Skip to content

C
cpython
Kaydet (Commit) 5049bcb1 authored tarafından Guido van Rossum's avatar Guido van Rossum
Dosyalara gözat
Seçenekler

another round (sigh :-( )

üst 6bb1adc7
Hide whitespace changes
Inline Side-by-side
Showing with 1716 additions and 1006 deletions
Doc/ext.tex
Dosyayı görüntüle @ 5049bcb1
\documentstyle[twoside,11pt,myformat]{report} \documentstyle[twoside,11pt,myformat]{report}
% XXX PM Modulator
\title{Extending and Embedding the Python Interpreter} \title{Extending and Embedding the Python Interpreter}
\input{boilerplate} \input{boilerplate}
...@@ -45,294 +47,333 @@ system supports this feature. ...@@ -45,294 +47,333 @@ system supports this feature.
It is quite easy to add non-standard built-in modules to Python, if It is quite easy to add non-standard built-in modules to Python, if
you know how to program in C. A built-in module known to the Python you know how to program in C. A built-in module known to the Python
programmer as \code{foo} is generally implemented by a file called programmer as \code{spam} is generally implemented by a file called
\file{foomodule.c}. All but the two most essential standard built-in \file{spammodule.c} (if the module name is very long, like
modules also adhere to this convention, and in fact some of them form \samp{spammify}, you can drop the \samp{module}, leaving a file name
excellent examples of how to create an extension. like \file{spammify.c}). The standard built-in modules also adhere to
this convention, and in fact some of them are excellent examples of
how to create an extension.
Extension modules can do two things that can't be done directly in Extension modules can do two things that can't be done directly in
Python: they can implement new data types (which are different from Python: they can implement new data types (which are different from
classes, by the way), and they can make system calls or call C library classes, by the way), and they can make system calls or call C library
functions. We'll see how both types of extension are implemented by functions.
examining the code for a Python curses interface.
Note: unless otherwise mentioned, all file references in this To support extensions, the Python API (Application Programmers
document are relative to the toplevel directory of the Python Interface) defines many functions, macros and variables that provide
distribution --- i.e. the directory that contains the \file{configure} access to almost every aspect of the Python run-time system.
script. Most of the Python API is imported by including the single header file
\code{"Python.h"}. All user-visible symbols defined by including this
file have a prefix of \samp{Py} or \samp{PY}, except those defined in
standard header files --- for convenience, and since they are needed by
the Python interpreter, \file{"Python.h"} includes a few standard
header files: \file{<stdio.h>}, \file{<string.h>}, \file{<errno.h>},
and \file{<stdlib.h>}. If the latter header file does not exist on
your system, it declares the functions \code{malloc()}, \code{free()}
and \code{realloc()} itself.
The compilation of an extension module depends on your system setup The compilation of an extension module depends on your system setup
and the intended use of the module; details are given in a later and the intended use of the module; details are given in a later
section. section.
Note: unless otherwise mentioned, all file references in this
document are relative to the Python toplevel directory
(the directory that contains the \file{configure} script).
\section{A Simple Example}
\section{A first look at the code} Let's create an extension module called \samp{spam}. Create a file
\samp{spammodule.c}. The first line of this file can be:
It is important not to be impressed by the size and complexity of \begin{verbatim}
the average extension module; much of this is straightforward #include "Python.h"
`boilerplate' code (starting right with the copyright notice)! \end{verbatim}
which pulls in the Python API (you can add a comment describing the
purpose of the module and a copyright notice if you like).
Let's skip the boilerplate and have a look at an interesting function Let's create a Python interface to the C library function
in \file{posixmodule.c} first: \code{system()}.\footnote{An interface for this function already
exists in the \code{posix} module --- it was chosen as a simple and
straightfoward example.} This function takes a zero-terminated
character string as argument and returns an integer. We will want
this function to be callable from Python as follows:
\begin{verbatim} \begin{verbatim}
static object * >>> import spam
posix_system(self, args) >>> status = spam.system("ls -l")
object *self; \end{verbatim}
object *args;
The next thing we add to our module file is the C function that will
be called when the Python expression \samp{spam.system(\var{string})}
is evaluated (well see shortly how it ends up being called):
\begin{verbatim}
static PyObject *
spam_system(self, args)
PyObject *self;
PyObject *args;
{ {
char *command; char *command;
int sts; int sts;
if (!getargs(args, "s", &command)) if (!PyArg_ParseTuple(args, "s", &command))
return NULL; return NULL;
sts = system(command); sts = system(command);
return mkvalue("i", sts); return Py_BuildValue("i", sts);
} }
\end{verbatim} \end{verbatim}
This is the prototypical top-level function in an extension module. There is a straightforward translation from the argument list in
It will be called (we'll see later how) when the Python program Python (here the single expression \code{"ls -l"}) to the arguments
executes statements like that are passed to the C function. The C function always has two
arguments, conventionally named \var{self} and \var{args}.
\begin{verbatim}
>>> import posix The \var{self} argument is only used when the C function implements a
>>> sts = posix.system('ls -l') builtin method --- this will be discussed later. In the example,
\end{verbatim} \var{self} will always be a \code{NULL} pointer, since we are defining
a function, not a method. (This is done so that the interpreter
There is a straightforward translation from the arguments to the call doesn't have to understand two different types of C functions.)
in Python (here the single expression \code{'ls -l'}) to the arguments that
are passed to the C function. The C function always has two The \var{args} argument will be a pointer to a Python tuple object
parameters, conventionally named \var{self} and \var{args}. The containing the arguments --- the length of the tuple will be the
\var{self} argument is used when the C function implements a builtin number of arguments. It is necessary to do full argument type
method---this will be discussed later. checking in each call, since otherwise the Python user would be able
In the example, \var{self} will always be a \code{NULL} pointer, since to cause the Python interpreter to crash (rather than raising an
we are defining a function, not a method (this is done so that the exception) by passing invalid arguments to a function in an extension
interpreter doesn't have to understand two different types of C module. Because argument checking and converting arguments to C are
functions). such common tasks, there's a general function in the Python
interpreter that combines them: \code{PyArg_ParseTuple()}. It uses a
The \var{args} parameter will be a pointer to a Python object, or template string to determine the types of the Python argument and the
\code{NULL} if the Python function/method was called without types of the C variables into which it should store the converted
arguments. It is necessary to do full argument type checking on each values (more about this later).
call, since otherwise the Python user would be able to cause the
Python interpreter to `dump core' by passing invalid arguments to a \code{PyArg_ParseTuple()} returns nonzero if all arguments have the
function in an extension module. Because argument checking and right type and its components have been stored in the variables whose
converting arguments to C are such common tasks, there's a general addresses are passed. It returns zero if an invalid argument was
function in the Python interpreter that combines them: passed. In the latter case it also raises an appropriate exception by
\code{getargs()}. It uses a template string to determine both the so the calling function can return \code{NULL} immediately. Here's
types of the Python argument and the types of the C variables into why:
which it should store the converted values.\footnote{There are
convenience macros \code{getnoarg()}, \code{getstrarg()},
\code{getintarg()}, etc., for many common forms of \code{getargs()} \section{Intermezzo: Errors and Exceptions}
templates. These are relics from the past; the recommended practice
is to call \code{getargs()} directly.} (More about this later.)
If \code{getargs()} returns nonzero, the argument list has the right
type and its components have been stored in the variables whose
addresses are passed. If it returns zero, an error has occurred. In
the latter case it has already raised an appropriate exception by so
the calling function should return \code{NULL} immediately --- see the
next section.
\section{Intermezzo: errors and exceptions}
An important convention throughout the Python interpreter is the An important convention throughout the Python interpreter is the
following: when a function fails, it should set an exception condition following: when a function fails, it should set an exception condition
and return an error value (often a \code{NULL} pointer). Exceptions and return an error value (usually a \code{NULL} pointer). Exceptions
are stored in a static global variable in \file{Python/errors.c}; if are stored in a static global variable inside the interpreter; if
this variable is \code{NULL} no exception has occurred. A second this variable is \code{NULL} no exception has occurred. A second
static global variable stores the `associated value' of the exception global variable stores the `associated value' of the exception
--- the second argument to \code{raise}. --- the second argument to \code{raise}. A third variable contains
the stack traceback in case the error originated in Python code.
