Skip to content

C
cython
Commit 73f04523 authored by scoder's avatar scoder Committed by GitHub
Browse files
Options

Merge pull request #1666 from jdemeyer/remove_old_doc 

Remove obsolete documentation
parents 5a1c4958 85065883
Show whitespace changes
Inline Side-by-side
Showing with 0 additions and 2672 deletions
Doc/About.html deleted 100644 → 0
View file @ 5a1c4958
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.51 (Macintosh; I; PPC) [Netscape]">
<title>About Cython</title>
</head>
<body>
<center>
<h1>
<hr width="100%">Cython</h1></center>
<center><i><font size=+1>A language for writing Python extension modules</font></i>
<hr width="100%"></center>
<h2>
What is Cython all about?</h2>
Cython is a language specially designed for writing Python extension modules.
It's designed to bridge the gap between the nice, high-level, easy-to-use
world of Python and the messy, low-level world of C.
<p>You may be wondering why anyone would want a special language for this.
Python is really easy to extend using C or C++, isn't it? Why not just
write your extension modules in one of those languages?
<p>Well, if you've ever written an extension module for Python, you'll
know that things are not as easy as all that. First of all, there is a
fair bit of boilerplate code to write before you can even get off the ground.
Then you're faced with the problem of converting between Python and C data
types. For the basic types such as numbers and strings this is not too
bad, but anything more elaborate and you're into picking Python objects
apart using the Python/C API calls, which requires you to be meticulous
about maintaining reference counts, checking for errors at every step and
cleaning up properly if anything goes wrong. Any mistakes and you have
a nasty crash that's very difficult to debug.
<p>Various tools have been developed to ease some of the burdens of producing
extension code, of which perhaps <a href="http://www.swig.org">SWIG</a>
is the best known. SWIG takes a definition file consisting of a mixture
of C code and specialised declarations, and produces an extension module.
It writes all the boilerplate for you, and in many cases you can use it
without knowing about the Python/C API. But you need to use API calls if
any substantial restructuring of the data is required between Python and
C.
<p>What's more, SWIG gives you no help at all if you want to create a new
built-in Python <i>type. </i>It will generate pure-Python classes which
wrap (in a slightly unsafe manner) pointers to C data structures, but creation
of true extension types is outside its scope.
<p>Another notable attempt at making it easier to extend Python is <a href="http://pyinline.sourceforge.net/">PyInline</a>
, inspired by a similar facility for Perl. PyInline lets you embed pieces
of C code in the midst of a Python file, and automatically extracts them
and compiles them into an extension. But it only converts the basic types
automatically, and as with SWIG,&nbsp; it doesn't address the creation
of new Python types.
<p>Cython aims to go far beyond what any of these previous tools provides.
Cython deals with the basic types just as easily as SWIG, but it also lets
you write code to convert between arbitrary Python data structures and
arbitrary C data structures, in a simple and natural way, without knowing
<i>anything</i> about the Python/C API. That's right -- <i>nothing at all</i>!
Nor do you have to worry about reference counting or error checking --
it's all taken care of automatically, behind the scenes, just as it is
in interpreted Python code. And what's more, Cython lets you define new
<i>built-in</i> Python types just as easily as you can define new classes
in Python.
<p>Sound too good to be true? Read on and find out how it's done.
<h2>
The Basics of Cython</h2>
The fundamental nature of Cython can be summed up as follows: <b>Cython is
Python with C data types</b>.
<p><i>Cython is Python:</i> Almost any piece of Python code is also valid
Cython code. (There are a few limitations, but this approximation will serve
for now.) The Cython compiler will convert it into C code which makes equivalent
calls to the Python/C API. In this respect, Cython is similar to the former
Python2C project (to which I would supply a reference except that it no
longer seems to exist).
<p><i>...with C data types.</i> But Cython is much more than that, because
parameters and variables can be declared to have C data types. Code which
manipulates Python values and C values can be freely intermixed, with conversions
occurring automatically wherever possible. Reference count maintenance
and error checking of Python operations is also automatic, and the full
power of Python's exception handling facilities, including the try-except
and try-finally statements, is available to you -- even in the midst of
manipulating C data.
<p>Here's a small example showing some of what can be done. It's a routine
for finding prime numbers. You tell it how many primes you want, and it
returns them as a Python list.
<blockquote><b><tt><font size=+1>primes.pyx</font></tt></b></blockquote>
<blockquote>
<pre>&nbsp;1&nbsp; def primes(int kmax):
&nbsp;2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int n, k, i
&nbsp;3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int p[1000]
&nbsp;4&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; result = []
&nbsp;5&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if kmax > 1000:
&nbsp;6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; kmax = 1000
&nbsp;7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; k = 0
&nbsp;8&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; n = 2
&nbsp;9&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; while k &lt; kmax:
10&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; i = 0
11&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; while i &lt; k and n % p[i] &lt;> 0:
12&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; i = i + 1
13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if i == k:
14&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; p[k] = n
15&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; k = k + 1
16&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; result.append(n)
17&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; n = n + 1
18&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; return result</pre>
</blockquote>
You'll see that it starts out just like a normal Python function definition,
except that the parameter <b>kmax</b> is declared to be of type <b>int</b>
. This means that the object passed will be converted to a C integer (or
a TypeError will be raised if it can't be).
<p>Lines 2 and 3 use the <b>cdef</b> statement to define some local C variables.
Line 4 creates a Python list which will be used to return the result. You'll
notice that this is done exactly the same way it would be in Python. Because
the variable <b>result</b> hasn't been given a type, it is assumed to hold
a Python object.
<p>Lines 7-9 set up for a loop which will test candidate numbers for primeness
until the required number of primes has been found. Lines 11-12, which
try dividing a candidate by all the primes found so far, are of particular
interest. Because no Python objects are referred to, the loop is translated
entirely into C code, and thus runs very fast.
<p>When a prime is found, lines 14-15 add it to the p array for fast access
by the testing loop, and line 16 adds it to the result list. Again, you'll
notice that line 16 looks very much like a Python statement, and in fact
it is, with the twist that the C parameter <b>n</b> is automatically converted
to a Python object before being passed to the <b>append</b> method. Finally,
at line 18, a normal Python <b>return</b> statement returns the result
list.
<p>Compiling primes.pyx with the Cython compiler produces an extension module
which we can try out in the interactive interpreter as follows:
<blockquote>
<pre>>>> import primes
>>> primes.primes(10)
[2, 3, 5, 7, 11, 13, 17, 19, 23, 29]
>>></pre>
</blockquote>
See, it works! And if you're curious about how much work Cython has saved
you, take a look at the <a href="primes.c">C code generated for this module</a>
.
<h2>
Language Details</h2>
For more about the Cython language, see the <a href="overview.html">Language
Overview</a> .
<h2>
Future Plans</h2>
Cython is not finished. Substantial tasks remaining include:
<ul>
<li>
Support for certain Python language features which are planned but not
yet implemented. See the <a href="overview.html#Limitations">Limitations</a>
section of the <a href="overview.html">Language Overview</a> for a current
list.</li>
</ul>
<ul>
<li>
C++ support. This could be a very big can of worms - careful thought required
before going there.</li>
</ul>
<ul>
<li>
Reading C/C++ header files directly would be very nice, but there are some
severe problems that I will have to find solutions for first, such as what
to do about preprocessor macros. My current thinking is to use a separate
tool to convert .h files into Cython declarations, possibly with some manual
intervention.</li>
</ul>
</body>
</html>
Doc/FAQ.html deleted 100644 → 0
View file @ 5a1c4958
<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.51 (Macintosh; I; PPC) [Netscape]"><title>FAQ.html</title></head>
<body>
<center> <h1> <hr width="100%">Cython FAQ
<hr width="100%"></h1>
</center>
<h2> Contents</h2>
<ul>
<li> <b><a href="#CallCAPI">How do I call Python/C API routines?</a></b></li>
<li> <b><a href="#NullBytes">How do I convert a C string containing null
bytes to a Python string?</a></b></li>
<li> <b><a href="#NumericAccess">How do I access the data inside a Numeric
array object?</a></b></li>
<li><b><a href="#Rhubarb">Cython says my extension type object has no attribute
'rhubarb', but I know it does. What gives?</a></b></li><li><a style="font-weight: bold;" href="#Quack">Python says my extension type has no method called 'quack', but I know it does. What gives?</a><br>
</li>
</ul>
<hr width="100%"> <h2> <a name="CallCAPI"></a>How do I call Python/C API routines?</h2>
Declare them as C functions inside a <tt>cdef extern from</tt> block.
Use the type name <tt>object</tt> for any parameters and return types which
are Python object references. Don't use the word <tt>const</tt> anywhere.
Here is an example which defines and uses the <tt>PyString_FromStringAndSize</tt> routine:
<blockquote><tt>cdef extern from "Python.h":</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; object PyString_FromStringAndSize(char *, int)</tt> <p><tt>cdef char buf[42]</tt> <br>
<tt>my_string = PyString_FromStringAndSize(buf, 42)</tt></p>
</blockquote>
<h2> <a name="NullBytes"></a>How do I convert a C string containing null
bytes to a Python string?</h2>
Put in a declaration for the <tt>PyString_FromStringAndSize</tt> API routine
and use that<tt>.</tt> See <a href="#CallCAPI">How do I call Python/C API
routines?</a> <h2> <a name="NumericAccess"></a>How do I access the data inside a Numeric
array object?</h2>
Use a <tt>cdef extern from</tt> block to include the Numeric header file
and declare the array object as an external extension type. The following
code illustrates how to do this:
<blockquote><tt>cdef extern from "Numeric/arrayobject.h":</tt> <p><tt>&nbsp;&nbsp;&nbsp; struct PyArray_Descr:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int type_num, elsize</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; char type</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp; ctypedef class Numeric.ArrayType [object PyArrayObject]</tt><tt>:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef char *data</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int nd</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int *dimensions,
*strides</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef object base</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef PyArray_Descr *descr</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef int flags<br>
</tt></p>
</blockquote>
<p>For more information about external extension types, see the <a href="extension_types.html#ExternalExtTypes">"External Extension Types"</a>
section of the <a href="extension_types.html">"Extension Types"</a> documentation
page.<br>
<tt> </tt> </p>
<h2><a name="Rhubarb"></a>Cython says my extension type object has no attribute
'rhubarb', but I know it does. What gives?</h2>
You're probably trying to access it through a reference which Cython thinks
is a generic Python object. You need to tell Cython that it's a reference
to your extension type by means of a declaration,<br>
for example,<br>
<blockquote><tt>cdef class Vegetables:</tt><br>
<tt>&nbsp; &nbsp; cdef int rhubarb</tt><br>
<br>
<tt>...</tt><br>
<tt>cdef Vegetables veg</tt><br>
<tt>veg.rhubarb = 42</tt><br>
</blockquote>
Also see the <a href="extension_types.html#ExtTypeAttrs">"Attributes"</a>
section of the <a href="extension_types.html">"Extension
Types"</a> documentation page.<br>
<h2><a name="Quack"></a>Python says my extension type has no method called 'quack', but I know it does. What gives?</h2>
You may have declared the method using <span style="font-family: monospace;">cdef</span> instead of <span style="font-family: monospace;">def</span>. Only functions and methods declared with <span style="font-family: monospace;">def</span> are callable from Python code.<br>
---
</body></html>
\ No newline at end of file
Doc/extension_types.html deleted 100644 → 0
View file @ 5a1c4958
<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Extension Types</title></head>
<body>
<h1> <hr width="100%">Extension Types
<hr width="100%"></h1>
<h2> Contents</h2>
<ul>
<li> <a href="#Introduction">Introduction</a></li>
<li> <a href="#ExtTypeAttrs">Attributes</a></li>
<li> <a href="#NotNone">Extension types and None</a></li>
<li> <a href="special_methods.html">Special methods</a></li>
<li> <a href="#Properties">Properties</a> <font style="color: rgb(0, 153, 0);" color="#ed181e">(NEW in
0.9)</font></li>
<li><a href="#SubclassingExtTypes">Subclassing</a></li>
<li> <a href="#CMethods">C Methods</a> <font style="color: rgb(0, 153, 0);" color="#ff0000">(NEW in 0.9)</font><br>
<a href="#ForwardDeclaringExtTypes">Forward-declaring extension types</a></li><li><a href="#WeakRefs">Making extension types weak-referenceable</a> <span style="color: rgb(255, 0, 0);">(NEW in 0.9.4)</span><br>
</li>
<li> <a href="#PublicAndExtern">Public and external extension types</a><font color="#2f8b20"><br>
</font></li>
<ul>
<li> <a href="#ExternalExtTypes">External extension types</a></li>
<li> <a href="#ImplicitImport">Implicit importing</a><font color="#2f8b20"><br>
</font></li>
<li> <a href="#TypeVsConstructor">Type names vs. constructor names</a></li>
<li> <a href="#PublicExtensionTypes">Public extension types</a></li>
<li> <a href="#NameSpecClause">Name specification clause</a></li>
</ul>
</ul>
<h2> <a name="Introduction"></a>Introduction</h2>
As well as creating normal user-defined classes with the Python <b>class</b>
statement, Cython also lets you create new built-in Python types, known as
<i>extension types</i>. You define an extension type using the <b>cdef class</b> statement. Here's an example:
<blockquote><tt>cdef class Shrubbery:</tt> <p><tt>&nbsp;&nbsp;&nbsp; cdef int width, height</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp; def __init__(self, w, h):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.width = w</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.height = h</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp; def describe(self):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; print "This shrubbery is",
self.width, \</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
"by", self.height, "cubits."</tt></p>
</blockquote>
As you can see, a Cython extension type definition looks a lot like a Python
class definition. Within it, you use the <b>def</b> statement to define
methods that can be called from Python code. You can even define many of
the special methods such as <tt>__init__</tt> as you would in Python.
