionary of the
type object. Each descriptor controls access to one attribute of the instance
object. Each of the tables is optional; if all three are NULL, instances of
the type will only have attributes that are inherited from their base type, and
should leave the <a title="tp_getattro" class="reference external" href="../c-api/typeobj.html#tp_getattro"><tt class="xref docutils literal">tp_getattro</tt></a> and <a title="tp_setattro" class="reference external" href="../c-api/typeobj.html#tp_setattro"><tt class="xref docutils literal">tp_setattro</tt></a> fields NULL as
well, allowing the base type to handle attributes.
The tables are declared as three fields of the type object:
<div class="highlight-c"><div class="highlight"><pre>struct PyMethodDef *tp_methods;
struct PyMemberDef *tp_members;
struct PyGetSetDef *tp_getset;
</pre></div>
</div>
If <a title="tp_methods" class="reference external" href="../c-api/typeobj.html#tp_methods"><tt class="xref docutils literal">tp_methods</tt></a> is not NULL, it must refer to an array of
<a title="PyMethodDef" class="reference external" href="../c-api/structures.html#PyMethodDef"><tt class="xref docutils literal">PyMethodDef</tt></a> structures. Each entry in the table is an instance of this
structure:
<div class="highlight-c"><div class="highlight"><pre>typedef struct PyMethodDef {
 char *ml_name; /* method name */
 PyCFunction ml_meth; /* implementation function */
 int ml_flags; /* flags */
 char *ml_doc; /* docstring */
} PyMethodDef;
</pre></div>
</div>
One entry should be defined for each method provided by the type; no entries are
needed for methods inherited from a base type. One additional entry is needed
at the end; it is a sentinel that marks the end of the array. The
<tt class="xref docutils literal">ml_name</tt> field of the sentinel must be NULL.
XXX Need to refer to some unified discussion of the structure fields, shared
with the next section.
The second table is used to define attributes which map directly to data stored
in the instance. A variety of primitive C types are supported, and access may
be read-only or read-write. The structures in the table are defined as:
<div class="highlight-c"><div class="highlight"><pre>typedef struct PyMemberDef {
 char *name;
 int type;
 int offset;
 int flags;
 char *doc;
} PyMemberDef;
</pre></div>
</div>
For each entry in the table, a <a class="reference external" href="../glossary.html#term-descriptor">descriptor</a> will be constructed and added to the
type which will be able to extract a value from the instance structure. The
<a title="type" class="reference external" href="../library/functions.html#type"><tt class="xref docutils literal">type</tt></a> field should contain one of the type codes defined in the
<tt class="docutils literal">structmember.h</tt> header; the value will be used to determine how to
convert Python values to and from C values. The <tt class="xref docutils literal">flags</tt> field is used to
store flags which control how the attribute can be accessed.
XXX Need to move some of this to a shared section!
The following flag constants are defined in <tt class="docutils literal">structmember.h</tt>; they may be
combined using bitwise-OR.
<table border="1" class="docutils">
<colgroup>
<col width="37%" />
<col width="63%" />
</colgroup>
<thead valign="bottom">
<tr><th class="head">Constant</th>
<th class="head">Meaning</th>
</tr>
</thead>
<tbody valign="top">
<tr><td><tt class="xref docutils literal">READONLY</tt></td>
<td>Never writable.</td>
</tr>
<tr><td><tt class="xref docutils literal">RO</tt></td>
<td>Shorthand for <tt class="xref docutils literal">READONLY</tt>.</td>