The file \file{errors.h} declares a host of functions to set various These three variables are the C equivalents of the Python variables
types of exceptions. The most common one is \code{err_setstr()} --- \code{sys.exc_type}, \code{sys.exc_value} and \code{sys.exc_traceback}
its arguments are an exception object (e.g. \code{RuntimeError} --- --- see the section on module \code{sys} in the Library Reference
actually it can be any string object) and a C string indicating the Manual. It is important to know about them to understand how errors
cause of the error (this is converted to a string object and stored as are passed around.
the `associated value' of the exception). Another useful function is
\code{err_errno()}, which only takes an exception argument and The Python API defines a host of functions to set various types of
constructs the associated value by inspection of the (UNIX) global exceptions. The most common one is \code{PyErr_SetString()} --- its
variable errno. The most general function is \code{err_set()}, which arguments are an exception object (e.g. \code{PyExc_RuntimeError} ---
takes two object arguments, the exception and its associated value. actually it can be any object that is a legal exception indicator),
You don't need to \code{INCREF()} the objects passed to any of these and a C string indicating the cause of the error (this is converted to
a string object and stored as the `associated value' of the
exception). Another useful function is \code{PyErr_SetFromErrno()},
which only takes an exception argument and constructs the associated
value by inspection of the (\UNIX{}) global variable \code{errno}. The
most general function is \code{PyErr_SetObject()}, which takes two
object arguments, the exception and its associated value. You don't
need to \code{Py_INCREF()} the objects passed to any of these
functions. functions.
You can test non-destructively whether an exception has been set with You can test non-destructively whether an exception has been set with
\code{err_occurred()}. However, most code never calls \code{PyErr_Occurred()} --- this returns the current exception object,
\code{err_occurred()} to see whether an error occurred or not, but or \code{NULL} if no exception has occurred. Most code never needs to
relies on error return values from the functions it calls instead. call \code{PyErr_Occurred()} to see whether an error occurred or not,
but relies on error return values from the functions it calls instead.
When a function that calls another function detects that the called When a function that calls another function detects that the called
function fails, it should return an error value (e.g. \code{NULL} or function fails, it should return an error value (e.g. \code{NULL} or
\code{-1}) but not call one of the \code{err_*} functions --- one has \code{-1}). It shouldn't call one of the \code{PyErr_*} functions ---
already been called. The caller is then supposed to also return an one has already been called. The caller is then supposed to also
error indication to {\em its} caller, again {\em without} calling return an error indication to {\em its} caller, again {\em without}
\code{err_*()}, and so on --- the most detailed cause of the error was calling \code{PyErr_*()}, and so on --- the most detailed cause of the
already reported by the function that first detected it. Once the error was already reported by the function that first detected it.
error has reached Python's interpreter main loop, this aborts the Once the error has reached Python's interpreter main loop, this aborts
currently executing Python code and tries to find an exception handler the currently executing Python code and tries to find an exception
specified by the Python programmer. handler specified by the Python programmer.
(There are situations where a module can actually give a more detailed (There are situations where a module can actually give a more detailed
error message by calling another \code{err_*} function, and in such error message by calling another \code{PyErr_*} function, and in such
cases it is fine to do so. As a general rule, however, this is not cases it is fine to do so. As a general rule, however, this is not
necessary, and can cause information about the cause of the error to necessary, and can cause information about the cause of the error to
be lost: most operations can fail for a variety of reasons.) be lost: most operations can fail for a variety of reasons.)
To ignore an exception set by a function call that failed, the To ignore an exception set by a function call that failed, the exception
exception condition must be cleared explicitly by calling condition must be cleared explicitly by calling \code{PyErr_Clear()}.
\code{err_clear()}. The only time C code should call The only time C code should call \code{PyErr_Clear()} is if it doesn't
\code{err_clear()} is if it doesn't want to pass the error on to the want to pass the error on to the interpreter but wants to handle it
interpreter but wants to handle it completely by itself (e.g. by completely by itself (e.g. by trying something else or pretending
trying something else or pretending nothing happened). nothing happened).
Finally, the function \code{err_get()} gives you both error variables
{\em and clears them}. Note that even if an error occurred the second
one may be \code{NULL}. You have to \code{XDECREF()} both when you
are finished with them. I doubt you will need to use this function.
Note that a failing \code{malloc()} call must also be turned into an Note that a failing \code{malloc()} call must also be turned into an
exception --- the direct caller of \code{malloc()} (or exception --- the direct caller of \code{malloc()} (or
\code{realloc()}) must call \code{err_nomem()} and return a failure \code{realloc()}) must call \code{PyErr_NoMemory()} and return a
indicator itself. All the object-creating functions failure indicator itself. All the object-creating functions
(\code{newintobject()} etc.) already do this, so only if you call (\code{PyInt_FromLong()} etc.) already do this, so only if you call
\code{malloc()} directly this note is of importance. \code{malloc()} directly this note is of importance.
Also note that, with the important exception of \code{getargs()}, Also note that, with the important exception of
functions that return an integer status usually return \code{0} or a \code{PyArg_ParseTuple()}, functions that return an integer status
positive value for success and \code{-1} for failure. usually return \code{0} or a positive value for success and \code{-1}
for failure (like \UNIX{} system calls).
Finally, be careful about cleaning up garbage (making \code{XDECREF()} Finally, be careful about cleaning up garbage (making \code{Py_XDECREF()}
or \code{DECREF()} calls for objects you have already created) when or \code{Py_DECREF()} calls for objects you have already created) when
you return an error! you return an error!
The choice of which exception to raise is entirely yours. There are The choice of which exception to raise is entirely yours. There are
predeclared C objects corresponding to all built-in Python exceptions, predeclared C objects corresponding to all built-in Python exceptions,
e.g. \code{ZeroDevisionError} which you can use directly. Of course, e.g. \code{PyExc_ZeroDevisionError} which you can use directly. Of
you should chose exceptions wisely --- don't use \code{TypeError} to course, you should chose exceptions wisely --- don't use
mean that a file couldn't be opened (that should probably be \code{PyExc_TypeError} to mean that a file couldn't be opened (that
\code{IOError}). If anything's wrong with the argument list the should probably be \code{PyExc_IOError}). If something's wrong with
\code{getargs()} function raises \code{TypeError}. If you have an the argument list, the \code{PyArg_ParseTuple()} function usually
argument whose value which must be in a particular range or must raises \code{PyExc_TypeError}. If you have an argument whose value
satisfy other conditions, \code{ValueError} is appropriate. which must be in a particular range or must satisfy other conditions,
\code{PyExc_ValueError} is appropriate.
You can also define a new exception that is unique to your module. You can also define a new exception that is unique to your module.
For this, you usually declare a static object variable at the For this, you usually declare a static object variable at the
beginning of your file, e.g. beginning of your file, e.g.
\begin{verbatim} \begin{verbatim}
static object *FooError; static PyObject *SpamError;
\end{verbatim} \end{verbatim}
and initialize it in your module's initialization function and initialize it in your module's initialization function
(\code{initfoo()}) with a string object, e.g. (leaving out the error (\code{initspam()}) with a string object, e.g. (leaving out the error
checking for simplicity): checking for simplicity):
\begin{verbatim} \begin{verbatim}
void void
initfoo() initspam()
{ {
object *m, *d; PyObject *m, *d;
m = initmodule("foo", foo_methods); m = Py_InitModule("spam", spam_methods);
d = getmoduledict(m); d = PyModule_GetDict(m);
FooError = newstringobject("foo.error"); SpamError = PyString_FromString("spam.error");
dictinsert(d, "error", FooError); PyDict_SetItemString(d, "error", SpamError);
} }
\end{verbatim} \end{verbatim}
Note that the Python name for the exception object is \code{spam.error}
--- it is conventional for module and exception names to be spelled in
lower case. It is also conventional that the \emph{value} of the
exception object is the same as its name, e.g.\ the string
\code{"spam.error"}.