<p>The main difference is that you can use the <b>cdef</b> statement to define
attributes. The attributes may be Python objects (either generic or of a particular
extension type), or they may be of any C data type. So you can use extension
types to wrap arbitrary C data structures and provide a Python-like interface
to them. </p>
<h2> <a name="ExtTypeAttrs"></a>Attributes</h2>
Attributes of an extension type are stored directly in the object's C struct.
The set of attributes is fixed at compile time; you can't add attributes
to an extension type instance at run time simply by assigning to them, as
you could with a Python class instance. (You can subclass the extension type
in Python and add attributes to instances of the subclass, however.)
<p>There are two ways that attributes of an extension type can be accessed:
by Python attribute lookup, or by direct access to the C struct from Cython
code. Python code is only able to access attributes of an extension type
by the first method, but Cython code can use either method. </p>
<p>By default, extension type attributes are only accessible by direct access,
not Python access, which means that they are not accessible from Python code.
To make them accessible from Python code, you need to declare them as <tt>public</tt> or <tt>readonly</tt>. For example, </p>
<blockquote><tt>cdef class Shrubbery:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; cdef public int width, height</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; cdef readonly float depth</tt></blockquote>
makes the <tt>width</tt> and <tt>height</tt> attributes readable and writable
from Python code, and the <tt>depth</tt> attribute readable but not writable.
<p>Note that you can only expose simple C types, such as ints, floats and
strings, for Python access. You can also expose Python-valued attributes,
although read-write exposure is only possible for generic Python attributes
(of type <tt>object</tt>). If the attribute is declared to be of an extension
type, it must be exposed <tt>readonly</tt>. </p>
<p>Note also that the <tt>public</tt> and <tt>readonly</tt> options apply
only to <i>Python</i> access, not direct access. All the attributes of an
extension type are always readable and writable by direct access. </p>
<p>Howerver, for direct access to be possible, the Cython compiler must know
that you have an instance of that type, and not just a generic Python object.
It knows this already in the case of the "self" parameter of the methods of
that type, but in other cases you will have to tell it by means of a declaration.
For example, </p>
<blockquote><tt>cdef widen_shrubbery(Shrubbery sh, extra_width):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
If you attempt to access an extension type attribute through a generic
object reference, Cython will use a Python attribute lookup. If the attribute
is exposed for Python access (using <tt>public</tt> or <tt>readonly</tt>)
then this will work, but it will be much slower than direct access.
<h2> <a name="NotNone"></a>Extension types and None</h2>
When you declare a parameter or C variable as being of an extension type,
Cython will allow it to take on the value None as well as values of its declared
type. This is analogous to the way a C pointer can take on the value NULL,
and you need to exercise the same caution because of it. There is no problem
as long as you are performing Python operations on it, because full dynamic
type checking will be applied. However, when you access C attributes of an
extension type (as in the <tt>widen_shrubbery</tt> function above), it's up
to you to make sure the reference you're using is not None -- in the interests
of efficiency, Cython does <i>not</i> check this.
<p>You need to be particularly careful when exposing Python functions which
take extension types as arguments. If we wanted to make <tt>widen_shrubbery</tt>
a Python function, for example, if we simply wrote </p>
<blockquote><tt>def widen_shrubbery(Shrubbery sh, extra_width): # <font color="#ed181e">This is</font></tt> <br>
<tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# <font color="#ed181e">dangerous!</font></tt></blockquote>
then users of our module could crash it by passing None for the <tt>sh</tt>
parameter.
<p>One way to fix this would be </p>
<blockquote><tt>def widen_shrubbery(Shrubbery sh, extra_width):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; if sh is None:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; raise TypeError</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
but since this is anticipated to be such a frequent requirement, Cython
provides a more convenient way. Parameters of a Python function declared
as an extension type can have a <b><tt>not None</tt></b> clause:
<blockquote><tt>def widen_shrubbery(Shrubbery sh not None, extra_width):</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp; sh.width = sh.width + extra_width</tt></blockquote>
Now the function will automatically check that <tt>sh</tt> is not None
along with checking that it has the right type.
<p>Note, however that the <tt>not None</tt> clause can <i>only</i> be used
in Python functions (defined with <tt>def</tt>) and not C functions (defined
with <tt>cdef</tt>). If you need to check whether a parameter to a C function
is None, you will need to do it yourself. </p>
<p>Some more things to note: </p>
<ul>
<li> The <b>self</b> parameter of a method of an extension type is guaranteed
never to be None.</li>
</ul>
<ul>
<li> When comparing a value with None, keep in mind that, if <tt>x</tt> is a Python object, <tt>x is None</tt> and <tt>x is not None</tt> are very
efficient because they translate directly to C pointer comparisons, whereas
<tt>x == None</tt> and <tt>x != None</tt>, or simply using <tt>x</tt> as a boolean value (as in <tt>if x: ...</tt>) will invoke Python operations
and therefore be much slower.</li>
</ul>
<h2> <a name="ExtTypeSpecialMethods"></a>Special methods</h2>
Although the principles are similar, there are substantial differences
between many of the <span style="font-family: monospace;">__xxx__</span> special methods of extension types and their
Python counterparts. There is a <a href="special_methods.html">separate page</a> devoted to this subject, and you should read it carefully before attempting
to use any special methods in your extension types.
<h2> <a name="Properties"></a>Properties</h2>
There is a special syntax for defining <b>properties</b> in an extension
class:
<blockquote><tt>cdef class Spam:</tt> <p><tt>&nbsp;&nbsp;&nbsp; property cheese:</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; "A doc string can go
here."</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __get__(self):</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# This is called when the property is read.</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
...</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __set__(self, value):</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# This is called when the property is written.</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
...</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; def __del__(self):</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;
# This is called when the property is deleted.</tt> <br>
&nbsp;</p>
</blockquote>
The <tt>__get__</tt>, <tt>__set__</tt> and <tt>__del__</tt> methods are
all optional; if they are omitted, an exception will be raised when the corresponding
operation is attempted.
<p>Here's a complete example. It defines a property which adds to a list
each time it is written to, returns the list when it is read, and empties
the list when it is deleted. <br>
&nbsp; </p>
<center> <table align="center" cellpadding="5">
<tbody>
<tr>
<td bgcolor="#ffaf18"><b><tt>cheesy.pyx</tt></b></td>
<td bgcolor="#5dbaca"><b>Test input</b></td>
</tr>
<tr>
<td rowspan="3" bgcolor="#ffaf18" valign="top"><tt>cdef class CheeseShop:</tt>
<p><tt>&nbsp; cdef object cheeses</tt> </p>
<p><tt>&nbsp; def __new__(self):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; self.cheeses = []</tt> </p>
<p><tt>&nbsp; property cheese:</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp; def __get__(self):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; return "We don't have: %s" % self.cheeses</tt>
</p>
<p><tt>&nbsp;&nbsp;&nbsp; def __set__(self, value):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.cheeses.append(value)</tt>
</p>
<p><tt>&nbsp;&nbsp;&nbsp; def __del__(self):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; del self.cheeses[:]</tt></p>
</td>
<td bgcolor="#5dbaca" valign="top"><tt>from cheesy import CheeseShop</tt>
<p><tt>shop = CheeseShop()</tt> <br>
<tt>print shop.cheese</tt> </p>
<p><tt>shop.cheese = "camembert"</tt> <br>
<tt>print shop.cheese</tt> </p>
<p><tt>shop.cheese = "cheddar"</tt> <br>
<tt>print shop.cheese</tt> </p>
<p><tt>del shop.cheese</tt> <br>
<tt>print shop.cheese</tt></p>
</td>
</tr>
<tr>
<td bgcolor="#8cbc1c"><b>Test output</b></td>
</tr>
<tr>
<td bgcolor="#8cbc1c"><tt>We don't have: []</tt> <br>
<tt>We don't have: ['camembert']</tt> <br>
<tt>We don't have: ['camembert', 'cheddar']</tt> <br>
<tt>We don't have: []</tt></td>
</tr>
</tbody> </table>
</center>
<h2> <a name="SubclassingExtTypes"></a>Subclassing</h2>
An extension type may inherit from a built-in type or another extension
type:
<blockquote><tt>cdef class Parrot:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt><tt></tt> <p><tt>cdef class Norwegian(Parrot):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></p>
</blockquote>
<p><br>
A complete definition of the base type must be available to Cython, so if
the base type is a built-in type, it must have been previously declared as
an <b>extern</b> extension type. If the base type is defined in another Cython
module, it must either be declared as an extern extension type or imported
using the <b><a href="sharing.html">cimport</a></b> statement. </p>
<p>An extension type can only have one base class (no multiple inheritance).
</p>
<p>Cython extension types can also be subclassed in Python. A Python class
can inherit from multiple extension types provided that the usual Python
rules for multiple inheritance are followed (i.e. the C layouts of all the
base classes must be compatible).<br>
</p>
<h2><a name="CMethods"></a>C methods</h2>
Extension types can have C methods as well as Python methods. Like C functions,
C methods are declared using <tt>cdef</tt> instead of <tt>def</tt>. C methods
are "virtual", and may be overridden in derived extension types.<br>
<br>
<table align="center" cellpadding="5">
<tbody>
<tr>
<td bgcolor="#ffaf18" valign="top" width="50%"><b><tt>pets.pyx</tt></b><br>
</td>
<td bgcolor="#8cbc1c" valign="top" width="30%"><b>Output</b><br>
</td>
</tr>
<tr>
<td bgcolor="#ffaf18" valign="top" width="50%"><tt>cdef class Parrot:<br>
<br>
&nbsp; cdef void describe(self):<br>
&nbsp; &nbsp; print "This parrot is resting."<br>
<br>
cdef class Norwegian(Parrot):<br>
<br>
&nbsp; cdef void describe(self):<br>
&nbsp; &nbsp; Parrot.describe(self)<br>
&nbsp; &nbsp; print "Lovely plumage!"<br>
<br>
<br>
cdef Parrot p1, p2<br>
p1 = Parrot()<br>
p2 = Norwegian()<br>
print "p1:"<br>
p1.describe()<br>
print "p2:"<br>
p2.describe()</tt> <br>
</td>
<td bgcolor="#8cbc1c" valign="top" width="30%"><tt>p1:<br>
This parrot is resting.<br>
p2:<br>
</tt><tt>This parrot is resting.<br>
</tt><tt> Lovely plumage!</tt><br>
</td>
</tr>
</tbody> </table>
<br>
The above example also illustrates that a C method can call an inherited
C method using the usual Python technique, i.e.<br>
<blockquote><tt>Parrot.describe(self)</tt><br>
</blockquote>
<h2><a name="ForwardDeclaringExtTypes"></a>Forward-declaring extension types</h2>
Extension types can be forward-declared, like struct and union types. This
will be necessary if you have two extension types that need to refer to
each other, e.g.
<blockquote><tt>cdef class Shrubbery # forward declaration</tt> <p><tt>cdef class Shrubber:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; cdef Shrubbery work_in_progress</tt> </p>
<p><tt>cdef class Shrubbery:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; cdef Shrubber creator</tt></p>
</blockquote>
If you are forward-declaring an exension type that has a base class, you
must specify the base class in both the forward declaration and its subsequent
definition, for example,<br>
<blockquote><tt>cdef class A(B)<br>
<br>
...<br>
<br>
cdef class A(B):<br>
&nbsp; &nbsp; # attributes and methods</tt><br>
</blockquote>
<h2><a name="WeakRefs"></a>Making extension types weak-referenceable</h2>By
default, extension types do not support having weak references made to
them. You can enable weak referencing by declaring a C attribute of
type <span style="font-family: monospace;">object</span> called <span style="font-family: monospace; font-weight: bold;">__weakref__</span>. For example,<br>
<br>
<div style="margin-left: 40px;"><span style="font-family: monospace;">cdef class ExplodingAnimal:</span><br style="font-family: monospace;">
<span style="font-family: monospace;">&nbsp;&nbsp;&nbsp; """This animal will self-destruct when it is</span><br>
<span style="font-family: monospace;">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; no longer strongly referenced."""</span><br>
<span style="font-family: monospace;">&nbsp;&nbsp;&nbsp; </span><br style="font-family: monospace;">
<span style="font-family: monospace;"></span><span style="font-family: monospace;">&nbsp;&nbsp;&nbsp; cdef object __weakref__</span><br>
</div>
<br>
<h2><a name="PublicAndExtern"></a>Public and external extension types</h2>
Extension types can be declared <b>extern</b> or <b>public</b>. An <a href="#ExternalExtTypes"><b>extern</b> extension type declaration</a> makes
an extension type defined in external C code available to a Cython module.
A <a href="#PublicExtensionTypes"><b>public</b> extension type declaration</a> makes an extension type defined in a Cython module available to external C
code.
<h3> <a name="ExternalExtTypes"></a>External extension types</h3>
An <b>extern</b> extension type allows you to gain access to the internals
of Python objects defined in the Python core or in a non-Cython extension
module.
<blockquote><b>NOTE:</b> In Cython versions before 0.8, <b>extern</b> extension
types were also used to reference extension types defined in another Cython
module. While you can still do that, Cython 0.8 and later provides a better
mechanism for this. See <a href="sharing.html">Sharing C Declarations Between
Cython Modules</a>.</blockquote>
Here is an example which will let you get at the C-level members of the
built-in <i>complex</i> object.