</tr>
<tr><td><tt class="xref docutils literal">READ_RESTRICTED</tt></td>
<td>Not readable in restricted mode.</td>
</tr>
<tr><td><tt class="xref docutils literal">WRITE_RESTRICTED</tt></td>
<td>Not writable in restricted mode.</td>
</tr>
<tr><td><tt class="xref docutils literal">RESTRICTED</tt></td>
<td>Not readable or writable in restricted mode.</td>
</tr>
</tbody>
</table>
An interesting advantage of using the <a title="tp_members" class="reference external" href="../c-api/typeobj.html#tp_members"><tt class="xref docutils literal">tp_members</tt></a> table to build
descriptors that are used at runtime is that any attribute defined this way can
have an associated doc string simply by providing the text in the table. An
application can use the introspection API to retrieve the descriptor from the
class object, and get the doc string using its <tt class="xref docutils literal">__doc__</tt> attribute.
As with the <a title="tp_methods" class="reference external" href="../c-api/typeobj.html#tp_methods"><tt class="xref docutils literal">tp_methods</tt></a> table, a sentinel entry with a <tt class="xref docutils literal">name</tt> value
of NULL is required.
</div>
<div class="section" id="type-specific-attribute-management">
<h4>2.2.3.2. Type-specific Attribute Management<a class="headerlink" href="#type-specific-attribute-management" title="Permalink to this headline">¶</a></h4>
For simplicity, only the <tt class="xref docutils literal">char*</tt> version will be demonstrated here; the
type of the name parameter is the only difference between the <tt class="xref docutils literal">char*</tt>
and <a title="PyObject" class="reference external" href="../c-api/structures.html#PyObject"><tt class="xref docutils literal">PyObject*</tt></a> flavors of the interface. This example effectively does
the same thing as the generic example above, but does not use the generic
support added in Python 2.2. The value in showing this is two-fold: it
demonstrates how basic attribute management can be done in a way that is
portable to older versions of Python, and explains how the handler functions are
called, so that if you do need to extend their functionality, you&#8217;ll understand
what needs to be done.
The <a title="tp_getattr" class="reference external" href="../c-api/typeobj.html#tp_getattr"><tt class="xref docutils literal">tp_getattr</tt></a> handler is called when the object requires an attribute
look-up. It is called in the same situations where the <a title="object.__getattr__" class="reference external" href="../reference/datamodel.html#object.__getattr__"><tt class="xref docutils literal">__getattr__()</tt></a>
method of a class would be called.
A likely way to handle this is (1) to implement a set of functions (such as
<tt class="xref docutils literal">newdatatype_getSize()</tt> and <tt class="xref docutils literal">newdatatype_setSize()</tt> in the example
below), (2) provide a method table listing these functions, and (3) provide a
getattr function that returns the result of a lookup in that table. The method
table uses the same structure as the <a title="tp_methods" class="reference external" href="../c-api/typeobj.html#tp_methods"><tt class="xref docutils literal">tp_methods</tt></a> field of the type
object.
Here is an example:
<div class="highlight-c"><div class="highlight"><pre>static PyMethodDef newdatatype_methods[] = {
 {&quot;getSize&quot;, (PyCFunction)newdatatype_getSize, METH_VARARGS,
 &quot;Return the current size.&quot;},
 {&quot;setSize&quot;, (PyCFunction)newdatatype_setSize, METH_VARARGS,
 &quot;Set the size.&quot;},
 {NULL, NULL, 0, NULL} /* sentinel */
};