\section{Back to the example}
Going back to \code{posix_system()}, you should now be able to \section{Back to the Example}
understand this bit:
Going back to our example function, you should now be able to
understand this statement:
\begin{verbatim} \begin{verbatim}
if (!getargs(args, "s", &command)) if (!PyArg_ParseTuple(args, "s", &command))
return NULL; return NULL;
\end{verbatim} \end{verbatim}
It returns \code{NULL} (the error indicator for functions of this It returns \code{NULL} (the error indicator for functions returning
kind) if an error is detected in the argument list, relying on the object pointers) if an error is detected in the argument list, relying
exception set by \code{getargs()}. Otherwise the string value of the on the exception set by \code{PyArg_ParseTuple()}. Otherwise the
argument has been copied to the local variable \code{command} --- this string value of the argument has been copied to the local variable
is in fact just a pointer assignment and you are not supposed to \code{command}. This is a pointer assignment and you are not supposed
modify the string to which it points. to modify the string to which it points (so in ANSI C, the variable
\code{command} should properly be declared as \code{const char
If a function is called with multiple arguments, the argument list *command}).
(the argument \code{args}) is turned into a tuple. If it is called
without arguments, \code{args} is \code{NULL}. \code{getargs()} knows
about this; see later.
The next statement in \code{posix_system()} is a call to the C library The next statement is a call to the \UNIX{} function \code{system()},
function \code{system()}, passing it the string we just got from passing it the string we just got from \code{PyArg_ParseTuple()}:
\code{getargs()}:
\begin{verbatim} \begin{verbatim}
sts = system(command); sts = system(command);
\end{verbatim} \end{verbatim}
Finally, \code{posix.system()} must return a value: the integer status Our \code{spam.system()} function must return a value: the integer
returned by the C library \code{system()} function. This is done \code{sts} which contains the return value of the \UNIX{}
using the function \code{mkvalue()}, which is something like the \code{system()} function. This is done using the function
inverse of \code{getargs()}: it takes a format string and a variable \code{Py_BuildValue()}, which is something like the inverse of
number of C values and returns a new Python object. \code{PyArg_ParseTuple()}: it takes a format string and an arbitrary
number of C values, and returns a new Python object. More info on
\code{Py_BuildValue()} is given later.
\begin{verbatim} \begin{verbatim}
return mkvalue("i", sts); return Py_BuildValue("i", sts);
\end{verbatim} \end{verbatim}
In this case, it returns an integer object (yes, even integers are In this case, it will return an integer object. (Yes, even integers
objects on the heap in Python!). More info on \code{mkvalue()} is are objects on the heap in Python!)
given later.
If you had a function that returned no useful argument (a.k.a. a If you have a C function that returns no useful argument (a function
procedure), you would need this idiom: returning \code{void}), the corresponding Python function must return
\code{None}. You need this idiom to do so:
\begin{verbatim} \begin{verbatim}
INCREF(None); Py_INCREF(Py_None);
return None; return Py_None;
\end{verbatim} \end{verbatim}
\code{None} is a unique Python object representing `no value'. It \code{Py_None} is the C name for the special Python object
differs from \code{NULL}, which means `error' in most contexts. \code{None}. It is a genuine Python object (not a \code{NULL}
pointer, which means `error' in most contexts, as we have seen).
\section{The module's function table} \section{The Module's Method Table and Initialization Function}
I promised to show how I made the function \code{posix_system()} I promised to show how \code{spam_system()} is called from Python
callable from Python programs. This is shown later in programs. First, we need to list its name and address in a ``method
\file{Modules/posixmodule.c}: table'':
\begin{verbatim} \begin{verbatim}
static struct methodlist posix_methods[] = { static PyMethodDef spam_methods[] = {
... ...
{"system", posix_system}, {"system", spam_system, 1},
... ...
{NULL, NULL} /* Sentinel */ {NULL, NULL} /* Sentinel */
}; };
\end{verbatim}
Note the third entry (\samp{1}). This is a flag telling the
interpreter the calling convention to be used for the C function. It
should normally always be \samp{1}; a value of \samp{0} means that an
obsolete variant of \code{PyArg_ParseTuple()} is used.
The method table must be passed to the interpreter in the module's
initialization function (which should be the only non-\code{static}
item defined in the module file):
\begin{verbatim}
void void
initposix() initspam()
{ {
(void) initmodule("posix", posix_methods); (void) Py_InitModule("spam", spam_methods);
} }
\end{verbatim} \end{verbatim}
(The actual \code{initposix()} is somewhat more complicated, but many When the Python program imports module \code{spam} for the first time,
extension modules can be as simple as shown here.) When the Python \code{initspam()} is called. It calls \code{Py_InitModule()}, which
program first imports module \code{posix}, \code{initposix()} is creates a ``module object'' (which is inserted in the dictionary
called, which calls \code{initmodule()} with specific parameters. \code{sys.modules} under the key \code{"spam"}), and inserts built-in
This creates a `module object' (which is inserted in the table function objects into the newly created module based upon the table
\code{sys.modules} under the key \code{'posix'}), and adds (an array of \code{PyMethodDef} structures) that was passed as its
built-in-function objects to the newly created module based upon the second argument. \code{Py_InitModule()} returns a pointer to the
table (of type struct methodlist) that was passed as its second
parameter. The function \code{initmodule()} returns a pointer to the
module object that it creates (which is unused here). It aborts with module object that it creates (which is unused here). It aborts with
a fatal error if the module could not be initialized satisfactorily, a fatal error if the module could not be initialized satisfactorily,
so you don't need to check for errors. so the caller doesn't need to check for errors.
\section{Compilation and linkage} \section{Compilation and Linkage}
There are two more things to do before you can use your new extension There are two more things to do before you can use your new extension
module: compiling and linking it with the Python system. If you use module: compiling and linking it with the Python system. If you use
...@@ -342,13 +383,13 @@ about this. ...@@ -342,13 +383,13 @@ about this.
If you can't use dynamic loading, or if you want to make your module a If you can't use dynamic loading, or if you want to make your module a
permanent part of the Python interpreter, you will have to change the permanent part of the Python interpreter, you will have to change the
configuration setup and rebuild the interpreter. Luckily, in the 1.0 configuration setup and rebuild the interpreter. Luckily, this is
release this is very simple: just place your file (named very simple: just place your file (\file{spammodule.c} for example) in
\file{foomodule.c} for example) in the \file{Modules} directory, add a the \file{Modules} directory, add a line to the file
line to the file \file{Modules/Setup} describing your file: \file{Modules/Setup} describing your file:
\begin{verbatim} \begin{verbatim}
foo foomodule.o spam spammodule.o
\end{verbatim} \end{verbatim}
and rebuild the interpreter by running \code{make} in the toplevel and rebuild the interpreter by running \code{make} in the toplevel
...@@ -357,8 +398,15 @@ subdirectory, but then you must first rebuilt the \file{Makefile} ...@@ -357,8 +398,15 @@ subdirectory, but then you must first rebuilt the \file{Makefile}
there by running \code{make Makefile}. (This is necessary each time there by running \code{make Makefile}. (This is necessary each time
you change the \file{Setup} file.) you change the \file{Setup} file.)
If your module requires additional libraries to link with, these can
be listed on the line in the \file{Setup} file as well, for instance:
\begin{verbatim}
spam spammodule.o -lX11
\end{verbatim}
\section{Calling Python functions from C} \section{Calling Python Functions From C}
So far we have concentrated on making C functions callable from So far we have concentrated on making C functions callable from
Python. The reverse is also useful: calling Python functions from C. Python. The reverse is also useful: calling Python functions from C.
...@@ -378,211 +426,259 @@ Calling a Python function is easy. First, the Python program must ...@@ -378,211 +426,259 @@ Calling a Python function is easy. First, the Python program must
somehow pass you the Python function object. You should provide a somehow pass you the Python function object. You should provide a
function (or some other interface) to do this. When this function is function (or some other interface) to do this. When this function is
called, save a pointer to the Python function object (be careful to called, save a pointer to the Python function object (be careful to
\code{INCREF()} it!) in a global variable --- or whereever you see fit. \code{Py_INCREF()} it!) in a global variable --- or whereever you see fit.