<blockquote><tt>cdef extern from "complexobject.h":</tt> <p><tt>&nbsp;&nbsp;&nbsp; struct Py_complex:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; double real</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; double imag</tt> </p>
<p><tt>&nbsp;&nbsp;&nbsp; ctypedef class __builtin__.complex [object PyComplexObject]:</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; cdef Py_complex cval</tt>
</p>
<p><tt># A function which uses the above type</tt> <br>
<tt>def spam(complex c):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; print "Real:", c.cval.real</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; print "Imag:", c.cval.imag</tt></p>
</blockquote>
Some important things to note are:
<ol>
<li> In this example, <b>ctypedef class</b> has been used. This is because,
in the Python header files, the <tt>PyComplexObject</tt> struct is declared
with<br>
<br>
<div style="margin-left: 40px;"><tt>ctypedef struct {</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt> <br>
<tt>} PyComplexObject;<br>
<br>
</tt></div>
</li><li>As well as the name of the extension type, the <i>module</i> in which
its type object can be found is also specified. See the <a href="#ImplicitImport">implicit importing</a> section below.&nbsp; <br>
<br>
</li>
<li> When declaring an external extension type, you don't declare
any methods. Declaration of methods is not required in order to call them,
because the calls are Python method calls. Also, as with structs and unions,
if your extension class declaration is inside a <i>cdef extern from</i> block,
you only need to declare those C members which you wish to access.</li>
</ol>
<h3> <a name="ImplicitImport"></a>Implicit importing</h3>
<blockquote><font color="#ef1f1d">Backwards Incompatibility Note</font>:
You will have to update any pre-0.8 Cython modules you have which use <b>extern</b>
extension types. I apologise for this, but for complicated reasons it proved
to be too difficult to continue supporting the old way of doing these while
introducing the new features that I wanted.</blockquote>
Cython 0.8 and later requires you to include a module name in an extern
extension class declaration, for example,
<blockquote><tt>cdef extern class MyModule.Spam:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
The type object will be implicitly imported from the specified module and
bound to the corresponding name in this module. In other words, in this
example an implicit
<ol>
<pre>from <tt>MyModule</tt> import Spam</pre>
</ol>
statement will be executed at module load time.
<p>The module name can be a dotted name to refer to a module inside a package
hierarchy, for example, </p>
<blockquote><tt>cdef extern class My.Nested.Package.Spam:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
You can also specify an alternative name under which to import the type
using an <b>as</b> clause, for example,
<ol>
<tt>cdef extern class My.Nested.Package.Spam as Yummy:</tt> <br>
<tt>&nbsp;&nbsp; ...</tt> </ol>
which corresponds to the implicit import statement
<ol>
<pre>from <tt>My.Nested.Package</tt> import <tt>Spam</tt> as <tt>Yummy</tt></pre>
</ol>
<h3> <a name="TypeVsConstructor"></a>Type names vs. constructor names</h3>
Inside a Cython module, the name of an extension type serves two distinct
purposes. When used in an expression, it refers to a module-level global
variable holding the type's constructor (i.e. its type-object). However,
it can also be used as a C type name to declare variables, arguments and
return values of that type.
<p>When you declare </p>
<blockquote><tt>cdef extern class MyModule.Spam:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
the name <tt>Spam</tt> serves both these roles. There may be other names
by which you can refer to the constructor, but only <tt>Spam</tt> can be
used as a type name. For example, if you were to explicity <tt>import MyModule</tt>,
you could use<tt> MyModule.Spam()</tt> to create a Spam instance, but you
wouldn't be able to use <tt>MyModule.Spam</tt> as a type name.
<p>When an <b>as</b> clause is used, the name specified in the <b>as</b>
clause also takes over both roles. So if you declare </p>
<blockquote><tt>cdef extern class MyModule.Spam as Yummy:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
then <tt>Yummy</tt> becomes both the type name and a name for the constructor.
Again, there are other ways that you could get hold of the constructor,
but only <tt>Yummy</tt> is usable as a type name.
<h3> <a name="PublicExtensionTypes"></a>Public extension types</h3>
An extension type can be declared <b>public</b>, in which case a <b>.h</b>
file is generated containing declarations for its object struct and type
object. By including the <b>.h</b> file in external C code that you write,
that code can access the attributes of the extension type.
<h3> <a name="NameSpecClause"></a>Name specification clause</h3>
The part of the class declaration in square brackets is a special feature
only available for <b>extern</b> or <b>public</b> extension types. The full
form of this clause is
<blockquote><tt>[object </tt><i>object_struct_name</i><tt>, type </tt><i>type_object_name</i><span style="font-family: monospace;"> ]</span></blockquote>
where <i>object_struct_name</i> is the name to assume for the type's C
struct, and <i>type_object_name</i> is the name to assume for the type's
statically declared type object. (The object and type clauses can be written
in either order.)
<p>If the extension type declaration is inside a <b>cdef extern from</b>
block, the <b>object</b> clause is required, because Cython must be able to
generate code that is compatible with the declarations in the header file.
Otherwise, for <b>extern</b> extension types, the <b>object</b> clause is
optional. </p>
<p>For <b>public</b> extension types, the <b>object</b> and <b>type</b> clauses
are both required, because Cython must be able to generate code that is compatible
with external C code. </p>
<p> </p>
<hr width="100%"> <br>
Back to the <a href="overview.html">Language Overview</a> <br>
&nbsp; <br>
<br>
</body></html>
\ No newline at end of file
Doc/index.html deleted 100644 → 0
View file @ 5a1c4958
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.51 (Macintosh; I; PPC) [Netscape]">
<title>Cython - Front Page</title>
</head>
<body>
&nbsp;
<table CELLSPACING=0 CELLPADDING=10 WIDTH="500" >
<tr>
<td VALIGN=TOP BGCOLOR="#FF9218"><font face="Arial,Helvetica"><font size=+4>Cython</font></font></td>
<td ALIGN=RIGHT VALIGN=TOP WIDTH="200" BGCOLOR="#5DBACA"><font face="Arial,Helvetica"><font size=+1>A
smooth blend of the finest Python&nbsp;</font></font>
<br><font face="Arial,Helvetica"><font size=+1>with the unsurpassed power&nbsp;</font></font>
<br><font face="Arial,Helvetica"><font size=+1>of raw C.</font></font></td>
</tr>
</table>
<blockquote><font size=+1>Welcome to Cython, a language for writing Python
extension modules. Cython makes creating an extension module is almost as
easy as creating a Python module! To find out more, consult one of the
edifying documents below.</font></blockquote>
<h1>
<font face="Arial,Helvetica"><font size=+2>Documentation</font></font></h1>
<blockquote>
<h2>
<font face="Arial,Helvetica"><font size=+1><a href="About.html">About Cython</a></font></font></h2>
<blockquote><font size=+1>Read this to find out what Cython is all about
and what it can do for you.</font></blockquote>
<h2>
<font face="Arial,Helvetica"><font size=+1><a href="overview.html">Language
Overview</a></font></font></h2>
<blockquote><font size=+1>A description of all the features of the Cython
language. This is the closest thing to a reference manual in existence
yet.</font></blockquote>
<h2>
<font face="Arial,Helvetica"><font size=+1><a href="FAQ.html">FAQ</a></font></font></h2>
<blockquote><font size=+1>Want to know how to do something in Cython? Check
here first<font face="Arial,Helvetica">.</font></font></blockquote>
</blockquote>
<h1>
<font face="Arial,Helvetica"><font size=+2>Other Resources</font></font></h1>
<blockquote>
<h2>
<font face="Arial,Helvetica"><font size=+1><a href="http://www.cosc.canterbury.ac.nz/~greg/python/Cython/mpj17-pyrex-guide/">Michael's
Quick Guide to Cython</a></font></font></h2>
<blockquote><font size=+1>This tutorial-style presentation will take you
through the steps of creating some Cython modules to wrap existing C libraries.
Contributed by <a href="mailto:mpj17@cosc.canterbury.ac.nz">Michael JasonSmith</a>.</font></blockquote>
<h2>
<font face="Arial,Helvetica"><font size=+1><a href="mailto:greg@cosc.canterbury.ac.nz">Mail
to the Author</a></font></font></h2>
<blockquote><font size=+1>If you have a question that's not answered by
anything here, you're not sure about something, or you have a bug to report
or a suggestion to make, or anything at all to say about Cython, feel free
to email me:<font face="Arial,Helvetica"> </font><tt><a href="mailto:greg@cosc.canterbury.ac.nz">greg@cosc.canterbury.ac.nz</a></tt></font></blockquote>
</blockquote>
</body>
</html>
Doc/overview.html deleted 100644 → 0
View file @ 5a1c4958
<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]">
<title>Cython Language Overview</title>
</head>
<body>
<h1> <hr width="100%">Overview of the Cython Language&nbsp; <hr width="100%"></h1>
This document informally describes the extensions to the Python language
made by Cython. Some day there will be a reference manual covering everything
in more detail. <br>
&nbsp;
<h2> Contents</h2>
<ul>
<li> <a href="#Basics">Basics</a></li>
<ul>
<li> <a href="#PyFuncsVsCFuncs">Python functions vs. C functions</a></li>
<li> <a href="#PyObjParams">Python objects as parameters</a></li>
<li> <a href="#CVarAndTypeDecls">C variable and type definitions</a></li><li><a href="#AutomaticTypeConversions">Automatic type conversions</a></li>
<ul>
<li><a href="#PyToCStringCaveats">Caveats when using a Python string in a C context</a></li>
</ul>
<li> <a href="#ScopeRules">Scope rules</a></li>
<li> <a href="#StatsAndExprs">Statements and expressions</a></li>
<ul>
<li> <a href="#ExprSyntaxDifferences">Differences between C and Cython
expressions<br>
</a></li>
<li> <a href="#ForFromLoop">Integer for-loops</a></li>
</ul>
<li> <a href="#ExceptionValues">Error return values</a></li>
<ul>
<li> <a href="#CheckingReturnValues">Checking return values of non-Cython
functions</a></li>
</ul>
<li> <a href="#IncludeStatement">The <tt>include</tt> statement</a></li>
</ul>
<li> <a href="#InterfacingWithExternal">Interfacing with External C Code</a></li>
<ul>
<li> <a href="#ExternDecls">External declarations</a></li>
<ul>
<li> <a href="#ReferencingHeaders">Referencing C header files</a></li>
<li> <a href="#StructDeclStyles">Styles of struct/union/enum declaration</a></li>
<li> <a href="#AccessingAPI">Accessing Python/C API routines</a></li>
<li> <a href="#CNameSpecs">Resolving naming conflicts - C name specifications</a></li>
</ul>
<li> <a href="#PublicDecls">Public declarations</a></li>
</ul>
<li> <a href="extension_types.html">Extension Types</a> <font color="#006600">(Section revised in 0.9)</font></li>
<li> <a href="sharing.html">Sharing Declarations Between Cython Modules</a>
<font color="#006600">(NEW in 0.8)</font></li>
<li> <a href="#Limitations">Limitations</a></li>
<ul>
<li> <a href="#Unsupported">Unsupported Python features</a></li>
<li> <a href="#SemanticDifferences">Semantic differences between Python
and Cython</a></li>
</ul>
</ul>
<h2> <hr width="100%"><a name="Basics"></a>Basics
<hr width="100%"></h2>
This section describes the basic features of the Cython language. The facilities
covered in this section allow you to create Python-callable functions that
manipulate C data structures and convert between Python and C data types.
Later sections will cover facilities for <a href="#InterfacingWithExternal">wrapping external C code</a>, <a href="extension_types.html">creating new Python types</a> and <a href="sharing.html">cooperation between Cython modules</a>.
<h3> <a name="PyFuncsVsCFuncs"></a>Python functions vs. C functions</h3>
There are two kinds of function definition in Cython:
<p><b>Python functions</b> are defined using the <b>def</b> statement, as
in Python. They take Python objects as parameters and return Python objects.
</p>
<p><b>C functions</b> are defined using the new <b>cdef</b> statement. They
take either Python objects or C values as parameters, and can return either
Python objects or C values. </p>
<p>Within a Cython module, Python functions and C functions can call each other
freely, but only Python functions can be called from outside the module by
interpreted Python code. So, any functions that you want to "export" from
your Cython module must be declared as Python functions using <span style="font-weight: bold;">def</span>. </p>
<p>Parameters of either type of function can be declared to have C data types,
using normal C declaration syntax. For example, </p>
<blockquote> <pre>def spam(int i, char *s):<br>&nbsp;&nbsp;&nbsp; ...</pre>
<pre>cdef int eggs(unsigned long l, float f):<br>&nbsp;&nbsp;&nbsp; ...</pre>
</blockquote>
When a parameter of a Python function is declared to have a C data type,
it is passed in as a Python object and automatically converted to a C value,
if possible. Automatic conversion is currently only possible for numeric
types and string types; attempting to use any other type for the parameter
of a Python function will result in a compile-time error.
<p>C functions, on the other hand, can have parameters of any type, since
they're passed in directly using a normal C function call. </p>
<h3> <a name="PyObjParams"></a>Python objects as parameters and return values</h3>
If no type is specified for a parameter or return value, <i>it is assumed
to be a Python object.</i> (Note that this is different from the C convention,
where it would default to <tt>int</tt>.) For example, the following defines
a C function that takes two Python objects as parameters and returns a Python
object:
<blockquote> <pre>cdef spamobjs(x, y):<br>&nbsp;&nbsp;&nbsp; ...</pre>
</blockquote>
Reference counting for these objects is performed automatically according
to the standard Python/C API rules (i.e. borrowed references are taken as
parameters and a new reference is returned).