static PyObject *
newdatatype_getattr(newdatatypeobject *obj, char *name)
{
 return Py_FindMethod(newdatatype_methods, (PyObject *)obj, name);
}
</pre></div>
</div>
The <a title="tp_setattr" class="reference external" href="../c-api/typeobj.html#tp_setattr"><tt class="xref docutils literal">tp_setattr</tt></a> handler is called when the <a title="object.__setattr__" class="reference external" href="../reference/datamodel.html#object.__setattr__"><tt class="xref docutils literal">__setattr__()</tt></a> or
<a title="object.__delattr__" class="reference external" href="../reference/datamodel.html#object.__delattr__"><tt class="xref docutils literal">__delattr__()</tt></a> method of a class instance would be called. When an
attribute should be deleted, the third parameter will be NULL. Here is an
example that simply raises an exception; if this were really all you wanted, the
<a title="tp_setattr" class="reference external" href="../c-api/typeobj.html#tp_setattr"><tt class="xref docutils literal">tp_setattr</tt></a> handler should be set to NULL.
<div class="highlight-c"><div class="highlight"><pre>static int
newdatatype_setattr(newdatatypeobject *obj, char *name, PyObject *v)
{
 (void)PyErr_Format(PyExc_RuntimeError, &quot;Read-only attribute: \%s&quot;, name);
 return -1;
}
</pre></div>
</div>
</div>
</div>
<div class="section" id="object-comparison">
<h3>2.2.4. Object Comparison<a class="headerlink" href="#object-comparison" title="Permalink to this headline">¶</a></h3>
<div class="highlight-c"><div class="highlight"><pre>cmpfunc tp_compare;
</pre></div>
</div>
The <a title="tp_compare" class="reference external" href="../c-api/typeobj.html#tp_compare"><tt class="xref docutils literal">tp_compare</tt></a> handler is called when comparisons are needed and the
object does not implement the specific rich comparison method which matches the
requested comparison. (It is always used if defined and the
<a title="PyObject_Compare" class="reference external" href="../c-api/object.html#PyObject_Compare"><tt class="xref docutils literal">PyObject_Compare()</tt></a> or <a title="PyObject_Cmp" class="reference external" href="../c-api/object.html#PyObject_Cmp"><tt class="xref docutils literal">PyObject_Cmp()</tt></a> functions are used, or if
<a title="cmp" class="reference external" href="../library/functions.html#cmp"><tt class="xref docutils literal">cmp()</tt></a> is used from Python.) It is analogous to the <a title="object.__cmp__" class="reference external" href="../reference/datamodel.html#object.__cmp__"><tt class="xref docutils literal">__cmp__()</tt></a> method.
This function should return <tt class="docutils literal">-1</tt> if obj1 is less than obj2, <tt class="docutils literal">0</tt> if they
are equal, and <tt class="docutils literal">1</tt> if obj1 is greater than obj2. (It was previously
allowed to return arbitrary negative or positive integers for less than and
greater than, respectively; as of Python 2.2, this is no longer allowed. In the
future, other return values may be assigned a different meaning.)
A <a title="tp_compare" class="reference external" href="../c-api/typeobj.html#tp_compare"><tt class="xref docutils literal">tp_compare</tt></a> handler may raise an exception. In this case it should
return a negative value. The caller has to test for the exception using
<a title="PyErr_Occurred" class="reference external" href="../c-api/exceptions.html#PyErr_Occurred"><tt class="xref docutils literal">PyErr_Occurred()</tt></a>.
Here is a sample implementation:
<div class="highlight-c"><div class="highlight"><pre>static int
newdatatype_compare(newdatatypeobject * obj1, newdatatypeobject * obj2)
{
 long result;