For example, the following function might be part of a module For example, the following function might be part of a module
definition: definition:
\begin{verbatim} \begin{verbatim}
static object *my_callback = NULL; static PyObject *my_callback = NULL;
static object * static PyObject *
my_set_callback(dummy, arg) my_set_callback(dummy, arg)
object *dummy, *arg; PyObject *dummy, *arg;
{ {
XDECREF(my_callback); /* Dispose of previous callback */ Py_XDECREF(my_callback); /* Dispose of previous callback */
my_callback = arg; Py_XINCREF(arg); /* Add a reference to new callback */
XINCREF(my_callback); /* Remember new callback */ my_callback = arg; /* Remember new callback */
/* Boilerplate for "void" return */ /* Boilerplate to return "None" */
INCREF(None); Py_INCREF(Py_None);
return None; return Py_None;
} }
\end{verbatim} \end{verbatim}
This particular function doesn't do any typechecking on its argument The macros \code{Py_XINCREF()} and \code{Py_XDECREF()} increment/decrement
--- that will be done by \code{call_object()}, which is a bit late but
at least protects the Python interpreter from shooting itself in its
foot. (The problem with typechecking functions is that there are at
least five different Python object types that can be called, so the
test would be somewhat cumbersome.)
The macros \code{XINCREF()} and \code{XDECREF()} increment/decrement
the reference count of an object and are safe in the presence of the reference count of an object and are safe in the presence of
\code{NULL} pointers. More info on them in the section on Reference \code{NULL} pointers. More info on them in the section on Reference
Counts below. Counts below.
Later, when it is time to call the function, you call the C function Later, when it is time to call the function, you call the C function
\code{call_object()}. This function has two arguments, both pointers \code{PyEval_CallObject()}. This function has two arguments, both
to arbitrary Python objects: the Python function, and the argument pointers to arbitrary Python objects: the Python function, and the
list. The argument list must always be a tuple object, whose length argument list. The argument list must always be a tuple object, whose
is the number of arguments. To call the Python function with no length is the number of arguments. To call the Python function with
arguments, you must pass an empty tuple. For example: no arguments, pass an empty tuple; to call it with one argument, pass
a singleton tuple. \code{Py_BuildValue()} returns a tuple when its
format string consists of zero or more format codes between
parentheses. For example:
\begin{verbatim} \begin{verbatim}
object *arglist; int arg;
object *result; PyObject *arglist;
PyObject *result;
...
arg = 123;
... ...
/* Time to call the callback */ /* Time to call the callback */
arglist = newtupleobject(0); arglist = Py_BuildValue("(i)", arg);
result = call_object(my_callback, arglist); result = PyEval_CallObject(my_callback, arglist);
DECREF(arglist); Py_DECREF(arglist);
\end{verbatim} \end{verbatim}
\code{call_object()} returns a Python object pointer: this is \code{PyEval_CallObject()} returns a Python object pointer: this is
the return value of the Python function. \code{call_object()} is the return value of the Python function. \code{PyEval_CallObject()} is
`reference-count-neutral' with respect to its arguments. In the `reference-count-neutral' with respect to its arguments. In the
example a new tuple was created to serve as the argument list, which example a new tuple was created to serve as the argument list, which
is \code{DECREF()}-ed immediately after the call. is \code{Py_DECREF()}-ed immediately after the call.
The return value of \code{call_object()} is `new': either it is a The return value of \code{PyEval_CallObject()} is ``new'': either it
brand new object, or it is an existing object whose reference count is a brand new object, or it is an existing object whose reference
has been incremented. So, unless you want to save it in a global count has been incremented. So, unless you want to save it in a
variable, you should somehow \code{DECREF()} the result, even global variable, you should somehow \code{Py_DECREF()} the result,
(especially!) if you are not interested in its value. even (especially!) if you are not interested in its value.
Before you do this, however, it is important to check that the return Before you do this, however, it is important to check that the return
value isn't \code{NULL}. If it is, the Python function terminated by raising value isn't \code{NULL}. If it is, the Python function terminated by raising
an exception. If the C code that called \code{call_object()} is an exception. If the C code that called \code{PyEval_CallObject()} is
called from Python, it should now return an error indication to its called from Python, it should now return an error indication to its
Python caller, so the interpreter can print a stack trace, or the Python caller, so the interpreter can print a stack trace, or the
calling Python code can handle the exception. If this is not possible calling Python code can handle the exception. If this is not possible
or desirable, the exception should be cleared by calling or desirable, the exception should be cleared by calling
\code{err_clear()}. For example: \code{PyErr_Clear()}. For example:
\begin{verbatim} \begin{verbatim}
if (result == NULL) if (result == NULL)
return NULL; /* Pass error back */ return NULL; /* Pass error back */
/* Here maybe use the result */ ...use result...
DECREF(result); Py_DECREF(result);
\end{verbatim} \end{verbatim}
Depending on the desired interface to the Python callback function, Depending on the desired interface to the Python callback function,
you may also have to provide an argument list to \code{call_object()}. you may also have to provide an argument list to \code{PyEval_CallObject()}.
In some cases the argument list is also provided by the Python In some cases the argument list is also provided by the Python
program, through the same interface that specified the callback program, through the same interface that specified the callback
function. It can then be saved and used in the same manner as the function. It can then be saved and used in the same manner as the
function object. In other cases, you may have to construct a new function object. In other cases, you may have to construct a new
tuple to pass as the argument list. The simplest way to do this is to tuple to pass as the argument list. The simplest way to do this is to
call \code{mkvalue()}. For example, if you want to pass an integral call \code{Py_BuildValue()}. For example, if you want to pass an integral
event code, you might use the following code: event code, you might use the following code:
\begin{verbatim} \begin{verbatim}
object *arglist; PyObject *arglist;
... ...
arglist = mkvalue("(l)", eventcode); arglist = Py_BuildValue("(l)", eventcode);
result = call_object(my_callback, arglist); result = PyEval_CallObject(my_callback, arglist);
DECREF(arglist); Py_DECREF(arglist);
if (result == NULL) if (result == NULL)
return NULL; /* Pass error back */ return NULL; /* Pass error back */
/* Here maybe use the result */ /* Here maybe use the result */
DECREF(result); Py_DECREF(result);
\end{verbatim} \end{verbatim}
Note the placement of DECREF(argument) immediately after the call, Note the placement of \code{Py_DECREF(argument)} immediately after the call,
before the error check! Also note that strictly spoken this code is before the error check! Also note that strictly spoken this code is
not complete: \code{mkvalue()} may run out of memory, and this should not complete: \code{Py_BuildValue()} may run out of memory, and this should
be checked. be checked.
\section{Format strings for {\tt getargs()}} \section{Format Strings for {\tt PyArg_ParseTuple()}}
The \code{getargs()} function is declared in \file{modsupport.h} as The \code{PyArg_ParseTuple()} function is declared as follows:
follows:
\begin{verbatim} \begin{verbatim}
int getargs(object *arg, char *format, ...); int PyArg_ParseTuple(PyObject *arg, char *format, ...);
\end{verbatim} \end{verbatim}
The remaining arguments must be addresses of variables whose type is The \var{arg} argument must be a tuple object containing an argument
list passed from Python to a C function. The \var{format} argument
must be a format string, whose syntax is explained below. The
remaining arguments must be addresses of variables whose type is
determined by the format string. For the conversion to succeed, the determined by the format string. For the conversion to succeed, the
\var{arg} object must match the format and the format must be exhausted. \var{arg} object must match the format and the format must be
Note that while \code{getargs()} checks that the Python object really exhausted.
is of the specified type, it cannot check the validity of the
addresses of C variables provided in the call: if you make mistakes Note that while \code{PyArg_ParseTuple()} checks that the Python
there, your code will probably dump core. arguments have the required types, it cannot check the validity of the
addresses of C variables passed to the call: if you make mistakes
A non-empty format string consists of a single `format unit'. A there, your code will probably crash or at least overwrite random bits
format unit describes one Python object; it is usually a single in memory. So be careful!
character or a parenthesized sequence of format units. The type of a
format units is determined from its first character, the `format A format string consists of zero or more ``format units''. A format
letter': unit describes one Python object; it is usually a single character or
a parenthesized sequence of format units. With a few exceptions, a
format unit that is not a parenthesized sequence normally corresponds
to a single address argument to \code{PyArg_ParseTuple()}. In the
following description, the quoted form is the format unit; the entry
in (round) parentheses is the Python object type that matches the
format unit; and the entry in [square] brackets is the type of the C
variable(s) whose address should be passed. (Use the \samp{\&}
operator to pass a variable's address.)