<p>The name <b>object</b> can also be used to explicitly declare something
as a Python object. This can be useful if the name being declared would otherwise
be taken as the name of a type, for example, </p>
<blockquote> <pre>cdef ftang(object int):<br>&nbsp;&nbsp;&nbsp; ...</pre>
</blockquote>
declares a parameter called <tt>int</tt> which is a Python object. You
can also use <b>object </b>as the explicit return type of a function, e.g.
<blockquote> <pre>cdef object ftang(object int):<br>&nbsp;&nbsp;&nbsp; ...</pre>
</blockquote>
In the interests of clarity, it is probably a good idea to always be explicit
about <b>object </b>parameters in C functions.
<h3> <a name="CVarAndTypeDecls"></a>C variable and type definitions</h3>
The <b>cdef</b> statement is also used to declare C variables, either
local or module-level:
<blockquote> <pre>cdef int i, j, k<br>cdef float f, g[42], *h</pre>
</blockquote>
and C struct, union or enum types:
<blockquote> <pre>cdef struct Grail:<br>&nbsp;&nbsp;&nbsp; int age<br>&nbsp;&nbsp;&nbsp; float volume</pre>
<pre>cdef union Food:<br>&nbsp;&nbsp;&nbsp; char *spam<br>&nbsp;&nbsp;&nbsp; float *eggs</pre>
<pre>cdef enum CheeseType:<br>&nbsp;&nbsp;&nbsp; cheddar, edam,&nbsp;<br>&nbsp;&nbsp;&nbsp; camembert</pre>
<pre>cdef enum CheeseState:<br>&nbsp;&nbsp;&nbsp; hard = 1<br>&nbsp;&nbsp;&nbsp; soft = 2<br>&nbsp;&nbsp;&nbsp; runny = 3</pre>
</blockquote>
There is currently no special syntax for defining a constant, but you
can use an anonymous enum declaration for this purpose, for example,
<blockquote><tt>cdef enum:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; tons_of_spam = 3</tt></blockquote>
Note that the words <span style="font-family: monospace;">struct</span>, <span style="font-family: monospace;">union</span> and <span style="font-family: monospace;">enum</span> are used only when <i>defining</i> a type, not when referring to it. For example, to declare a variable pointing
to a Grail you would write
<blockquote> <pre>cdef Grail *gp</pre>
</blockquote>
and <i>not</i>
<blockquote> <pre>cdef struct Grail *gp <font color="#ed181e"># WRONG</font></pre>
</blockquote>
There is also a <b>ctypedef</b> statement for giving names to types, e.g.
<blockquote> <pre>ctypedef unsigned long ULong</pre>
<pre>ctypedef int *IntPtr<br></pre></blockquote>
<h3><a name="AutomaticTypeConversions"></a>Automatic type conversions</h3>
In most situations, automatic conversions will be performed for the
basic numeric and string types when a Python object is used in a
context requiring a C value, or vice versa. The following table
summarises the conversion possibilities.<br>
<br>
<table style="margin-left: auto; margin-right: auto; width: 10%; text-align: left;" border="1" cellpadding="4" cellspacing="0">
<tbody>
<tr>
<th style="vertical-align: top; width: 40%; white-space: nowrap;">C types<br>
</th>
<th style="vertical-align: top; width: 150px; white-space: nowrap;">From Python types<br>
</th>
<th style="vertical-align: top; width: 150px; white-space: nowrap;">To Python types<br>
</th>
</tr>
<tr>
<td colspan="1" rowspan="1" style="vertical-align: top; width: 40%; white-space: nowrap;">[unsigned] char<br>
[unsigned] short<br>
int, long</td>
<td colspan="1" rowspan="1" style="vertical-align: top; width: 150px; white-space: nowrap;">int, long<br>
</td>
<td colspan="1" rowspan="1" style="vertical-align: top; width: 150px; white-space: nowrap;">int<br>
</td>
</tr>
<tr>
</tr>
<tr>
<td colspan="1" rowspan="1" style="vertical-align: top; width: 40%; white-space: nowrap;">unsigned int<br>
unsigned long<br>
[unsigned] long long<br>
</td>
<td colspan="1" rowspan="1" style="vertical-align: top; white-space: nowrap;">int, long<br>
<br>
</td>
<td colspan="1" rowspan="1" style="vertical-align: top; white-space: nowrap;">long<br>
<br>
</td>
</tr>
<tr>
</tr>
<tr>
<td style="vertical-align: top; width: 40%; white-space: nowrap;">float, double, long double<br>
</td>
<td style="vertical-align: top; width: 150px; white-space: nowrap;">int, long, float<br>
</td>
<td style="vertical-align: top; width: 150px; white-space: nowrap;">float<br>
</td>
</tr>
<tr>
<td style="vertical-align: top; width: 40%; white-space: nowrap;">char *<br>
</td>
<td style="vertical-align: top; width: 150px; white-space: nowrap;">str<span style="font-style: italic;"></span><br>
</td>
<td style="vertical-align: top; width: 150px; white-space: nowrap;">str<br>
</td>
</tr>
</tbody>
</table>
<br>
<h4><a name="PyToCStringCaveats"></a>Caveats when using a Python string in a C context</h4>
You need to be careful when using a Python string in a context expecting a <span style="font-family: monospace;">char *</span>.
In this situation, a pointer to the contents of the Python string is
used, which is only valid as long as the Python string exists. So you
need to make sure that a reference to the original Python string is
held for as long as the C string is needed. If you can't guarantee that
the Python string will live long enough, you will need to copy the C
string.<br>
<br>
Cython detects and prevents <span style="font-style: italic;">some</span> mistakes of this kind. For instance, if you attempt something like<br>
<pre style="margin-left: 40px;">cdef char *s<br>s = pystring1 + pystring2</pre>
then Cython will produce the error message "<span style="font-weight: bold;">Obtaining char * from temporary Python value</span>".
The reason is that concatenating the two Python strings produces a new
Python string object that is referenced only by a temporary internal
variable that Cython generates. As soon as the statement has finished,
the temporary variable will be decrefed and the Python string
deallocated, leaving <span style="font-family: monospace;">s</span> dangling. Since this code could not possibly work, Cython refuses to compile it.<br>
<br>
The solution is to assign the result of the concatenation to a Python variable, and then obtain the char * from that, i.e.<br>
<pre style="margin-left: 40px;">cdef char *s<br>p = pystring1 + pystring2<br>s = p<br></pre>
It is then your responsibility to hold the reference <span style="font-family: monospace;">p</span> for as long as necessary.<br>
<br>
Keep in mind that the rules used to detect such errors are only
heuristics. Sometimes Cython will complain unnecessarily, and sometimes
it will fail to detect a problem that exists. Ultimately, you need to
understand the issue and be careful what you do.<br>
<blockquote>
</blockquote>
<h3> <a name="ScopeRules"></a>Scope rules</h3>
Cython determines whether a variable belongs to a local scope, the module
scope, or the built-in scope <i>completely statically.</i> As with Python,
assigning to a variable which is not otherwise declared implicitly declares
it to be a Python variable residing in the scope where it is assigned. Unlike
Python, however, a name which is referred to but not declared or assigned
is assumed to reside in the <i>builtin scope, </i>not the module scope.
Names added to the module dictionary at run time will not shadow such names.
<p>You can use a <b>global</b> statement at the module level to explicitly
declare a name to be a module-level name when there would otherwise not be
any indication of this, for example, </p>
<blockquote><tt>global __name__</tt> <br>
<tt>print __name__</tt></blockquote>
Without the <b>global</b> statement, the above would print the name of
the builtins module.<br>
<br>
Note: A consequence of these rules is that the module-level scope behaves
the same way as a Python local scope if you refer to a variable before assigning
to it. In particular, tricks such as the following will <i>not</i> work
in Cython:<br>
<blockquote> <pre>try:<br>&nbsp; x = True<br>except NameError:<br>&nbsp; True = 1<br></pre>
</blockquote>
because, due to the assignment, the True will always be looked up in the
module-level scope. You would have to do something like this instead:<br>
<blockquote> <pre>import __builtin__<br>try:<br> True = __builtin__.True<br>except AttributeError:<br> True = 1<br></pre>
</blockquote>
<hr width="100%">
<h3> <a name="StatsAndExprs"></a>Statements and expressions</h3>
Control structures and expressions follow Python syntax for the most part.
When applied to Python objects, they have the same semantics as in Python
(unless otherwise noted). Most of the Python operators can also be applied
to C values, with the obvious semantics.
<p>If Python objects and C values are mixed in an expression, conversions
are performed automatically between Python objects and C numeric or string
types. </p>
<p>Reference counts are maintained automatically for all Python objects, and
all Python operations are automatically checked for errors, with appropriate
action taken. </p>
<h4> <a name="ExprSyntaxDifferences"></a>Differences between C and Cython
expressions</h4>
There
are some differences in syntax and semantics between C expressions and
Cython expressions, particularly in the area of C constructs which have
no direct equivalent in Python.<br>
<ul>
<li>An integer literal without an <span style="font-family: monospace; font-weight: bold;">L</span> suffix is treated as a C constant, and will be truncated to whatever size your C compiler thinks appropriate. With an <span style="font-family: monospace; font-weight: bold;">L</span> suffix, it will be converted to Python long integer (even if it would be small enough to fit into a C int).<br>
<br>
</li>
<li> There is no <b><tt>-&gt;</tt></b> operator in Cython. Instead of <tt>p-&gt;x</tt>,
use <tt>p.x</tt></li>
&nbsp; <li> There is no <b><tt>*</tt></b> operator in Cython. Instead of
<tt>*p</tt>, use <tt>p[0]</tt></li>
&nbsp; <li> There is an <b><tt>&amp;</tt></b> operator, with the same semantics
as in C.</li>
&nbsp; <li> The null C pointer is called <b><tt>NULL</tt></b>, not 0 (and
<tt>NULL</tt> is a reserved word).</li>
&nbsp; <li> Character literals are written with a <b>c</b> prefix, for
example:</li>
<ul>
<pre>c'X'</pre>
</ul>
<li> Type casts are written <b><tt>&lt;type&gt;value</tt></b> , for example:</li>
<ul>
<pre>cdef char *p, float *q<br>p = &lt;char*&gt;q</pre>
</ul>
<i><b>Warning</b>: Don't attempt to use a typecast to convert between
Python and C data types -- it won't do the right thing. Leave Cython to perform
the conversion automatically.</i>
</ul>
<h4> <a name="ForFromLoop"></a>Integer for-loops</h4>
You should be aware that a for-loop such as
<blockquote><tt>for i in range(n):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
won't be very fast, even if <tt>i</tt> and <tt>n</tt> are declared as
C integers, because <tt>range</tt> is a Python function. For iterating over
ranges of integers, Cython has another form of for-loop:
<blockquote><tt>for i from 0 &lt;= i &lt; n:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
If the loop variable and the lower and upper bounds are all C integers,
this form of loop will be much faster, because Cython will translate it into
pure C code.
<p>Some things to note about the <tt>for-from</tt> loop: </p>
<ul>
<li> The target expression must be a variable name.</li>
<li> The name between the lower and upper bounds must be the same as
the target name.</li>
<li> The direction of iteration is determined by the relations. If they
are both from the set {<tt>&lt;</tt>, <tt>&lt;=</tt>} then it is upwards;
if they are both from the set {<tt>&gt;</tt>, <tt>&gt;=</tt>} then it is
downwards. (Any other combination is disallowed.)</li>
</ul>
Like other Python looping statements, <tt>break</tt> and <tt>continue</tt> may be used in the body, and the loop may have an <tt>else</tt> clause.
<h2> <hr width="100%"></h2>
<h3> <a name="ExceptionValues"></a>Error return values</h3>
If you don't do anything special, a function declared with <b>cdef</b> that does not return a Python object has no way of reporting Python exceptions
to its caller. If an exception is detected in such a function, a warning
message is printed and the exception is ignored.
<p>If you want a C function that does not return a Python object to be able
to propagate exceptions to its caller, you need to declare an <b>exception
value</b> for it. Here is an example: </p>
<blockquote><tt>cdef int spam() except -1:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
With this declaration, whenever an exception occurs inside <tt>spam</tt>,
it will immediately return with the value <tt>-1</tt>. Furthermore, whenever
a call to <tt>spam</tt> returns <tt>-1</tt>, an exception will be assumed
to have occurred and will be propagated.
<p>When you declare an exception value for a function, you should never explicitly
return that value. If all possible return values are legal and you can't
reserve one entirely for signalling errors, you can use an alternative form
of exception value declaration: </p>
<blockquote><tt>cdef int spam() except? -1:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
The "?" indicates that the value <tt>-1</tt> only indicates a <i>possible</i> error. In this case, Cython generates a call to <tt>PyErr_Occurred</tt>if the
exception value is returned, to make sure it really is an error.
<p>There is also a third form of exception value declaration: </p>
<blockquote><tt>cdef int spam() except *:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
This form causes Cython to generate a call to <tt>PyErr_Occurred</tt> after
<i>every</i> call to spam, regardless of what value it returns. If you have
a function returning <tt>void</tt> that needs to propagate errors, you will
have to use this form, since there isn't any return value to test.