if (obj1-&gt;obj_UnderlyingDatatypePtr-&gt;size &lt;
 obj2-&gt;obj_UnderlyingDatatypePtr-&gt;size) {
 result = -1;
 }
 else if (obj1-&gt;obj_UnderlyingDatatypePtr-&gt;size &gt;
 obj2-&gt;obj_UnderlyingDatatypePtr-&gt;size) {
 result = 1;
 }
 else {
 result = 0;
 }
 return result;
}
</pre></div>
</div>
</div>
<div class="section" id="abstract-protocol-support">
<h3>2.2.5. Abstract Protocol Support<a class="headerlink" href="#abstract-protocol-support" title="Permalink to this headline">¶</a></h3>
Python supports a variety of abstract &#8216;protocols;&#8217; the specific interfaces
provided to use these interfaces are documented in <a class="reference external" href="../c-api/abstract.html#abstract">Abstract Objects Layer</a>.
A number of these abstract interfaces were defined early in the development of
the Python implementation. In particular, the number, mapping, and sequence
protocols have been part of Python since the beginning. Other protocols have
been added over time. For protocols which depend on several handler routines
from the type implementation, the older protocols have been defined as optional
blocks of handlers referenced by the type object. For newer protocols there are
additional slots in the main type object, with a flag bit being set to indicate
that the slots are present and should be checked by the interpreter. (The flag
bit does not indicate that the slot values are non-NULL. The flag may be set
to indicate the presence of a slot, but a slot may still be unfilled.)
<div class="highlight-c"><div class="highlight"><pre>PyNumberMethods tp_as_number;
PySequenceMethods tp_as_sequence;
PyMappingMethods tp_as_mapping;
</pre></div>
</div>
If you wish your object to be able to act like a number, a sequence, or a
mapping object, then you place the address of a structure that implements the C
type <a title="PyNumberMethods" class="reference external" href="../c-api/typeobj.html#PyNumberMethods"><tt class="xref docutils literal">PyNumberMethods</tt></a>, <a title="PySequenceMethods" class="reference external" href="../c-api/typeobj.html#PySequenceMethods"><tt class="xref docutils literal">PySequenceMethods</tt></a>, or
<a title="PyMappingMethods" class="reference external" href="../c-api/typeobj.html#PyMappingMethods"><tt class="xref docutils literal">PyMappingMethods</tt></a>, respectively. It is up to you to fill in this
structure with appropriate values. You can find examples of the use of each of
these in the <tt class="docutils literal">Objects</tt> directory of the Python source distribution.
<div class="highlight-c"><div class="highlight"><pre>hashfunc tp_hash;
</pre></div>
</div>
This function, if you choose to provide it, should return a hash number for an
instance of your data type. Here is a moderately pointless example:
<div class="highlight-c"><div class="highlight"><pre>static long
newdatatype_hash(newdatatypeobject *obj)
{
 long result;
 result = obj-&gt;obj_UnderlyingDatatypePtr-&gt;size;
 result = result * 3;
 return result;
}
</pre></div>
</div>
<div class="highlight-c"><div class="highlight"><pre>ternaryfunc tp_call;
</pre></div>
</div>
This function is called when an instance of your data type is &#8220;called&#8221;, for
example, if <tt class="docutils literal">obj1</tt> is an instance of your data type and the Python script
contains <tt class="docutils literal">obj1('hello')</tt>, the <a title="tp_call" class="reference external" href="../c-api/typeobj.html#tp_call"><tt class="xref docutils literal">tp_call</tt></a> handler is invoked.
This function takes three arguments:
<ol class="arabic simple">
<li>arg1 is the instance of the data type which is the subject of the call. If
the call is <tt class="docutils literal">obj1('hello')</tt>, then arg1 is <tt class="docutils literal">obj1</tt>.</li>
<li>arg2 is a tuple containing the arguments to the call. You can use
<a title="PyArg_ParseTuple" class="reference external" href="../c-api/arg.html#PyArg_ParseTuple"><tt class="xref docutils literal">PyArg_ParseTuple()</tt></a> to extract the arguments.</li>
<li>arg3 is a dictionary of keyword arguments that were passed. If this is
non-NULL and you support keyword arguments, use
<a title="PyArg_ParseTupleAndKeywords" class="reference external" href="../c-api/arg.html#PyArg_ParseTupleAndKeywords"><tt class="xref docutils literal">PyArg_ParseTupleAndKeywords()</tt></a> to extract the arguments. If you do not
want to support keyword arguments and this is non-NULL, raise a
<a title="exceptions.TypeError" class="reference external" href="../library/exceptions.html#exceptions.TypeError"><tt class="xref docutils literal">TypeError</tt></a> with a message saying that keyword arguments are not supported.</li>
</ol>
Here is a desultory example of the implementation of the call function.
<div class="highlight-c"><div class="highlight"><pre>/* Implement the call function.
 * obj1 is the instance receiving the call.
 * obj2 is a tuple containing the arguments to the call, in this
 * case 3 strings.
 */
static PyObject *
newdatatype_call(newdatatypeobject *obj, PyObject *args, PyObject *other)
{
 PyObject *result;
 char *arg1;
 char *arg2;
 char *arg3;