\begin{description} \begin{description}
\item[\samp{s} (string)] \item[\samp{s} (string) [char *]]
The Python object must be a string object. The C argument must be a Convert a Python string to a C pointer to a character string. You
\code{(char**)} (i.e. the address of a character pointer), and a pointer must not provide storage for the string itself; a pointer to an
to the C string contained in the Python object is stored into it. You existing string is stored into the character pointer variable whose
must not provide storage to store the string; a pointer to an existing address you pass. The C string is null-terminated. The Python string
string is stored into the character pointer variable whose address you must not contain embedded null bytes; if it does, a \code{TypeError}
pass. If the next character in the format string is \samp{\#}, exception is raised.
another C argument of type \code{(int*)} must be present, and the
length of the Python string (not counting the trailing zero byte) is \item[\samp{s\#} (string) {[char *, int]}]
stored into it. This variant on \code{'s'} stores into two C variables, the first one
a pointer to a character string, the second one its length. In this
\item[\samp{z} (string or zero, i.e. \code{NULL})] case the Python string may contain embedded null bytes.
Like \samp{s}, but the object may also be None. In this case the
string pointer is set to \code{NULL} and if a \samp{\#} is present the \item[\samp{z} (string or \code{None}) {[char *]}]
size is set to 0. Like \samp{s}, but the Python object may also be \code{None}, in which
case the C pointer is set to \code{NULL}.
\item[\samp{b} (byte, i.e. char interpreted as tiny int)]
The object must be a Python integer. The C argument must be a \item[\samp{z\#} (string or \code{None}) {[char *, int]}]
\code{(char*)}. This is to \code{'s\#'} as \code{'z'} is to \code{'s'}.
\item[\samp{h} (half, i.e. short)] \item[\samp{b} (integer) {[char]}]
The object must be a Python integer. The C argument must be a Convert a Python integer to a tiny int, stored in a C \code{char}.
\code{(short*)}.
\item[\samp{h} (integer) {[short int]}]
\item[\samp{i} (int)] Convert a Python integer to a C \code{short int}.
The object must be a Python integer. The C argument must be an
\code{(int*)}. \item[\samp{i} (integer) {[int]}]
Convert a Python integer to a plain C \code{int}.
\item[\samp{l} (long)]
The object must be a (plain!) Python integer. The C argument must be \item[\samp{l} (integer) {[long int]}]
a \code{(long*)}. Convert a Python integer to a C \code{long int}.
\item[\samp{c} (char)] \item[\samp{c} (string of length 1) {[char]}]
The Python object must be a string of length 1. The C argument must Convert a Python character, represented as a string of length 1, to a
be a \code{(char*)}. (Don't pass an \code{(int*)}!) C \code{char}.
\item[\samp{f} (float)] \item[\samp{f} (float) {[float]}]
The object must be a Python int or float. The C argument must be a Convert a Python floating point number to a C \code{float}.
\code{(float*)}.
\item[\samp{d} (float) {[double]}]
\item[\samp{d} (double)] Convert a Python floating point number to a C \code{double}.
The object must be a Python int or float. The C argument must be a
\code{(double*)}. \item[\samp{O} (object) {[PyObject *]}]
Store a Python object (without any conversion) in a C object pointer.
\item[\samp{S} (string object)] The C program thus receives the actual object that was passed. The
The object must be a Python string. The C argument must be an object's reference count is not increased. The pointer stored is not
\code{(object**)} (i.e. the address of an object pointer). The C \code{NULL}.
program thus gets back the actual string object that was passed, not
just a pointer to its array of characters and its size as for format \item[\samp{O!} (object) {[\var{typeobject}, PyObject *]}]
character \samp{s}. The reference count of the object has not been Store a Python object in a C object pointer. This is similar to
increased. \samp{O}, but takes two C arguments: the first is the address of a
Python type object, the second is the address of the C variable (of
\item[\samp{O} (object)] type \code{PyObject *}) into which the object pointer is stored.
The object can be any Python object, including None, but not If the Python object does not have the required type, a
\code{NULL}. The C argument must be an \code{(object**)}. This can be \code{TypeError} exception is raised.
used if an argument list must contain objects of a type for which no
format letter exist: the caller must then check that it has the right \item[\samp{O\&} (object) {[\var{converter}, \var{anything}]}]
type. The reference count of the object has not been increased. Convert a Python object to a C variable through a \var{converter}
function. This takes two arguments: the first is a function, the
\item[\samp{(} (tuple)] second is the address of a C variable (of arbitrary type), converted
The object must be a Python tuple. Following the \samp{(} character to \code{void *}. The \var{converter} function in turn is called as
in the format string must come a number of format units describing the follows:
elements of the tuple, followed by a \samp{)} character. Tuple
format units may be nested. (There are no exceptions for empty and \code{\var{status} = \var{converter}(\var{object}, \var{address});}
singleton tuples; \samp{()} specifies an empty tuple and \samp{(i)} a
singleton of one integer. Normally you don't want to use the latter, where \var{object} is the Python object to be converted and
since it is hard for the Python user to specify. \var{address} is the \code{void *} argument that was passed to
\code{PyArg_ConvertTuple()}. The returned \var{status} should be
\code{1} for a successful conversion and \code{0} if the conversion
has failed. When the conversion fails, the \var{converter} function
should raise an exception.
\item[\samp{S} (string) {[PyStringObject *]}]
Like \samp{O} but raises a \code{TypeError} exception that the object
is a string object. The C variable may also be declared as
\code{PyObject *}.
\item[\samp{(\var{items})} (tuple) {[\var{matching-items}]}]
The object must be a Python tuple whose length is the number of format
units in \var{items}. The C arguments must correspond to the
individual format units in \var{items}. Format units for tuples may
be nested.
\end{description} \end{description}
More format characters will probably be added as the need arises. It It is possible to pass Python long integers where integers are
should (but currently isn't) be allowed to use Python long integers requested; however no proper range checking is done -- the most
whereever integers are expected, and perform a range check. (A range significant bits are silently truncated when the receiving field is
check is in fact always necessary for the \samp{b}, \samp{h} and too small to receive the value (actually, the semantics are inherited
\samp{i} format letters, but this is currently not implemented.) from downcasts in C --- your milage may vary).
A few other characters have a meaning in a format string. These may
not occur inside nested parentheses. They are:
\begin{description}
\item[\samp{|}]
Indicates that the remaining arguments in the Python argument list are
optional. The C variables corresponding to optional arguments should
be initialized to their default value --- when an optional argument is
not specified, the \code{PyArg_ParseTuple} does not touch the contents
of the corresponding C variable(s).
\item[\samp{:}]
The list of format units ends here; the string after the colon is used
as the function name in error messages (the ``associated value'' of
the exceptions that \code{PyArg_ParseTuple} raises).
\item[\samp{;}]
The list of format units ends here; the string after the colon is used
as the error message \emph{instead} of the default error message.
Clearly, \samp{:} and \samp{;} mutually exclude each other.