<p>Some things to note: </p>
<ul>
<li> Currently, exception values can only declared for functions returning
an integer, float or pointer type, and the value must be a <i>literal</i>,
not an expression (although it can be negative). The only possible pointer
exception value is <tt>NULL</tt>. Void functions can only use the <tt>except
*</tt> form.</li>
<br>
&nbsp; <li> The exception value specification is part of the signature
of the function. If you're passing a pointer to a function as a parameter
or assigning it to a variable, the declared type of the parameter or variable
must have the same exception value specification (or lack thereof). Here
is an example of a pointer-to-function declaration with an exception value:</li>
<ul>
<pre><tt>int (*grail)(int, char *) except -1</tt></pre>
</ul>
<li> You don't need to (and shouldn't) declare exception values for functions
which return Python objects. Remember that a function with no declared return
type implicitly returns a Python object.</li>
</ul>
<h4> <a name="CheckingReturnValues"></a>Checking return values of non-Cython
functions</h4>
It's important to understand that the <tt>except</tt> clause does <i>not</i> cause an error to be <i>raised</i> when the specified value is returned. For
example, you can't write something like
<blockquote> <pre>cdef extern FILE *fopen(char *filename, char *mode) except NULL <font color="#ed181e"># WRONG!</font></pre>
</blockquote>
and expect an exception to be automatically raised if a call to fopen
returns NULL. The except clause doesn't work that way; its only purpose
is for <i>propagating</i> exceptions that have already been raised, either
by a Cython function or a C function that calls Python/C API routines. To
get an exception from a non-Python-aware function such as fopen, you will
have to check the return value and raise it yourself, for example,
<blockquote> <pre>cdef FILE *p<br>p = fopen("spam.txt", "r")<br>if p == NULL:<br>&nbsp;&nbsp;&nbsp; raise SpamError("Couldn't open the spam file")</pre>
</blockquote>
<h4> <hr width="100%"></h4>
<h4> <a name="IncludeStatement"></a>The <tt>include</tt> statement</h4>
For convenience, a large Cython module can be split up into a number of
files which are put together using the <b>include</b> statement, for example
<blockquote> <pre>include "spamstuff.pxi"</pre>
</blockquote>
The contents of the named file are textually included at that point. The
included file can contain any complete top-level Cython statements, including
other <b>include</b> statements. The <b>include</b> statement itself can
only appear at the top level of a file.
<p>The <b>include</b> statement can also be used in conjunction with <a href="#PublicDecls"><b>public</b> declarations</a> to make C functions and
variables defined in one Cython module accessible to another. However, note
that some of these uses have been superseded by the facilities described
in <a href="sharing.html">Sharing Declarations Between Cython Modules</a>,
and it is expected that use of the <b>include</b> statement for this purpose
will be phased out altogether in future versions. </p>
<h2> <hr width="100%"><a name="InterfacingWithExternal"></a>Interfacing with External
C Code
<hr width="100%"></h2>
One of the main uses of Cython is wrapping existing libraries of C code.
This is achieved by using <a href="#ExternDecls">external declarations</a> to declare the C functions and variables from the library that you want to
use.
<p>You can also use <a href="#PublicDecls">public declarations</a> to make
C functions and variables defined in a Cython module available to external
C code. The need for this is expected to be less frequent, but you might
want to do it, for example, if you are embedding Python in another application
as a scripting language. Just as a Cython module can be used as a bridge to
allow Python code to call C code, it can also be used to allow C code to
call Python code. </p>
<h3> <a name="ExternDecls"></a>External declarations</h3>
By default, C functions and variables declared at the module level are
local to the module (i.e. they have the C <b>static</b> storage class). They
can also be declared <b>extern</b> to specify that they are defined elsewhere,
for example:
<blockquote> <pre>cdef extern int spam_counter</pre>
<pre>cdef extern void order_spam(int tons)</pre>
</blockquote>
<blockquote> </blockquote>
<h4> <a name="ReferencingHeaders"></a>Referencing C header files</h4>
When you use an extern definition on its own as in the examples above,
Cython includes a declaration for it in the generated C file. This can cause
problems if the declaration doesn't exactly match the declaration that will
be seen by other C code. If you're wrapping an existing C library, for example,
it's important that the generated C code is compiled with exactly the same
declarations as the rest of the library.
<p>To achieve this, you can tell Cython that the declarations are to be found
in a C header file, like this: </p>
<blockquote> <pre>cdef extern from "spam.h":</pre>
<pre>&nbsp;&nbsp;&nbsp; int spam_counter</pre>
<pre>&nbsp;&nbsp;&nbsp; void order_spam(int tons)</pre>
</blockquote>
The <b>cdef extern from</b> clause does three things:
<ol>
<li> It directs Cython to place a <b>#include</b> statement for the named
header file in the generated C code.<br>
</li>
&nbsp; <li> It prevents Cython from generating any C code for the declarations
found in the associated block.<br>
</li>
&nbsp; <li> It treats all declarations within the block as though they
started with <b>cdef extern</b>.</li>
</ol>
It's important to understand that Cython does <i>not</i> itself read the
C header file, so you still need to provide Cython versions of any declarations
from it that you use. However, the Cython declarations don't always have to
exactly match the C ones, and in some cases they shouldn't or can't. In particular:
<ol>
<li> Don't use <b>const</b>. Cython doesn't know anything about const,
so just leave it out. Most of the time this shouldn't cause any problem,
although on rare occasions you might have to use a cast.<sup><a href="#Footnote1"> 1</a></sup><br>
</li>
&nbsp; <li> Leave out any platform-specific extensions to C declarations
such as <b>__declspec()</b>.<br>
</li>
&nbsp; <li> If the header file declares a big struct and you only want
to use a few members, you only need to declare the members you're interested
in. Leaving the rest out doesn't do any harm, because the C compiler will
use the full definition from the header file.<br>
<br>
In some cases, you might not need <i>any</i> of the struct's members, in
which case you can just put <tt>pass</tt> in the body of the struct declaration,
e.g.<br>
<br>
<tt>&nbsp; &nbsp; cdef extern from "foo.h":<br>
&nbsp; &nbsp; &nbsp; &nbsp; struct spam:<br>
&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; pass</tt><br>
<br>
Note that you can only do this inside a <b>cdef extern from</b> block; struct
declarations anywhere else must be non-empty.<br>
<br>
</li>
<li> If the header file uses typedef names such as <b>size_t </b>to refer
to platform-dependent flavours of numeric types, you will need a corresponding
<b>ctypedef</b> statement, but you don't need to match the type exactly,
just use something of the right general kind (int, float, etc). For example,</li>
<ol>
<pre>ctypedef int size_t</pre>
</ol>
will work okay whatever the actual size of a size_t is (provided the header
file defines it correctly). <br>
&nbsp; <li> If the header file uses macros to define constants, translate
them into a dummy <b>enum</b> declaration.<br>
</li>
&nbsp; <li> If the header file defines a function using a macro, declare
it as though it were an ordinary function, with appropriate argument and
result types.</li>
</ol>
A few more tricks and tips:
<ul>
<li> If you want to include a C header because it's needed by another
header, but don't want to use any declarations from it, put <tt><font size="+1">pass</font></tt> in the extern-from block:</li>
</ul>
<ul>
<ul>
<tt>cdef extern from "spam.h":</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; pass</tt> </ul>
</ul>
<ul>
<li> If you want to include some external declarations, but don't want
to specify a header file (because it's included by some other header that
you've already included) you can put <tt>*</tt> in place of the header file
name:</li>
</ul>
<blockquote> <blockquote><tt>cdef extern from *:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></blockquote>
</blockquote>
<h4> <a name="StructDeclStyles"></a>Styles of struct, union and enum declaration</h4>
There are two main ways that structs, unions and enums can be declared
in C header files: using a tag name, or using a typedef. There are also some
variations based on various combinations of these.
<p>It's important to make the Cython declarations match the style used in the
header file, so that Cython can emit the right sort of references to the type
in the code it generates. To make this possible, Cython provides two different
syntaxes for declaring a struct, union or enum type. The style introduced
above corresponds to the use of a tag name. To get the other style, you prefix
the declaration with <b>ctypedef</b>, as illustrated below. </p>
<p>The following table shows the various possible styles that can be found
in a header file, and the corresponding Cython declaration that you should
put in the <b>cdef exern from </b>block. Struct declarations are used as
an example; the same applies equally to union and enum declarations. </p>
<p>Note that in all the cases below, you refer to the type in Cython code simply
as <tt><font size="+1">Foo</font></tt>, not <tt><font size="+1">struct Foo</font></tt>.
<br>
&nbsp; <table cellpadding="5">
<tbody>
<tr bgcolor="#8cbc1c" valign="top">
<td bgcolor="#8cbc1c">&nbsp;</td>
<td bgcolor="#ff9933" nowrap="nowrap"><b>C code</b></td>
<td bgcolor="#66cccc" valign="top"><b>Possibilities for corresponding
Cython code</b></td>
<td bgcolor="#99cc33" valign="top"><b>Comments</b></td>
</tr>
<tr bgcolor="#8cbc1c" valign="top">
<td>1</td>
<td bgcolor="#ff9900"><tt>struct Foo {</tt> <br>
<tt>&nbsp; ...</tt> <br>
<tt>};</tt></td>
<td bgcolor="#66cccc"><tt>cdef struct Foo:</tt> <br>
<tt>&nbsp; ...</tt></td>
<td>Cython will refer to the type as <tt>struct Foo </tt>in the generated
C code<tt>.</tt></td>
</tr>
<tr bgcolor="#8cbc1c" valign="top">
<td valign="top">2</td>
<td bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct {</tt> <br>
<tt>&nbsp; ...</tt> <br>
<tt>} Foo;</tt></td>
<td bgcolor="#66cccc" valign="top"><tt>ctypedef struct Foo:</tt> <br>
<tt>&nbsp; ...</tt></td>
<td valign="top">Cython will refer to the type simply as <tt>Foo</tt>
in the generated C code.</td>
</tr>
<tr bgcolor="#8cbc1c" valign="top">
<td rowspan="2">3</td>
<td rowspan="2" bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct
foo {</tt> <br>
<tt>&nbsp; ...</tt> <br>
<tt>} Foo;</tt></td>
<td bgcolor="#66cccc" nowrap="nowrap" valign="top"><tt>cdef struct foo:</tt> <br>
<tt>&nbsp; ...</tt> <br>
<tt>ctypedef foo Foo #optional</tt></td>
<td rowspan="2" valign="top">If the C header uses both a tag and a typedef
with <i>different</i> names, you can use either form of declaration in Cython
(although if you need to forward reference the type, you'll have to use
the first form).</td>
</tr>
<tr>
<td bgcolor="#66cccc"><tt>ctypedef struct Foo:</tt> <br>
<tt>&nbsp; ...</tt></td>
</tr>
<tr bgcolor="#8cbc1c" valign="top">
<td>4</td>
<td bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct Foo {</tt> <br>
<tt>&nbsp; ...</tt> <br>
<tt>} Foo;</tt></td>
<td bgcolor="#66cccc" valign="top"><tt>cdef struct Foo:</tt> <br>
<tt>&nbsp; ...</tt></td>
<td>If the header uses the <i>same</i> name for the tag and the typedef,
you won't be able to include a <b>ctypedef</b> for it -- but then, it's not
necessary.</td>
</tr>
</tbody> </table>
</p>
<h4> <a name="AccessingAPI"></a>Accessing Python/C API routines</h4>
One particular use of the <b>cdef extern from</b> statement is for gaining
access to routines in the Python/C API. For example,
<blockquote> <pre>cdef extern from "Python.h":</pre>
<pre>&nbsp;&nbsp;&nbsp; object PyString_FromStringAndSize(char *s, int len)</pre>
</blockquote>
will allow you to create Python strings containing null bytes.
<p> </p>
<hr width="100%">
<h3> <a name="CNameSpecs"></a>Resolving naming conflicts - C name specifications</h3>
Each Cython module has a single module-level namespace for both Python
and C names. This can be inconvenient if you want to wrap some external
C functions and provide the Python user with Python functions of the same
names.
<p>Cython 0.8 provides a couple of different ways of solving this problem.
The best way, especially if you have many C functions to wrap, is probably
to put the extern C function declarations into a different namespace using
the facilities described in the section on <a href="sharing.html">sharing
declarations between Cython modules</a>. </p>
<p>The other way is to use a <b>c name specification</b> to give different
Cython and C names to the C function. Suppose, for example, that you want
to wrap an external function called <tt>eject_tomato</tt>. If you declare
it as </p>
<blockquote> <pre>cdef extern void c_eject_tomato "eject_tomato" (float speed)</pre>
</blockquote>
then its name inside the Cython module will be <tt>c_eject_tomato</tt>,
whereas its name in C will be <tt>eject_tomato</tt>. You can then wrap it
with
<blockquote> <pre>def eject_tomato(speed):<br>&nbsp; c_eject_tomato(speed)</pre>
</blockquote>
so that users of your module can refer to it as <tt>eject_tomato</tt>.
<p>Another use for this feature is referring to external names that happen
to be Cython keywords. For example, if you want to call an external function
called <tt>print</tt>, you can rename it to something else in your Cython
module. </p>
<p>As well as functions, C names can be specified for variables, structs,
unions, enums, struct and union members, and enum values. For example, </p>
<blockquote> <pre>cdef extern int one "ein", two "zwei"<br>cdef extern float three "drei"<br><br>cdef struct spam "SPAM":<br>&nbsp; int i "eye"</pre>
<tt>cdef enum surprise "inquisition":</tt> <br>
<tt>&nbsp; first "alpha"</tt> <br>
<tt>&nbsp; second "beta" = 3</tt></blockquote>
<hr width="100%">
<h3> <a name="PublicDecls"></a>Public Declarations</h3>
You can make C variables and functions defined in a Cython module accessible
to external C code (or another Cython module) using the <b><tt>public</tt></b> keyword, as follows:
<blockquote><tt>cdef public int spam # public variable declaration</tt> <p><tt>cdef public void grail(int num_nuns): # public function declaration</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; ...</tt></p>
</blockquote>
If there are any <tt>public</tt> declarations in a Cython module, a <b>.h</b> file is generated containing equivalent C declarations for inclusion in other
C code.