if (!PyArg_ParseTuple(args, &quot;sss:call&quot;, &amp;arg1, &amp;arg2, &amp;arg3)) {
 return NULL;
 }
 result = PyString_FromFormat(
 &quot;Returning -- value: [\%d] arg1: [\%s] arg2: [\%s] arg3: [\%s]\n&quot;,
 obj-&gt;obj_UnderlyingDatatypePtr-&gt;size,
 arg1, arg2, arg3);
 printf(&quot;\%s&quot;, PyString_AS_STRING(result));
 return result;
}
</pre></div>
</div>
XXX some fields need to be added here...
<div class="highlight-c"><div class="highlight"><pre>/* Added in release 2.2 */
/* Iterators */
getiterfunc tp_iter;
iternextfunc tp_iternext;
</pre></div>
</div>
These functions provide support for the iterator protocol. Any object which
wishes to support iteration over its contents (which may be generated during
iteration) must implement the <tt class="docutils literal">tp_iter</tt> handler. Objects which are returned
by a <tt class="docutils literal">tp_iter</tt> handler must implement both the <tt class="docutils literal">tp_iter</tt> and <tt class="docutils literal">tp_iternext</tt>
handlers. Both handlers take exactly one parameter, the instance for which they
are being called, and return a new reference. In the case of an error, they
should set an exception and return NULL.
For an object which represents an iterable collection, the <tt class="docutils literal">tp_iter</tt> handler
must return an iterator object. The iterator object is responsible for
maintaining the state of the iteration. For collections which can support
multiple iterators which do not interfere with each other (as lists and tuples
do), a new iterator should be created and returned. Objects which can only be
iterated over once (usually due to side effects of iteration) should implement
this handler by returning a new reference to themselves, and should also
implement the <tt class="docutils literal">tp_iternext</tt> handler. File objects are an example of such an
iterator.
Iterator objects should implement both handlers. The <tt class="docutils literal">tp_iter</tt> handler should
return a new reference to the iterator (this is the same as the <tt class="docutils literal">tp_iter</tt>
handler for objects which can only be iterated over destructively). The
<tt class="docutils literal">tp_iternext</tt> handler should return a new reference to the next object in the
iteration if there is one. If the iteration has reached the end, it may return
NULL without setting an exception or it may set <a title="exceptions.StopIteration" class="reference external" href="../library/exceptions.html#exceptions.StopIteration"><tt class="xref docutils literal">StopIteration</tt></a>; avoiding
the exception can yield slightly better performance. If an actual error occurs,
it should set an exception and return NULL.
</div>
<div class="section" id="weak-reference-support">
<h3>2.2.6. Weak Reference Support<a class="headerlink" href="#weak-reference-support" title="Permalink to this headline">¶</a></h3>
One of the goals of Python&#8217;s weak-reference implementation is to allow any type
to participate in the weak reference mechanism without incurring the overhead on
those objects which do not benefit by weak referencing (such as numbers).
For an object to be weakly referencable, the extension must include a
<a title="PyObject" class="reference external" href="../c-api/structures.html#PyObject"><tt class="xref docutils literal">PyObject*</tt></a> field in the instance structure for the use of the weak
reference mechanism; it must be initialized to NULL by the object&#8217;s
constructor. It must also set the <a title="tp_weaklistoffset" class="reference external" href="../c-api/typeobj.html#tp_weaklistoffset"><tt class="xref docutils literal">tp_weaklistoffset</tt></a> field of the
corresponding type object to the offset of the field. For example, the instance
type is defined with the following structure:
<div class="highlight-c"><div class="highlight"><pre>typedef struct {
 PyObject_HEAD
 PyClassObject *in_class; /* The class object */
 PyObject *in_dict; /* A dictionary */
 PyObject *in_weakreflist; /* List of weak references */
} PyInstanceObject;
</pre></div>
</div>
The statically-declared type object for instances is defined this way:
<div class="highlight-c"><div class="highlight"><pre>PyTypeObject PyInstance_Type = {
 PyObject_HEAD_INIT(&amp;PyType_Type)
 0,
 &quot;module.instance&quot;,