\end{description}
Some example calls: Some example calls:
...@@ -593,186 +689,444 @@ Some example calls: ...@@ -593,186 +689,444 @@ Some example calls:
char *s; char *s;
int size; int size;
ok = getargs(args, ""); /* No arguments */ ok = PyArg_ParseTuple(args, ""); /* No arguments */
/* Python call: f() */ /* Python call: f() */
ok = getargs(args, "s", &s); /* A string */ ok = PyArg_ParseTuple(args, "s", &s); /* A string */
/* Possible Python call: f('whoops!') */ /* Possible Python call: f('whoops!') */
ok = getargs(args, "(lls)", &k, &l, &s); /* Two longs and a string */ ok = PyArg_ParseTuple(args, "lls", &k, &l, &s); /* Two longs and a string */
/* Possible Python call: f(1, 2, 'three') */ /* Possible Python call: f(1, 2, 'three') */
ok = getargs(args, "((ii)s#)", &i, &j, &s, &size); ok = PyArg_ParseTuple(args, "(ii)s#", &i, &j, &s, &size);
/* A pair of ints and a string, whose size is also returned */ /* A pair of ints and a string, whose size is also returned */
/* Possible Python call: f(1, 2, 'three') */ /* Possible Python call: f(1, 2, 'three') */
{
char *file;
char *mode = "r";
int bufsize = 0;
ok = PyArg_ParseTuple(args, "s|si", &file, &mode, &bufsize);
/* A string, and optionally another string and an integer */
/* Possible Python calls:
f('spam')
f('spam', 'w')
f('spam', 'wb', 100000) */
}
{ {
int left, top, right, bottom, h, v; int left, top, right, bottom, h, v;
ok = getargs(args, "(((ii)(ii))(ii))", ok = PyArg_ParseTuple(args, "((ii)(ii))(ii)",
&left, &top, &right, &bottom, &h, &v); &left, &top, &right, &bottom, &h, &v);
/* A rectangle and a point */ /* A rectangle and a point */
/* Possible Python call: /* Possible Python call:
f( ((0, 0), (400, 300)), (10, 10)) */ f(((0, 0), (400, 300)), (10, 10)) */
} }
\end{verbatim} \end{verbatim}
Note that the `top level' of a non-empty format string must consist of
a single unit; strings like \samp{is} and \samp{(ii)s\#} are not valid
format strings. (But \samp{s\#} is.) If you have multiple arguments,
the format must therefore always be enclosed in parentheses, as in the
examples \samp{((ii)s\#)} and \samp{(((ii)(ii))(ii)}. (The current
implementation does not complain when more than one unparenthesized
format unit is given. Sorry.)
The \code{getargs()} function does not support variable-length \section{The {\tt Py_BuildValue()} Function}
argument lists. In simple cases you can fake these by trying several
calls to This function is the counterpart to \code{PyArg_ParseTuple()}. It is
\code{getargs()} until one succeeds, but you must take care to call declared as follows:
\code{err_clear()} before each retry. For example:
\begin{verbatim} \begin{verbatim}
static object *my_method(self, args) object *self, *args; { PyObject *Py_BuildValue(char *format, ...);
int i, j, k;
if (getargs(args, "(ii)", &i, &j)) {
k = 0; /* Use default third argument */
}
else {
err_clear();
if (!getargs(args, "(iii)", &i, &j, &k))
return NULL;
}
/* ... use i, j and k here ... */
INCREF(None);
return None;
}
\end{verbatim} \end{verbatim}
(It is possible to think of an extension to the definition of format It recognizes a set of format units similar to the ones recognized by
strings to accommodate this directly, e.g. placing a \samp{|} in a \code{PyArg_ParseTuple()}, but the arguments (which are input to the
tuple might specify that the remaining arguments are optional. function, not output) must not be pointers, just values. It returns a
\code{getargs()} should then return one more than the number of new Python object, suitable for returning from a C function called
variables stored into.) from Python.
One difference with \code{PyArg_ParseTuple()}: while the latter
requires its first argument to be a tuple (since Python argument lists
are always represented as tuples internally), \code{BuildValue()} does
not always build a tuple. It builds a tuple only if its format string
contains two or more format units. If the format string is empty, it
returns \code{None}; if it contains exactly one format unit, it
returns whatever object is described by that format unit. To force it
to return a tuple of size 0 or one, parenthesize the format string.
In the following description, the quoted form is the format unit; the
entry in (round) parentheses is the Python object type that the format
unit will return; and the entry in [square] brackets is the type of
the C value(s) to be passed.
The characters space, tab, colon and comma are ignored in format
strings (but not within format units such as \samp{s\#}). This can be
used to make long format strings a tad more readable.
\begin{description}
\item[\samp{s} (string) {[char *]}]
Convert a null-terminated C string to a Python object. If the C
string pointer is \code{NULL}, \code{None} is returned.
\item[\samp{s\#} (string) {[char *, int]}]
Convert a C string and its length to a Python object. If the C string
pointer is \code{NULL}, the length is ignored and \code{None} is
returned.
\item[\samp{z} (string or \code{None}) {[char *]}]
Same as \samp{s}.
\item[\samp{z\#} (string or \code{None}) {[char *, int]}]
Same as \samp{s\#}.
\item[\samp{i} (integer) {[int]}]
Convert a plain C \code{int} to a Python integer object.
Advanced users note: If you set the `varargs' flag in the method list \item[\samp{b} (integer) {[char]}]
for a function, the argument will always be a tuple (the `raw argument Same as \samp{i}.
list'). In this case you must enclose single and empty argument lists
in parentheses, e.g. \samp{(s)} and \samp{()}.
\item[\samp{h} (integer) {[short int]}]
Same as \samp{i}.
\section{The {\tt mkvalue()} function} \item[\samp{l} (integer) {[long int]}]
Convert a C \code{long int} to a Python integer object.
\item[\samp{c} (string of length 1) {[char]}]
Convert a C \code{int} representing a character to a Python string of
length 1.
\item[\samp{d} (float) {[double]}]
Convert a C \code{double} to a Python floating point number.
\item[\samp{f} (float) {[float]}]
Same as \samp{d}.
\item[\samp{O} (object) {[PyObject *]}]
Pass a Python object untouched (except for its reference count, which
is incremented by one). If the object passed in is a \code{NULL}
pointer, it is assumed that this was caused because the call producing
the argument found an error and set an exception. Therefore,
\code{Py_BuildValue()} will return \code{NULL} but won't raise an
exception. If no exception has been raised yet,
\code{PyExc_SystemError} is set.
\item[\samp{S} (object) {[PyObject *]}]
Same as \samp{O}.
\item[\samp{O\&} (object) {[\var{converter}, \var{anything}]}]
Convert \var{anything} to a Python object through a \var{converter}
function. The function is called with \var{anything} (which should be
compatible with \code{void *}) as its argument and should return a
``new'' Python object, or \code{NULL} if an error occurred.
\item[\samp{(\var{items})} (tuple) {[\var{matching-items}]}]
Convert a sequence of C values to a Python tuple with the same number
of items.
\item[\samp{[\var{items}]} (list) {[\var{matching-items}]}]
Convert a sequence of C values to a Python list with the same number
of items.
\item[\samp{\{\var{items}\}} (dictionary) {[\var{matching-items}]}]
Convert a sequence of C values to a Python dictionary. Each pair of
consecutive C values adds one item to the dictionary, serving as key
and value, respectively.
\end{description}
This function is the counterpart to \code{getargs()}. It is declared If there is an error in the format string, the
in \file{Include/modsupport.h} as follows: \code{PyExc_SystemError} exception is raised and \code{NULL} returned.
Examples (to the left the call, to the right the resulting Python value):
\begin{verbatim}
Py_BuildValue("") None
Py_BuildValue("i", 123) 123
Py_BuildValue("ii", 123, 456) (123, 456)
Py_BuildValue("s", "hello") 'hello'
Py_BuildValue("ss", "hello", "world") ('hello', 'world')
Py_BuildValue("s#", "hello", 4) 'hell'
Py_BuildValue("()") ()
Py_BuildValue("(i)", 123) (123,)
Py_BuildValue("(ii)", 123, 456) (123, 456)
Py_BuildValue("(i,i)", 123, 456) (123, 456)
Py_BuildValue("[i,i]", 123, 456) [123, 456]
Py_BuildValue("{s:i,s:i}", "abc", 123, "def", 456)
{'abc': 123, 'def': 456}
Py_BuildValue("((ii)(ii)) (ii)", 1, 2, 3, 4, 5, 6)
(((1, 2), (3, 4)), (5, 6))
\end{verbatim}
\section{Reference Counts}
\subsection{Introduction}
In languages like C or \Cpp{}, the programmer is responsible for
dynamic allocation and deallocation of memory on the heap. In C, this
is done using the functions \code{malloc()} and \code{free()}. In
\Cpp{}, the operators \code{new} and \code{delete} are used with
essentially the same meaning; they are actually implemented using
\code{malloc()} and \code{free()}, so we'll restrict the following
discussion to the latter.