<p>Cython also generates a <b>.pxi</b> file containing Cython versions of the
declarations for inclusion in another Cython module using the <b><a href="#IncludeStatement">include</a></b> statement. If you use this, you
will need to arrange for the module using the declarations to be linked
against the module defining them, and for both modules to be available to
the dynamic linker at run time. I haven't tested this, so I can't say how
well it will work on the various platforms. </p>
<blockquote>NOTE: If all you want to export is an extension type, there is
now a better way -- see <a href="sharing.html">Sharing Declarations Between
Cython Modules</a>.</blockquote>
<h2> <hr width="100%">Extension Types
<hr width="100%"></h2>
One of the most powerful features of Cython is the ability to easily create
new built-in Python types, called <b>extension types</b>. This is a major
topic in itself, so there is a&nbsp; <a href="extension_types.html">separate
page</a> devoted to it.
<h2> <hr width="100%">Sharing Declarations Between Cython Modules
<hr width="100%"></h2>
Cython 0.8 introduces a substantial new set of facilities allowing a Cython
module to easily import and use C declarations and extension types from another
Cython module. You can now create a set of co-operating Cython modules just
as easily as you can create a set of co-operating Python modules. There is
a <a href="sharing.html">separate page</a> devoted to this topic.
<h2> <hr width="100%"><a name="Limitations"></a>Limitations
<hr width="100%"></h2>
<h3> <a name="Unsupported"></a>Unsupported Python features</h3>
Cython is not quite a full superset of Python. The following restrictions
apply:
<blockquote> <li> Function definitions (whether using <b>def</b> or <b>cdef</b>)
cannot be nested within other function definitions.<br>
</li>
&nbsp; <li> Class definitions can only appear at the top level of a module,
not inside a function.<br>
</li>
&nbsp; <li> The<tt> import *</tt> form of import is not allowed anywhere
(other forms of the import statement are fine, though).<br>
</li>
&nbsp; <li> Generators cannot be defined in Cython.<br>
<br>
</li>
<li> The <tt>globals()</tt> and <tt>locals()</tt> functions cannot be
used.</li>
</blockquote>
The above restrictions will most likely remain, since removing them would
be difficult and they're not really needed for Cython's intended applications.
<p>There are also some temporary limitations, which may eventually be lifted, including:
</p>
<blockquote> <li> Class and function definitions cannot be placed inside
control structures.<br>
</li>
&nbsp; <li> In-place arithmetic operators (+=, etc) are not yet supported.<br>
</li>
&nbsp; <li> List comprehensions are not yet supported.<br>
</li>
&nbsp; <li> There is no support for Unicode.<br>
</li>
&nbsp; <li> Special methods of extension types cannot have functioning
docstrings.<br>
<br>
</li>
<li> The use of string literals as comments is not recommended at present,
because Cython doesn't optimize them away, and won't even accept them in
places where executable statements are not allowed.</li>
</blockquote>
<h3> <a name="SemanticDifferences"></a>Semantic differences between Python
and Cython</h3>
<h4> Behaviour of class scopes</h4>
In Python, referring to a method of a class inside the class definition,
i.e. while the class is being defined, yields a plain function object, but
in Cython it yields an unbound method<sup><font size="-2"><a href="#Footnote2">2</a></font></sup>. A consequence of this is that the
usual idiom for using the classmethod and staticmethod functions, e.g.
<blockquote> <pre>class Spam:</pre>
<pre>&nbsp; def method(cls):<br>&nbsp;&nbsp;&nbsp; ...</pre>
<pre>&nbsp; method = classmethod(method)</pre>
</blockquote>
will not work in Cython. This can be worked around by defining the function
<i>outside</i> the class, and then assigning the result of classmethod or
staticmethod inside the class, i.e.
<blockquote> <pre>def Spam_method(cls):<br>&nbsp; ...</pre>
<pre>class Spam:</pre>
<pre>&nbsp; method = classmethod(Spam_method)</pre>
</blockquote>
<h1> <hr width="100%"><font size="+0">Footnotes</font> <hr width="100%"></h1>
<a name="Footnote1"></a>1. A problem with const could arise if you have
something like
<blockquote> <pre>cdef extern from "grail.h":<br>&nbsp; char *nun</pre>
</blockquote>
where grail.h actually contains
<blockquote> <pre>extern const char *nun;</pre>
</blockquote>
and you do
<blockquote> <pre>cdef void languissement(char *s):<br>&nbsp; #something that doesn't change s</pre>
<pre>...</pre>
<pre>languissement(nun)</pre>
</blockquote>
which will cause the C compiler to complain. You can work around it by
casting away the constness:
<blockquote> <pre>languissement(&lt;char *&gt;nun)</pre>
</blockquote>
<hr width="100%"><a name="Footnote2"></a>2. The reason for the different behaviour
of class scopes is that Cython-defined Python functions are PyCFunction objects,
not PyFunction objects, and are not recognised by the machinery that creates
a bound or unbound method when a function is extracted from a class. To get
around this, Cython wraps each method in an unbound method object itself before
storing it in the class's dictionary. <br>
&nbsp; <br>
<br>
</body></html>
\ No newline at end of file
Doc/primes.c deleted 100644 → 0
View file @ 5a1c4958
#include "Python.h"
static PyObject *__Pyx_UnpackItem(PyObject *, int);
static int __Pyx_EndUnpack(PyObject *, int);
static int __Pyx_PrintItem(PyObject *);
static int __Pyx_PrintNewline(void);
static void __Pyx_ReRaise(void);
static void __Pyx_RaiseWithTraceback(PyObject *, PyObject *, PyObject *);
static PyObject *__Pyx_Import(PyObject *name, PyObject *from_list);
static PyObject *__Pyx_GetExcValue(void);
static PyObject *__Pyx_GetName(PyObject *dict, char *name);
static PyObject *__pyx_m;
static PyObject *__pyx_d;
static PyObject *__pyx_b;
PyObject *__pyx_f_primes(PyObject *__pyx_self, PyObject *__pyx_args); /*proto*/
PyObject *__pyx_f_primes(PyObject *__pyx_self, PyObject *__pyx_args) {
int __pyx_v_kmax;
int __pyx_v_n;
int __pyx_v_k;
int __pyx_v_i;
int (__pyx_v_p[1000]);
PyObject *__pyx_v_result;
PyObject *__pyx_r;
PyObject *__pyx_1 = 0;
int __pyx_2;
int __pyx_3;
int __pyx_4;
PyObject *__pyx_5 = 0;
PyObject *__pyx_6 = 0;
if (!PyArg_ParseTuple(__pyx_args, "i", &__pyx_v_kmax)) return 0;
__pyx_v_result = Py_None; Py_INCREF(__pyx_v_result);
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":2 */
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":3 */
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":4 */
__pyx_1 = PyList_New(0); if (!__pyx_1) goto __pyx_L1;
Py_DECREF(__pyx_v_result);
__pyx_v_result = __pyx_1;
__pyx_1 = 0;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":5 */
__pyx_2 = (__pyx_v_kmax > 1000);
if (__pyx_2) {
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":6 */
__pyx_v_kmax = 1000;
goto __pyx_L2;
}
__pyx_L2:;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":7 */
__pyx_v_k = 0;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":8 */
__pyx_v_n = 2;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":9 */
while (1) {
__pyx_L3:;
__pyx_2 = (__pyx_v_k < __pyx_v_kmax);
if (!__pyx_2) break;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":10 */
__pyx_v_i = 0;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":11 */
while (1) {
__pyx_L5:;
if (__pyx_3 = (__pyx_v_i < __pyx_v_k)) {
__pyx_3 = ((__pyx_v_n % (__pyx_v_p[__pyx_v_i])) != 0);
}
if (!__pyx_3) break;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":12 */
__pyx_v_i = (__pyx_v_i + 1);
}
__pyx_L6:;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":13 */
__pyx_4 = (__pyx_v_i == __pyx_v_k);
if (__pyx_4) {
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":14 */
(__pyx_v_p[__pyx_v_k]) = __pyx_v_n;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":15 */
__pyx_v_k = (__pyx_v_k + 1);
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":16 */
__pyx_1 = PyObject_GetAttrString(__pyx_v_result, "append"); if (!__pyx_1) goto __pyx_L1;
__pyx_5 = PyInt_FromLong(__pyx_v_n); if (!__pyx_5) goto __pyx_L1;
__pyx_6 = PyTuple_New(1); if (!__pyx_6) goto __pyx_L1;
PyTuple_SET_ITEM(__pyx_6, 0, __pyx_5);
__pyx_5 = 0;
__pyx_5 = PyObject_CallObject(__pyx_1, __pyx_6); if (!__pyx_5) goto __pyx_L1;
Py_DECREF(__pyx_6); __pyx_6 = 0;
Py_DECREF(__pyx_5); __pyx_5 = 0;
goto __pyx_L7;
}
__pyx_L7:;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":17 */
__pyx_v_n = (__pyx_v_n + 1);
}
__pyx_L4:;
/* "ProjectsA:Python:Pyrex:Demos:primes.pyx":18 */
Py_INCREF(__pyx_v_result);
__pyx_r = __pyx_v_result;
goto __pyx_L0;
__pyx_r = Py_None; Py_INCREF(__pyx_r);
goto __pyx_L0;
__pyx_L1:;
Py_XDECREF(__pyx_1);
Py_XDECREF(__pyx_5);
Py_XDECREF(__pyx_6);
__pyx_r = 0;
__pyx_L0:;
Py_DECREF(__pyx_v_result);
return __pyx_r;
}
static struct PyMethodDef __pyx_methods[] = {
{"primes", (PyCFunction)__pyx_f_primes, METH_VARARGS, 0},
{0, 0, 0, 0}
};
void initprimes(void); /*proto*/
void initprimes(void) {
__pyx_m = Py_InitModule4("primes", __pyx_methods, 0, 0, PYTHON_API_VERSION);
__pyx_d = PyModule_GetDict(__pyx_m);
__pyx_b = PyImport_AddModule("__builtin__");
PyDict_SetItemString(__pyx_d, "__builtins__", __pyx_b);
}
/* Runtime support code */
Doc/sharing.html deleted 100644 → 0
View file @ 5a1c4958
<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Sharing Declarations Between Cython Modules</title></head>
<body>
<h1> <hr width="100%">Sharing Declarations Between Cython Modules
<hr width="100%"></h1>
This section describes a new set of facilities introduced in Cython 0.8
for making C declarations and extension types in one Cython module available
for use in another Cython module. These facilities are closely modelled on
the Python import mechanism, and can be thought of as a compile-time version
of it.
<h2> Contents</h2>
<ul>
<li> <a href="#DefAndImpFiles">Definition and Implementation files</a></li>
<ul>
<li> <a href="#WhatDefFileContains">What a Definition File contains</a></li>
<li> <a href="#WhatImpFileContains">What an Implementation File contains</a></li>
</ul>
<li> <a href="#CImportStatement">The <tt>cimport</tt> statement</a></li>
<ul>
<li> <a href="#SearchPaths">Search paths for definition files</a></li>
<li> <a href="#ResolvingNamingConflicts">Using <tt>cimport</tt> to resolve
naming conflicts</a></li>
</ul>
<li> <a href="#SharingExtensionTypes">Sharing extension types</a></li>
</ul>
<h2> <a name="DefAndImpFiles"></a>Definition and Implementation files</h2>
A Cython module can be split into two parts: a <i>definition file</i> with
a <tt>.pxd</tt> suffix, containing C declarations that are to be available
to other Cython modules, and an <i>implementation file</i> with a <tt>.pyx</tt>
suffix, containing everything else. When a module wants to use something
declared in another module's definition file, it imports it using the <a href="#CImportStatement"><b>cimport</b> statement</a>.
<h3> <a name="WhatDefFileContains"></a>What a Definition File contains</h3>
A definition file can contain:
<ul>
<li> Any kind of C type declaration.</li>
<li> <b>extern</b> C function or variable declarations.</li>
<li> The definition part of an extension type (<a href="#SharingExtensionTypes">see below</a>).</li>
</ul>
It cannot currently contain any non-extern C function or variable declarations
(although this may be possible in a future version).
<p>It cannot contain the implementations of any C or Python functions, or
any Python class definitions, or any executable statements. </p>
<blockquote>NOTE: You don't need to (and shouldn't) declare anything in a
declaration file <b>public</b> in order to make it available to other Cython
modules; its mere presence in a definition file does that. You only need a
public declaration if you want to make something available to external C code.</blockquote>
<h3> <a name="WhatImpFileContains"></a>What an Implementation File contains</h3>
An implementation file can contain any kind of Cython statement, although
there are some restrictions on the implementation part of an extension type
if the corresponding definition file also defines that type (see below).