/* Lots of stuff omitted for brevity... */

Py_TPFLAGS_DEFAULT, /* tp_flags */
 0, /* tp_doc */
 0, /* tp_traverse */
 0, /* tp_clear */
 0, /* tp_richcompare */
 offsetof(PyInstanceObject, in_weakreflist), /* tp_weaklistoffset */
};
</pre></div>
</div>
The type constructor is responsible for initializing the weak reference list to
NULL:
<div class="highlight-c"><div class="highlight"><pre>static PyObject *
instance_new() {
 /* Other initialization stuff omitted for brevity */

self-&gt;in_weakreflist = NULL;

if (inst-&gt;in_weakreflist != NULL)
 PyObject_ClearWeakRefs((PyObject *) inst);

/* Proceed with object destruction normally. */
}
</pre></div>
</div>
</div>
<div class="section" id="more-suggestions">
<h3>2.2.7. More Suggestions<a class="headerlink" href="#more-suggestions" title="Permalink to this headline">¶</a></h3>
Remember that you can omit most of these functions, in which case you provide
<tt class="docutils literal">0</tt> as a value. There are type definitions for each of the functions you must
provide. They are in <tt class="docutils literal">object.h</tt> in the Python include directory that
comes with the source distribution of Python.
In order to learn how to implement any specific method for your new data type,
do the following: Download and unpack the Python source distribution. Go the
<tt class="docutils literal">Objects</tt> directory, then search the C source files for <tt class="docutils literal">tp_</tt> plus the
function you want (for example, <tt class="docutils literal">tp_print</tt> or <tt class="docutils literal">tp_compare</tt>). You will find
examples of the function you want to implement.
When you need to verify that an object is an instance of the type you are
implementing, use the <a title="PyObject_TypeCheck" class="reference external" href="../c-api/object.html#PyObject_TypeCheck"><tt class="xref docutils literal">PyObject_TypeCheck()</tt></a> function. A sample of its use
might be something like the following:
<div class="highlight-c"><div class="highlight"><pre>if (! PyObject_TypeCheck(some_object, &amp;MyType)) {
 PyErr_SetString(PyExc_TypeError, &quot;arg #1 not a mything&quot;);
 return NULL;
}
</pre></div>
</div>
Footnotes
<table class="docutils footnote" frame="void" id="id7" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id2">[1]</a></td><td>This is true when we know that the object is a basic type, like a string or a
float.</td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id8" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id3">[2]</a></td><td>We relied on this in the <a title="tp_dealloc" class="reference external" href="../c-api/typeobj.html#tp_dealloc"><tt class="xref docutils literal">tp_dealloc</tt></a> handler in this example, because our
type doesn&#8217;t support garbage collection. Even if a type supports garbage
collection, there are calls that can be made to &#8220;untrack&#8221; the object from
garbage collection, however, these calls are advanced and not covered here.</td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id9" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id4">[3]</a></td><td>We now know that the first and last members are strings, so perhaps we could be
less careful about decrementing their reference counts, however, we accept
instances of string subclasses. Even though deallocating normal strings won&#8217;t
call back into our objects, we can&#8217;t guarantee that deallocating an instance of
a string subclass won&#8217;t call back into our objects.</td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id10" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id5">[4]</a></td><td>Even in the third version, we aren&#8217;t guaranteed to avoid cycles. Instances of
string subclasses are allowed and string subclasses could allow cycles even if
normal strings don&#8217;t.</td></tr>
</tbody>
</table>
</div>
</div>
</div>