Every block of memory allocated with \code{malloc()} should eventually
be returned to the pool of available memory by exactly one call to
\code{free()}. It is important to call \code{free()} at the right
time. If a block's address is forgotten but \code{free()} is not
called for it, the memory it occupies cannot be reused until the
program terminates. This is called a \dfn{memory leak}. On the other
hand, if a program calls \code{free()} for a block and then continues
to use the block, it creates a conflict with re-use of the block
through another \code{malloc()} call. This is called \dfn{using freed
memory} has the same bad consequences as referencing uninitialized
data --- core dumps, wrong results, mysterious crashes.
Common causes of memory leaks are unusual paths through the code. For
instance, a function may allocate a block of memory, do some
calculation, and then free the block again. Now a change in the
requirements for the function may add a test to the calculation that
detects an error condition and can return prematurely from the
function. It's easy to forget to free the allocated memory block when
taking this premature exit, especially when it is added later to the
code. Such leaks, once introduced, often go undetected for a long
time: the error exit is taken only in a small fraction of all calls,
and most modern machines have plenty of virtual memory, so the leak
only becomes apparent in a long-running process that uses the leaking
function frequently. Therefore, it's important to prevent leaks from
happening by having a coding convention or strategy that minimizes
this kind of errors.
Since Python makes heavy use of \code{malloc()} and \code{free()}, it
needs a strategy to avoid memory leaks as well as the use of freed
memory. The chosen method is called \dfn{reference counting}. The
principle is simple: every object contains a counter, which is
incremented when a reference to the object is stored somewhere, and
which is decremented when a reference to it is deleted. When the
counter reaches zero, the last reference to the object has been
deleted and the object is freed.
An alternative strategy is called \dfn{automatic garbage collection}.
(Sometimes, reference counting is also referred to as a garbage
collection strategy, hence my use of ``automatic'' to distinguish the
two.) The big advantage of automatic garbage collection is that the
user doesn't need to call \code{free()} explicitly. (Another claimed
advantage is an improvement in speed or memory usage --- this is no
hard fact however.) The disadvantage is that for C, there is no
truly portable automatic garbage collector, while reference counting
can be implemented portably (as long as the functions \code{malloc()}
and \code{free()} are available --- which the C Standard guarantees).
Maybe some day a sufficiently portable automatic garbage collector
will be available for C. Until then, we'll have to live with
reference counts.
\subsection{Reference Counting in Python}
There are two macros, \code{Py_INCREF(x)} and \code{Py_DECREF(x)},
which handle the incrementing and decrementing of the reference count.
\code{Py_DECREF()} also frees the object when the count reaches zero.
For flexibility, it doesn't call \code{free()} directly --- rather, it
makes a call through a function pointer in the object's \dfn{type
object}. For this purpose (and others), every object also contains a
pointer to its type object.
The big question now remains: when to use \code{Py_INCREF(x)} and
\code{Py_DECREF(x)}? Let's first introduce some terms. Nobody
``owns'' an object; however, you can \dfn{own a reference} to an
object. An object's reference count is now defined as the number of
owned references to it. The owner of a reference is responsible for
calling \code{Py_DECREF()} when the reference is no longer needed.
Ownership of a reference can be transferred. There are three ways to
dispose of an owned reference: pass it on, store it, or call
\code{Py_DECREF()}. Forgetting to dispose of an owned reference creates
a memory leak.
It is also possible to \dfn{borrow}\footnote{The metaphor of
``borrowing'' a reference is not completely correct: the owner still
has a copy of the reference.} a reference to an object. The borrower
of a reference should not call \code{Py_DECREF()}. The borrower must
not hold on to the object longer than the owner from which it was
borrowed. Using a borrowed reference after the owner has disposed of
it risks using freed memory and should be avoided
completely.\footnote{Checking that the reference count is at least 1
\strong{does not work} --- the reference count itself could be in
freed memory and may thus be reused for another object!}
The advantage of borrowing over owning a reference is that you don't
need to take care of disposing of the reference on all possible paths
through the code --- in other words, with a borrowed reference you
don't run the risk of leaking when a premature exit is taken. The
disadvantage of borrowing over leaking is that there are some subtle
situations where in seemingly correct code a borrowed reference can be
used after the owner from which it was borrowed has in fact disposed
of it.
A borrowed reference can be changed into an owned reference by calling
\code{Py_INCREF()}. This does not affect the status of the owner from
which the reference was borrowed --- it creates a new owned reference,
and gives full owner responsibilities (i.e., the new owner must
dispose of the reference properly, as well as the previous owner).
\subsection{Ownership Rules}
Whenever an object reference is passed into or out of a function, it
is part of the function's interface specification whether ownership is
transferred with the reference or not.
Most functions that return a reference to an object pass on ownership
with the reference. In particular, all functions whose function it is
to create a new object, e.g.\ \code{PyInt_FromLong()} and
\code{Py_BuildValue()}, pass ownership to the receiver. Even if in
fact, in some cases, you don't receive a reference to a brand new
object, you still receive ownership of the reference. For instance,
\code{PyInt_FromLong()} maintains a cache of popular values and can
return a reference to a cached item.
Many functions that extract objects from other objects also transfer
ownership with the reference, for instance
\code{PyObject_GetAttrString()}. The picture is less clear, here,
however, since a few common routines are exceptions:
\code{PyTuple_GetItem()}, \code{PyList_GetItem()} and
\code{PyDict_GetItem()} (and \code{PyDict_GetItemString()}) all return
references that you borrow from the tuple, list or dictionary.
The function \code{PyImport_AddModule()} also returns a borrowed
reference, even though it may actually create the object it returns:
this is possible because an owned reference to the object is stored in
\code{sys.modules}.
When you pass an object reference into another function, in general,
the function borrows the reference from you --- if it needs to store
it, it will use \code{Py_INCREF()} to become an independent owner.
There are exactly two important exceptions to this rule:
\code{PyTuple_SetItem()} and \code{PyList_SetItem()}. These functions
take over ownership of the item passed to them --- even if they fail!
(Note that \code{PyDict_SetItem()} and friends don't take over
ownership --- they are ``normal''.)
When a C function is called from Python, it borrows references to its
arguments from the caller. The caller owns a reference to the object,
so the borrowed reference's lifetime is guaranteed until the function
returns. Only when such a borrowed reference must be stored or passed
on, it must be turned into an owned reference by calling
\code{Py_INCREF()}.
The object reference returned from a C function that is called from
Python must be an owned reference --- ownership is tranferred from the
function to its caller.
\subsection{Thin Ice}
There are a few situations where seemingly harmless use of a borrowed
reference can lead to problems. These all have to do with implicit
invocations of the interpreter, which can cause the owner of a
reference to dispose of it.
The first and most important case to know about is using
\code{Py_DECREF()} on an unrelated object while borrowing a reference
to a list item. For instance:
\begin{verbatim}
bug(PyObject *list) {
PyObject *item = PyList_GetItem(list, 0);
PyList_SetItem(list, 1, PyInt_FromLong(0L));
PyObject_Print(item, stdout, 0); /* BUG! */
}
\end{verbatim}
This function first borrows a reference to \code{list[0]}, then
replaces \code{list[1]} with the value \code{0}, and finally prints
the borrowed reference. Looks harmless, right? But it's not!
Let's follow the control flow into \code{PyList_SetItem()}. The list
owns references to all its items, so when item 1 is replaced, it has
to dispose of the original item 1. Now let's suppose the original
item 1 was an instance of a user-defined class, and let's further
suppose that the class defined a \code{__del__()} method. If this
class instance has a reference count of 1, disposing of it will call
its \code{__del__()} method.
Since it is written in Python, the \code{__del__()} method can execute
arbitrary Python code. Could it perhaps do something to invalidate
the reference to \code{item} in \code{bug()}? You bet! Assuming that
the list passed into \code{bug()} is accessible to the
\code{__del__()} method, it could execute a statement to the effect of
\code{del list[0]}, and assuming this was the last reference to that
object, it would free the memory associated with it, thereby
invalidating \code{item}.