<h2> <a name="CImportStatement"></a>The <tt>cimport</tt> statement</h2>
The <b>cimport</b> statement is used in a definition or implementation
file to gain access to names declared in another definition file. Its syntax
exactly parallels that of the normal Python import statement:
<blockquote><tt>cimport </tt><i>module</i><tt> [, </tt><i>module</i><tt>...]</tt></blockquote>
<blockquote><tt>from </tt><i>module</i><tt> cimport </tt><i>name</i><tt>
[as </tt><i>name</i><tt>] [, </tt><i>name</i><tt> [as </tt><i>name</i><tt>]
...]</tt></blockquote>
Here is an example. The file on the left is a definition file which exports
a C data type. The file on the right is an implementation file which imports
and uses it. <br>
&nbsp; <table cellpadding="5" cols="2" width="100%">
<tbody>
<tr>
<td bgcolor="#ffcc00" width="40%"><b><tt>dishes.pxd</tt></b></td>
<td bgcolor="#5dbaca"><b><tt>restaurant.pyx</tt></b></td>
</tr>
<tr align="left" valign="top">
<td bgcolor="#ffcc18" width="40%"><tt>cdef enum otherstuff:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; sausage, eggs, lettuce</tt> <p><tt>cdef struct spamdish:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; int oz_of_spam</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; otherstuff filler</tt></p>
</td>
<td bgcolor="#5dbaca"><tt>cimport dishes</tt> <br>
<tt>from dishes cimport spamdish</tt> <p><tt>cdef void prepare(spamdish *d):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; d.oz_of_spam = 42</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; d.filler = dishes.sausage</tt> </p>
<p><tt>def serve():</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; spamdish d</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; prepare(&amp;d)</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; print "%d oz spam, filler no. %d" % \</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (d-&gt;oz_of_spam,
d-&gt;filler)</tt></p>
</td>
</tr>
</tbody> </table>
<p>It is important to understand that the <b>cimport</b> statement can <i>only</i>
be used to import C data types, external C functions and variables, and extension
types. It cannot be used to import any Python objects, and (with one exception)
it doesn't imply any Python import at run time. If you want to refer to any
Python names from a module that you have cimported, you will have to include
a regular <b>import</b> statement for it as well. </p>
<p>The exception is that when you use <b>cimport</b> to import an extension
type, its type object is imported at run time and made available by the
name under which you imported it. Using <b>cimport</b> to import extension
types is covered in more detail <a href="#SharingExtensionTypes">below</a>.
</p>
<h3> <a name="SearchPaths"></a>Search paths for definition files</h3>
When you <b>cimport</b> a module called <tt>modulename</tt>, the Cython
compiler searches for a file called <tt>modulename.pxd</tt> along the search
path for include files, as specified by <b>-I</b> command line options.
<p>Also, whenever you compile a file <tt>modulename.pyx</tt>, the corresponding
definition file <tt>modulename.pxd</tt> is first searched for along the
same path, and if found, it is processed before processing the <tt>.pyx</tt>
file. </p>
<h3> <a name="ResolvingNamingConflicts"></a>Using cimport to resolve naming
conflicts</h3>
The cimport mechanism provides a clean and simple way to solve the problem
of wrapping external C functions with Python functions of the same name.
All you need to do is put the extern C declarations into a .pxd file for
an imaginary module, and cimport that module. You can then refer to the C
functions by qualifying them with the name of the module. Here's an example:
<br>
&nbsp; <table cellpadding="5" cols="2" width="100%">
<tbody>
<tr>
<td bgcolor="#ffcc00" width="50%"><b><tt>c_lunch.pxd</tt></b></td>
<td bgcolor="#5dbaca"><b><tt>lunch.pyx</tt></b></td>
</tr>
<tr align="left" valign="top">
<td bgcolor="#ffcc18" width="50%"><tt>cdef extern from "lunch.h":</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp; void eject_tomato(float)</tt></td>
<td bgcolor="#5dbaca"><tt>cimport c_lunch</tt> <p><tt>def eject_tomato(float speed):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; c_lunch.eject_tomato(speed)</tt></p>
</td>
</tr>
</tbody> </table>
<p>You don't need any <tt>c_lunch.pyx</tt> file, because the only things
defined in <tt>c_lunch.pxd</tt> are extern C entities. There won't be any
actual <tt>c_lunch</tt> module at run time, but that doesn't matter -- <tt>c_lunch</tt>
has done its job of providing an additional namespace at compile time. </p>
<h2> <a name="SharingExtensionTypes"></a>Sharing Extension Types</h2>
An extension type declaration can also be split into two parts, one in
a definition file and the other in the corresponding implementation file.
<br>
<br>
The definition part of the extension type can only declare C attributes
and C methods, not Python methods, and it must declare <i>all</i> of that
type's C attributes and C methods.<br>
<br>
The implementation part must implement all of the C methods declared in
the definition part, and may not add any further C attributes. It may also
define Python methods.
<p>Here is an example of a module which defines and exports an extension
type, and another module which uses it. <br>
&nbsp; <table cellpadding="5" cols="2" width="100%">
<tbody>
<tr>
<td bgcolor="#ffcc18" width="30%"><b><tt>Shrubbing.pxd</tt></b></td>
<td bgcolor="#5dbaca" width="50%"><b><tt>Shrubbing.pyx</tt></b></td>
</tr>
<tr align="left" valign="top">
<td bgcolor="#ffcc18" width="30%"><tt>cdef class Shrubbery:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; cdef int width</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; cdef int length</tt></td>
<td bgcolor="#5dbaca" width="50%"><tt>cdef class Shrubbery:</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; def __new__(self, int w, int l):</tt> <br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.width = w</tt>
<br>
<tt>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; self.length = l</tt>
<p><tt>def standard_shrubbery():</tt> <br>
<tt>&nbsp;&nbsp;&nbsp; return Shrubbery(3, 7)</tt></p>
</td>
</tr>
<tr>
<td colspan="2" bgcolor="#8cbc1c" width="30%"><b><tt>Landscaping.pyx</tt></b></td>
</tr>
<tr>
<td colspan="2" bgcolor="#99cc00" width="30%"><tt>cimport Shrubbing</tt>
<br>
<tt>import Shrubbing</tt> <p><tt>cdef Shrubbing.Shrubbery sh</tt> <br>
<tt>sh = Shrubbing.standard_shrubbery()</tt> <br>
<tt>print "Shrubbery size is %d x %d" % (sh.width, sh.height)</tt>
<br>
&nbsp;</p>
</td>
</tr>
</tbody> </table>
</p>
<p>Some things to note about this example: </p>
<ul>
<li> There is a <tt>cdef class Shrubbery</tt> declaration in both Shrubbing.pxd
and Shrubbing.pyx. When the Shrubbing module is compiled, these two declarations
are combined into one.</li>
&nbsp; <li> In Landscaping.pyx, the <tt>cimport Shrubbing</tt> declaration
allows us to refer to the Shrubbery type as <tt>Shrubbing.Shrubbery</tt>.
But it doesn't bind the name <tt>Shrubbery</tt> in Landscaping's module namespace
at run time, so to access <tt>Shrubbery.standard_shrubbery</tt> we also
need to <tt>import Shrubbing</tt>.</li>
</ul>
<hr width="100%">Back to the <a href="overview.html">Language Overview</a>
<br>
<br>
</body></html>
Doc/special_methods.html deleted 100644 → 0
View file @ 5a1c4958
<!DOCTYPE doctype PUBLIC "-//w3c//dtd html 4.0 transitional//en">
<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.61 (Macintosh; I; PPC) [Netscape]"><title>Special Methods of Extenstion Types</title></head>
<body>
<h1> <hr width="100%">Special Methods of Extension Types
<hr width="100%"></h1>
This page describes the special methods currently supported by Cython extension
types. A complete list of all the special methods appears in the table at
the bottom. Some of these methods behave differently from their Python counterparts
or have no direct Python counterparts, and require special mention.
<p><span style="font-weight: bold;">Note:</span><i> Everything said on this page applies only to </i><span style="font-weight: bold;">extension</span><i style="font-weight: bold;">
</i><span style="font-weight: bold;">types</span><i>, defined with the </i><span style="font-weight: bold; font-family: monospace;">cdef class</span><i> statement. It doesn't apply&nbsp;
to classes defined with the Python </i><span style="font-family: monospace;">class</span><i><span style="font-family: monospace;"> </span>statement, where the normal
Python rules apply.</i> </p>
<h2><small>Declaration</small></h2>Special methods of extension types must be declared with <span style="font-family: monospace; font-weight: bold;">def</span>, <span style="font-style: italic;">not</span> <span style="font-family: monospace;">cdef</span>.<br>
<h2><font size="+1">Docstrings</font></h2>
Currently, docstrings are not fully supported in special methods of extension
types. You can place a docstring in the source to serve as a comment, but
it won't show up in the corresponding <span style="font-family: monospace;">__doc__</span> attribute at run time. (This
is a Python limitation -- there's nowhere in the PyTypeObject data structure
to put such docstrings.)
<h2> <font size="+1">Initialisation methods: <tt>__new__</tt> and <tt>__init__</tt></font></h2>
There are two methods concerned with initialising the object<tt>.</tt>
<p>The <b><tt>__new__</tt></b> method is where you should perform basic C-level
initialisation of the object, including allocation of any C data structures
that your object will own. You need to be careful what you do in the __new__
method, because the object may not yet be a valid Python object when it is
called. Therefore, you must not invoke any Python operations which might touch
the object; in particular, do not try to call any of its methods. </p>
<p>Unlike the corresponding method in Python, your <tt>__new__</tt> method
is <i>not</i> responsible for <i>creating</i> the object. By the time your
<tt>__new__</tt> method is called, memory has been allocated for the object
and any C attributes it has have been initialised to 0 or null. (Any Python
attributes have also been initialised to <tt>None</tt>, but you probably shouldn't
rely on that.) Your <tt>__new__</tt> method is guaranteed to be called exactly
once.<br>
<br>
If your extension type has a base type, the <tt>__new__</tt> method of the
base type is automatically called <i>before</i> your <tt>__new__</tt> method
is called; you cannot explicitly call the inherited <tt>__new__</tt> method.
If you need to pass a modified argument list to the base type, you will have
to do the relevant part of the initialisation in the <tt>__init__</tt> method
instead (where the normal rules for calling inherited methods apply).<br>
</p>
<p>Note that the first parameter of the <tt>__new__</tt> method is the object
to be initialised, not the class of the object as it is in Python. </p>
<p>Any initialisation which cannot safely be done in the <tt>__new__</tt>
method should be done in the <b><tt>__init__</tt></b> method. By the time
<tt>__init__</tt> is called, the object is a fully valid Python object and
all operations are safe. Under some circumstances it is possible for <tt>__init__</tt>
to be called more than once or not to be called at all, so your other methods
should be designed to be robust in such situations. </p>
<p>Keep in mind that any arguments passed to the constructor will be passed
to the <tt>__new__</tt> method as well as the <tt>__init__</tt> method.
If you anticipate subclassing your extension type in Python, you may find
it useful to give the <tt>__new__</tt> method * and ** arguments so that
it can accept and ignore extra arguments. Otherwise, any Python subclass
which has an <tt>__init__</tt> with a different signature will have to override
<tt>__new__</tt> as well as <tt>__init__</tt>, which the writer of a Python
class wouldn't expect to have to do. </p>
<h2> <font size="+1">Finalization method: <tt>__dealloc__</tt><tt></tt></font></h2>
The counterpart to the <tt>__new__</tt> method is the <b><tt>__dealloc__</tt></b>
method, which should perform the inverse of the <tt>__new__</tt> method.
Any C data structures that you allocated in your <tt>__new__</tt> method
should be freed in your <tt>__dealloc__</tt> method.
<p>You need to be careful what you do in a <tt>__dealloc__</tt> method. By
the time your <tt>__dealloc__</tt> method is called, the object may already
have been partially destroyed and may not be in a valid state as far as Python
is concerned, so you should avoid invoking any Python operations which might
touch the object. In particular, don't call any other methods of the object
or do anything which might cause the object to be resurrected. It's best if
you stick to just deallocating C data. </p>
<p>You don't need to worry about deallocating Python attributes of your object,
because that will be done for you by Cython after your <tt>__dealloc__</tt>
method returns.<br>
<br>
<b>Note:</b> There is no <tt>__del__</tt> method for extension types. (Earlier
versions of the Cython documentation stated that there was, but this turned
out to be incorrect.)<br>
</p>
<h2><font size="+1">Arithmetic methods</font></h2>
Arithmetic operator methods, such as <tt>__add__</tt>, behave differently
from their Python counterparts. There are no separate "reversed" versions
of these methods (<tt>__radd__</tt>, etc.) Instead, if the first operand
cannot perform the operation, the <i>same</i> method of the second operand
is called, with the operands in the <i>same order</i>.
<p>This means that you can't rely on the first parameter of these methods
being "self", and you should test the types of both operands before deciding
what to do. If you can't handle the combination of types you've been given,
you should return <tt>NotImplemented</tt>. </p>
<p>This also applies to the in-place arithmetic method <tt>__ipow__</tt>.
