<!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 <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>
<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> ...</pre>
<pre>cdef int eggs(unsigned long l, float f):<br> ...</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> ...</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> ...</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> ...</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> int age<br> float volume</pre>
<pre>cdef union Food:<br> char *spam<br> float *eggs</pre>
<pre>cdef enum CheeseType:<br> cheddar, edam, <br> camembert</pre>
<pre>cdef enum CheeseState:<br> hard = 1<br> soft = 2<br> 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> 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> x = True<br>except NameError:<br> 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>-></tt></b> operator in Cython. Instead of <tt>p->x</tt>,
use <tt>p.x</tt></li>
<li> There is no <b><tt>*</tt></b> operator in Cython. Instead of
<tt>*p</tt>, use <tt>p[0]</tt></li>
<li> There is an <b><tt>&</tt></b> operator, with the same semantics
as in C.</li>
<li> The null C pointer is called <b><tt>NULL</tt></b>, not 0 (and
<tt>NULL</tt> is a reserved word).</li>
<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><type>value</tt></b> , for example:</li>
<ul>
<pre>cdef char *p, float *q<br>p = <char*>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> ...</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 <= i < n:</tt> <br>
<tt> ...</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><</tt>, <tt><=</tt>} then it is upwards;
if they are both from the set {<tt>></tt>, <tt>>=</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> ...</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> ...</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> ...</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>
<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> 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> int spam_counter</pre>
<pre> 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>
<li> It prevents Cython from generating any C code for the declarations
found in the associated block.<br>
</li>
<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>
<li> Leave out any platform-specific extensions to C declarations
such as <b>__declspec()</b>.<br>
</li>
<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> cdef extern from "foo.h":<br>
struct spam:<br>
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>
<li> If the header file uses macros to define constants, translate
them into a dummy <b>enum</b> declaration.<br>
</li>
<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> 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> ...</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>
<table cellpadding="5">
<tbody>
<tr bgcolor="#8cbc1c" valign="top">
<td bgcolor="#8cbc1c"> </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> ...</tt> <br>
<tt>};</tt></td>
<td bgcolor="#66cccc"><tt>cdef struct Foo:</tt> <br>
<tt> ...</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> ...</tt> <br>
<tt>} Foo;</tt></td>
<td bgcolor="#66cccc" valign="top"><tt>ctypedef struct Foo:</tt> <br>
<tt> ...</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> ...</tt> <br>
<tt>} Foo;</tt></td>
<td bgcolor="#66cccc" nowrap="nowrap" valign="top"><tt>cdef struct foo:</tt> <br>
<tt> ...</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> ...</tt></td>
</tr>
<tr bgcolor="#8cbc1c" valign="top">
<td>4</td>
<td bgcolor="#ff9900" nowrap="nowrap"><tt>typedef struct Foo {</tt> <br>
<tt> ...</tt> <br>
<tt>} Foo;</tt></td>
<td bgcolor="#66cccc" valign="top"><tt>cdef struct Foo:</tt> <br>
<tt> ...</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> 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> 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> int i "eye"</pre>
<tt>cdef enum surprise "inquisition":</tt> <br>
<tt> first "alpha"</tt> <br>
<tt> 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> ...</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 <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>
<li> Class definitions can only appear at the top level of a module,
not inside a function.<br>
</li>
<li> The<tt> import *</tt> form of import is not allowed anywhere
(other forms of the import statement are fine, though).<br>
</li>
<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>
<li> In-place arithmetic operators (+=, etc) are not yet supported.<br>
</li>
<li> List comprehensions are not yet supported.<br>
</li>
<li> There is no support for Unicode.<br>
</li>
<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> def method(cls):<br> ...</pre>
<pre> 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> ...</pre>
<pre>class Spam:</pre>
<pre> 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> 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> #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(<char *>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>
<br>
<br>
</body></html>
distrib
>
Mandriva
>
current
>
i586
>
media
>
main-updates
>
by-pkgid
>
78c129e4a46a64625b53d61bd2b90fab
>
files
>
366