</div>
 </div>
 </div>
 <div class="sphinxsidebar">
 <div class="sphinxsidebarwrapper">
 <h3><a href="../contents.html">Table Of Contents</a></h3>
 <ul>
<li><a class="reference external" href="#">2. Defining New Types</a><ul>
<li><a class="reference external" href="#the-basics">2.1. The Basics</a><ul>
<li><a class="reference external" href="#adding-data-and-methods-to-the-basic-example">2.1.1. Adding data and methods to the Basic example</a></li>
<li><a class="reference external" href="#providing-finer-control-over-data-attributes">2.1.2. Providing finer control over data attributes</a></li>
<li><a class="reference external" href="#supporting-cyclic-garbage-collection">2.1.3. Supporting cyclic garbage collection</a></li>
<li><a class="reference external" href="#subclassing-other-types">2.1.4. Subclassing other types</a></li>
</ul>
</li>
<li><a class="reference external" href="#type-methods">2.2. Type Methods</a><ul>
<li><a class="reference external" href="#finalization-and-de-allocation">2.2.1. Finalization and De-allocation</a></li>
<li><a class="reference external" href="#object-presentation">2.2.2. Object Presentation</a></li>
<li><a class="reference external" href="#attribute-management">2.2.3. Attribute Management</a><ul>
<li><a class="reference external" href="#generic-attribute-management">2.2.3.1. Generic Attribute Management</a></li>
<li><a class="reference external" href="#type-specific-attribute-management">2.2.3.2. Type-specific Attribute Management</a></li>
</ul>
</li>
<li><a class="reference external" href="#object-comparison">2.2.4. Object Comparison</a></li>
<li><a class="reference external" href="#abstract-protocol-support">2.2.5. Abstract Protocol Support</a></li>
<li><a class="reference external" href="#weak-reference-support">2.2.6. Weak Reference Support</a></li>
<li><a class="reference external" href="#more-suggestions">2.2.7. More Suggestions</a></li>
</ul>
</li>
</ul>
</li>
</ul>

<h4>Previous topic</h4>
 <a href="extending.html"
 title="previous chapter">1. Extending Python with C or C++</a>
 <h4>Next topic</h4>
 <a href="building.html"
 title="next chapter">3. Building C and C++ Extensions with distutils</a>
<h3>This Page</h3>
<ul class="this-page-menu">
 <li><a href="../bugs.html">Report a Bug</a></li>
 <li><a href="../_sources/extending/newtypes.txt"
 rel="nofollow">Show Source</a></li>
</ul>

<div id="searchbox" style="display: none">
 <h3>Quick search</h3>
 <form class="search" action="../search.html" method="get">
 <input type="text" name="q" size="18" />
 <input type="submit" value="Go" />
 <input type="hidden" name="check_keywords" value="yes" />
 <input type="hidden" name="area" value="default" />
 </form>
 
 Enter search terms or a module, class or function name.
 
 </div>
 <script type="text/javascript">$('#searchbox').show(0);</script>
 </div>
 </div>
 <div class="clearer"></div>
 </div>
 <div class="related">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="../genindex.html" title="General Index"
 >index</a></li>
 <li class="right" >
 <a href="../modindex.html" title="Global Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="building.html" title="3. Building C and C++ Extensions with distutils"
 >next</a> |</li>
 <li class="right" >
 <a href="extending.html" title="1. Extending Python with C or C++"
 >previous</a> |</li>
 <li><img src="../_static/py.png" alt=""
 style="vertical-align: middle; margin-top: -1px"/></li>
 <li><a href="../index.html">Python v2.6.6 documentation</a> &raquo;</li>

<li><a href="index.html" >Extending and Embedding the Python Interpreter</a> &raquo;</li> 
 </ul>
 </div>
 <div class="footer">
 &copy; <a href="../copyright.html">Copyright</a> 1990-2011, Python Software Foundation.
 
 The Python Software Foundation is a non-profit corporation. 
 <a href="http://www.python.org/psf/donations/">Please donate.</a>
 
 Last updated on Jul 20, 2011.
 <a href="../bugs.html">Found a bug</a>?
 
 Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 0.6.6.
 </div>

</body>
</html>

Back to File Manager