It doesn't apply to any of the <i>other</i> in-place methods (<tt>__iadd__</tt>,
etc.) which always take self as the first argument. </p>
<h2> <font size="+1">Rich comparisons</font></h2>
There are no separate methods for the individual rich comparison operations
(<tt>__eq__</tt>, <tt>__le__</tt>, etc.) Instead there is a single method
<tt>__richcmp__</tt> which takes an integer indicating which operation is
to be performed, as follows:
<ul>
<ul>
&nbsp; <table nosave="" border="0" cellpadding="5" cellspacing="0">
<tbody>
<tr nosave="">
<td nosave="" bgcolor="#ffcc33" width="30"> <div align="right">&lt;</div>
</td>
<td nosave="" bgcolor="#66ffff" width="30">0</td>
<td><br>
</td>
<td nosave="" bgcolor="#ffcc33" width="30"> <div align="right">==</div>
</td>
<td nosave="" bgcolor="#66ffff" width="30">2</td>
<td><br>
</td>
<td nosave="" bgcolor="#ffcc33" width="30"> <div align="right">&gt;</div>
</td>
<td nosave="" bgcolor="#66ffff" width="30">4</td>
</tr>
<tr nosave="">
<td nosave="" bgcolor="#ffcc33"> <div align="right">&lt;=</div>
</td>
<td nosave="" bgcolor="#66ffff">1</td>
<td><br>
</td>
<td nosave="" bgcolor="#ffcc33"> <div align="right">!=</div>
</td>
<td nosave="" bgcolor="#66ffff">3</td>
<td><br>
</td>
<td nosave="" bgcolor="#ffcc33"> <div align="right">&gt;=</div>
</td>
<td nosave="" bgcolor="#66ffff">5</td>
</tr>
</tbody> </table>
</ul>
</ul>
<h2> <font size="+1">The __next__ method</font></h2>
Extension types wishing to implement the iterator interface should define
a method called <b><tt>__next__</tt></b>, <i>not</i> <tt>next</tt>. The Python
system will automatically supply a <tt>next</tt> method which calls your
<span style="font-family: monospace;">__next__</span>. <b>Do NOT explicitly give your type a <tt>next</tt> method</b>,
or bad things could happen (see note 3).
<h2> <font size="+1">Special Method Table</font></h2>
This table lists all of the special methods together with their parameter
and return types. A parameter named <b>self</b> is of the type the method
belongs to. Other untyped parameters are generic Python objects.
<p>You don't have to declare your method as taking these parameter types.
If you declare different types, conversions will be performed as necessary.
<br>
&nbsp; <table nosave="" bgcolor="#ccffff" border="1" cellpadding="5" cellspacing="0">
<tbody>
<tr nosave="" bgcolor="#ffcc33">
<td nosave=""><b>Name</b></td>
<td><b>Parameters</b></td>
<td><b>Return type</b></td>
<td><b>Description</b></td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>General</b></td>
</tr>
<tr>
<td><tt>__new__</tt></td>
<td>self, ...</td>
<td>&nbsp;</td>
<td>Basic initialisation (no direct Python equivalent)</td>
</tr>
<tr>
<td><tt>__init__</tt></td>
<td>self, ...</td>
<td>&nbsp;</td>
<td>Further initialisation</td>
</tr>
<tr>
<td><tt>__dealloc__</tt></td>
<td>self</td>
<td>&nbsp;</td>
<td>Basic deallocation (no direct Python equivalent)</td>
</tr>
<tr>
<td><tt>__cmp__</tt></td>
<td>x, y</td>
<td>int</td>
<td>3-way comparison</td>
</tr>
<tr>
<td><tt>__richcmp__</tt></td>
<td>x, y, int op</td>
<td>object</td>
<td>Rich comparison (no direct Python equivalent)</td>
</tr>
<tr>
<td><tt>__str__</tt></td>
<td>self</td>
<td>object</td>
<td>str(self)</td>
</tr>
<tr>
<td><tt>__repr__</tt></td>
<td>self</td>
<td>object</td>
<td>repr(self)</td>
</tr>
<tr nosave="">
<td nosave=""><tt>__hash__</tt></td>
<td>self</td>
<td>int</td>
<td>Hash function</td>
</tr>
<tr>
<td><tt>__call__</tt></td>
<td>self, ...</td>
<td>object</td>
<td>self(...)</td>
</tr>
<tr>
<td><tt>__iter__</tt></td>
<td>self</td>
<td>object</td>
<td>Return iterator for sequence</td>
</tr>
<tr>
<td><tt>__getattr__</tt></td>
<td>self, name</td>
<td>object</td>
<td>Get attribute</td>
</tr>
<td><tt>__getattribute__</tt></td>
<td>self, name</td>
<td>object</td>
<td>Get attribute, unconditionally</td>
</tr>
<tr>
<td><tt>__setattr__</tt></td>
<td>self, name, val</td>
<td>&nbsp;</td>
<td>Set attribute</td>
</tr>
<tr>
<td><tt>__delattr__</tt></td>
<td>self, name</td>
<td>&nbsp;</td>
<td>Delete attribute</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>Arithmetic operators</b></td>
</tr>
<tr>
<td><tt>__add__</tt></td>
<td>x, y</td>
<td>object</td>
<td>binary + operator</td>
</tr>
<tr>
<td><tt>__sub__</tt></td>
<td>x, y</td>
<td>object</td>
<td>binary - operator</td>
</tr>
<tr>
<td><tt>__mul__</tt></td>
<td>x, y</td>
<td>object</td>
<td>* operator</td>
</tr>
<tr>
<td><tt>__div__</tt></td>
<td>x, y</td>
<td>object</td>
<td>/&nbsp; operator for old-style division</td>
</tr>
<tr>
<td><tt>__floordiv__</tt></td>
<td>x, y</td>
<td>object</td>
<td>//&nbsp; operator</td>
</tr>
<tr>
<td><tt>__truediv__</tt></td>
<td>x, y</td>
<td>object</td>
<td>/&nbsp; operator for new-style division</td>
</tr>
<tr>
<td><tt>__mod__</tt></td>
<td>x, y</td>
<td>object</td>
<td>% operator</td>
</tr>
<tr>
<td><tt>__divmod__</tt></td>
<td>x, y</td>
<td>object</td>
<td>combined div and mod</td>
</tr>
<tr>
<td><tt>__pow__</tt></td>
<td>x, y, z</td>
<td>object</td>
<td>** operator or pow(x, y, z)</td>
</tr>
<tr>
<td><tt>__neg__</tt></td>
<td>self</td>
<td>object</td>
<td>unary - operator</td>
</tr>
<tr>
<td><tt>__pos__</tt></td>
<td>self</td>
<td>object</td>
<td>unary + operator</td>
</tr>
<tr>
<td><tt>__abs__</tt></td>
<td>self</td>
<td>object</td>
<td>absolute value</td>
</tr>
<tr>
<td><tt>__nonzero__</tt></td>
<td>self</td>
<td>int</td>
<td>convert to boolean</td>
</tr>
<tr>
<td><tt>__invert__</tt></td>
<td>self</td>
<td>object</td>
<td>~ operator</td>
</tr>
<tr>
<td><tt>__lshift__</tt></td>
<td>x, y</td>
<td>object</td>
<td>&lt;&lt; operator</td>
</tr>
<tr>
<td><tt>__rshift__</tt></td>
<td>x, y</td>
<td>object</td>
<td>&gt;&gt; operator</td>
</tr>
<tr>
<td><tt>__and__</tt></td>
<td>x, y</td>
<td>object</td>
<td>&amp; operator</td>
</tr>
<tr>
<td><tt>__or__</tt></td>
<td>x, y</td>
<td>object</td>
<td>| operator</td>
</tr>
<tr>
<td><tt>__xor__</tt></td>
<td>x, y</td>
<td>object</td>
<td>^ operator</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>Numeric conversions</b></td>
</tr>
<tr>
<td><tt>__int__</tt></td>
<td>self</td>
<td>object</td>
<td>Convert to integer</td>
</tr>
<tr>
<td><tt>__long__</tt></td>
<td>self</td>
<td>object</td>
<td>Convert to long integer</td>
</tr>
<tr>
<td><tt>__float__</tt></td>
<td>self</td>
<td>object</td>
<td>Convert to float</td>
</tr>
<tr>
<td><tt>__oct__</tt></td>
<td>self</td>
<td>object</td>
<td>Convert to octal</td>
</tr>
<tr>
<td><tt>__hex__</tt></td>
<td>self</td>
<td>object</td>
<td>Convert to hexadecimal</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>In-place arithmetic operators</b></td>
</tr>
<tr>
<td><tt>__iadd__</tt></td>
<td>self, x</td>
<td>object</td>
<td>+= operator</td>
</tr>
<tr>
<td><tt>__isub__</tt></td>
<td>self, x</td>
<td>object</td>
<td>-= operator</td>
</tr>
<tr>
<td><tt>__imul__</tt></td>
<td>self, x</td>
<td>object</td>
<td>*= operator</td>
</tr>
<tr>
<td><tt>__idiv__</tt></td>
<td>self, x</td>
<td>object</td>
<td>/= operator for old-style division</td>
</tr>
<tr>
<td><tt>__ifloordiv__</tt></td>
<td>self, x</td>
<td>object</td>
<td>//= operator</td>
</tr>
<tr>
<td><tt>__itruediv__</tt></td>
<td>self, x</td>
<td>object</td>
<td>/= operator for new-style division</td>
</tr>
<tr>
<td><tt>__imod__</tt></td>
<td>self, x</td>
<td>object</td>
<td>%= operator</td>
</tr>
<tr>
<td><tt>__ipow__</tt></td>
<td>x, y, z</td>
<td>object</td>
<td>**= operator</td>
</tr>
<tr>
<td><tt>__ilshift__</tt></td>
<td>self, x</td>
<td>object</td>
<td>&lt;&lt;= operator</td>
</tr>
<tr>
<td><tt>__irshift__</tt></td>
<td>self, x</td>
<td>object</td>
<td>&gt;&gt;= operator</td>
</tr>
<tr>
<td><tt>__iand__</tt></td>
<td>self, x</td>
<td>object</td>
<td>&amp;= operator</td>
</tr>
<tr>
<td><tt>__ior__</tt></td>
<td>self, x</td>
<td>object</td>
<td>|= operator</td>
</tr>
<tr>
<td><tt>__ixor__</tt></td>
<td>self, x</td>
<td>object</td>
<td>^= operator</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>Sequences and mappings</b></td>
</tr>
<tr>
<td><tt>__len__</tt></td>
<td>self</td>
<td>int</td>
<td>len(self)</td>
</tr>
<tr>
<td><tt>__getitem__</tt></td>
<td>self, x</td>
<td>object</td>
<td>self[x]</td>
</tr>
<tr>
<td><tt>__setitem__</tt></td>
<td>self, x, y</td>
<td>&nbsp;</td>
<td>self[x] = y</td>
</tr>
<tr>
<td><tt>__delitem__</tt></td>
<td>self, x</td>
<td>&nbsp;</td>
<td>del self[x]</td>
</tr>
<tr>
<td><tt>__getslice__</tt></td>
<td>self, int i, int j</td>
<td>object</td>
<td>self[i:j]</td>
</tr>
<tr>
<td><tt>__setslice__</tt></td>
<td>self, int i, int j, x</td>
<td>&nbsp;</td>
<td>self[i:j] = x</td>
</tr>
<tr>
<td><tt>__delslice__</tt></td>
<td>self, int i, int j</td>
<td>&nbsp;</td>
<td>del self[i:j]</td>
</tr>
<tr>
<td><tt>__contains__</tt></td>
<td>self, x</td>
<td>int</td>
<td>x in self</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>Iterators</b></td>
</tr>
<tr>
<td><tt>__next__</tt></td>
<td>self</td>
<td>object</td>
<td>Get next item (called <tt>next</tt> in Python)</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>Buffer interface</b>&nbsp; (no Python equivalents
- see note 1)</td>
</tr>
<tr>
<td><tt>__getreadbuffer__</tt></td>
<td>self, int i, void **p</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr>
<td><tt>__getwritebuffer__</tt></td>
<td>self, int i, void **p</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr>
<td><tt>__getsegcount__</tt></td>
<td>self, int *p</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr>
<td><tt>__getcharbuffer__</tt></td>
<td>self, int i, char **p</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
</tr>
<tr nosave="" bgcolor="#66ffff">
<td colspan="4" nosave=""><b>Descriptor objects</b>&nbsp; (no Python equivalents
- see note 2)</td>
</tr>
<tr>
<td><tt>__get__</tt></td>
<td>self, instance, class</td>
<td>object</td>
<td>Get value of attribute</td>
</tr>
<tr>
<td><tt>__set__</tt></td>
<td>self, instance, value</td>
<td>&nbsp;</td>
<td>Set value of attribute</td>
</tr>
<tr>
<td style="font-family: monospace;">__delete__</td>
<td>self, instance</td>
<td>&nbsp;</td>
<td>Delete attribute</td>
</tr>
</tbody> </table>
</p>
<p>Note 1: The buffer interface is intended for use by C code and is not
directly accessible from Python. It is described in the <a href="http://www.python.org/doc/current/api/api.html">Python/C API Reference
Manual</a> under sections <a href="http://www.python.org/doc/current/api/abstract-buffer.html">6.6</a>
and <a href="http://www.python.org/doc/current/api/buffer-structs.html">10.6</a>.
</p>
<p>Note 2: Descriptor objects are part of the support mechanism for new-style
Python classes. See the <a href="http://www.python.org/doc/2.2.1/whatsnew/sect-rellinks.html#SECTION000320000000000000000">discussion
of descriptors in the Python documentation</a>. See also <a href="http://www.python.org/peps/pep-0252.html">PEP 252, "Making Types Look
More Like Classes"</a>, and <a href="http://www.python.org/peps/pep-0253.html">PEP 253, "Subtyping Built-In
Types"</a>. </p>
<p>Note 3: If your type defines a <tt>__new__</tt> method, any method called
<tt>new</tt> that you define will be overwritten with the system-supplied
<tt>new</tt> at module import time. </p>
<br>
<br>
</body></html>
\ No newline at end of file
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7