The solution, once you know the source of the problem, is easy:
temporarily increment the reference count. The correct version of the
function reads:
\begin{verbatim} \begin{verbatim}
object *mkvalue(char *format, ...); no_bug(PyObject *list) {
PyObject *item = PyList_GetItem(list, 0);
Py_INCREF(item);
PyList_SetItem(list, 1, PyInt_FromLong(0L));
PyObject_Print(item, stdout, 0);
Py_DECREF(item);
}
\end{verbatim} \end{verbatim}
It supports exactly the same format letters as \code{getargs()}, but This is a true story. An older version of Python contained variants
the arguments (which are input to the function, not output) must not of this bug and someone spent a considerable amount of time in a C
be pointers, just values. If a byte, short or float is passed to a debugger to figure out why his \code{__del__()} methods would fail...
varargs function, it is widened by the compiler to int or double, so
\samp{b} and \samp{h} are treated as \samp{i} and \samp{f} is The second case of problems with a borrowed reference is a variant
treated as \samp{d}. \samp{S} is treated as \samp{O}, \samp{s} is involving threads. Normally, multiple threads in the Python
treated as \samp{z}. \samp{z\#} and \samp{s\#} are supported: a interpreter can't get in each other's way, because there is a global
second argument specifies the length of the data (negative means use lock protecting Python's entire object space. However, it is possible
\code{strlen()}). \samp{S} and \samp{O} add a reference to their to temporarily release this lock using the macro
argument (so you should \code{DECREF()} it if you've just created it \code{Py_BEGIN_ALLOW_THREADS}, and to re-acquire it using
and aren't going to use it again). \code{Py_END_ALLOW_THREADS}. This is common around blocking I/O
calls, to let other threads use the CPU while waiting for the I/O to
If the argument for \samp{O} or \samp{S} is a \code{NULL} pointer, it is complete. Obviously, the following function has the same problem as
assumed that this was caused because the call producing the argument the previous one:
found an error and set an exception. Therefore, \code{mkvalue()} will
return \code{NULL} but won't set an exception if one is already set.
If no exception is set, \code{SystemError} is set.
If there is an error in the format string, the \code{SystemError}
exception is set, since it is the calling C code's fault, not that of
the Python user who sees the exception.
Example:
\begin{verbatim} \begin{verbatim}
return mkvalue("(ii)", 0, 0); bug(PyObject *list) {
PyObject *item = PyList_GetItem(list, 0);
Py_BEGIN_ALLOW_THREADS
...some blocking I/O call...
Py_END_ALLOW_THREADS
PyObject_Print(item, stdout, 0); /* BUG! */
}
\end{verbatim} \end{verbatim}
returns a tuple containing two zeros. (Outer parentheses in the \subsection{NULL Pointers}
format string are actually superfluous, but you can use them for
compatibility with \code{getargs()}, which requires them if more than In general, functions that take object references as arguments don't
one argument is expected.) expect you to pass them \code{NULL} pointers, and will dump core (or
cause later core dumps) if you do so. Functions that return object
references generally return \code{NULL} only to indicate that an
\section{Reference counts} exception occurred. The reason for not testing for \code{NULL}
arguments is that functions often pass the objects they receive on to
Here's a useful explanation of \code{INCREF()} and \code{DECREF()} other function --- if each function were to test for \code{NULL},
(after an original by Sjoerd Mullender). there would be a lot of redundant tests and the code would run slower.
Use \code{XINCREF()} or \code{XDECREF()} instead of \code{INCREF()} or It is better to test for \code{NULL} only at the ``source'', i.e.\
\code{DECREF()} when the argument may be \code{NULL} --- the versions when a pointer that may be \code{NULL} is received, e.g.\ from
without \samp{X} are faster but wull dump core when they encounter a \code{malloc()} or from a function that may raise an exception.
\code{NULL} pointer.
The macros \code{Py_INCREF()} and \code{Py_DECREF()}
The basic idea is, if you create an extra reference to an object, you don't check for \code{NULL} pointers --- however, their variants
must \code{INCREF()} it, if you throw away a reference to an object, \code{Py_XINCREF()} and \code{Py_XDECREF()} do.
you must \code{DECREF()} it. Functions such as
\code{newstringobject()}, \code{newsizedstringobject()}, The macros for checking for a particular object type
\code{newintobject()}, etc. create a reference to an object. If you (\code{Py\var{type}_Check()}) don't check for \code{NULL} pointers ---
want to throw away the object thus created, you must use again, there is much code that calls several of these in a row to test
\code{DECREF()}. an object against various different expected types, and this would
generate redundant tests. There are no variants with \code{NULL}
If you put an object into a tuple or list using \code{settupleitem()} checking.
or \code{setlistitem()}, the idea is that you usually don't want to
keep a reference of your own around, so Python does not The C function calling mechanism guarantees that the argument list
\code{INCREF()} the elements. It does \code{DECREF()} the old value. passed to C functions (\code{args} in the examples) is never
This means that if you put something into such an object using the \code{NULL} --- in fact it guarantees that it is always a tuple.%
functions Python provides for this, you must \code{INCREF()} the \footnote{These guarantees don't hold when you use the ``old'' style
object if you also want to keep a separate reference to the object around. calling convention --- this is still found in much existing code.}
Also, if you replace an element, you should \code{INCREF()} the old
element first if you want to keep it. If you didn't \code{INCREF()} It is a severe error to ever let a \code{NULL} pointer ``escape'' to
it before you replaced it, you are not allowed to look at it anymore, the Python user.
since it may have been freed.
Returning an object to Python (i.e. when your C function returns) \section{Writing Extensions in \Cpp{}}
creates a reference to an object, but it does not change the reference
count. When your code does not keep another reference to the object,
you should not \code{INCREF()} or \code{DECREF()} it (assuming it is a
newly created object). When you do keep a reference around, you
should \code{INCREF()} the object. Also, when you return a global
object such as \code{None}, you should \code{INCREF()} it.
If you want to return a tuple, you should consider using
\code{mkvalue()}. This function creates a new tuple with a reference
count of 1 which you can return. If any of the elements you put into
the tuple are objects (format codes \samp{O} or \samp{S}), they
are \code{INCREF()}'ed by \code{mkvalue()}. If you don't want to keep
references to those elements around, you should \code{DECREF()} them
after having called \code{mkvalue()}.
Usually you don't have to worry about arguments. They are
\code{INCREF()}'ed before your function is called and
\code{DECREF()}'ed after your function returns. When you keep a
reference to an argument, you should \code{INCREF()} it and
\code{DECREF()} when you throw it away. Also, when you return an
argument, you should \code{INCREF()} it, because returning the
argument creates an extra reference to it.
If you use \code{getargs()} to parse the arguments, you can get a
reference to an object (by using \samp{O} in the format string). This
object was not \code{INCREF()}'ed, so you should not \code{DECREF()}
it. If you want to keep the object, you must \code{INCREF()} it
yourself.
If you create your own type of objects, you should use \code{NEWOBJ()}
to create the object. This sets the reference count to 1. If you
want to throw away the object, you should use \code{DECREF()}. When
the reference count reaches zero, your type's \code{dealloc()}
function is called. In it, you should \code{DECREF()} all object to
which you keep references in your object, but you should not use
\code{DECREF()} on your object. You should use \code{DEL()} instead.
\section{Writing extensions in \Cpp{}}
It is possible to write extension modules in \Cpp{}. Some restrictions It is possible to write extension modules in \Cpp{}. Some restrictions
apply: since the main program (the Python interpreter) is compiled and apply: since the main program (the Python interpreter) is compiled and
...@@ -782,8 +1136,9 @@ indirectly (i.e. via function pointers) by the Python interpreter will ...@@ -782,8 +1136,9 @@ indirectly (i.e. via function pointers) by the Python interpreter will
have to be declared using \code{extern "C"}; this applies to all have to be declared using \code{extern "C"}; this applies to all
`methods' as well as to the module's initialization function. `methods' as well as to the module's initialization function.
It is unnecessary to enclose the Python header files in It is unnecessary to enclose the Python header files in
\code{extern "C" \{...\}} --- they do this already. \code{extern "C" \{...\}} --- they use this form already if the symbol
\samp{__cplusplus} is defined (all recent C++ compilers define this
symbol).
\chapter{Embedding Python in another application} \chapter{Embedding Python in another application}
...@@ -797,16 +1152,16 @@ interpreter to run some Python code. ...@@ -797,16 +1152,16 @@ interpreter to run some Python code.
So if you are embedding Python, you are providing your own main So if you are embedding Python, you are providing your own main
program. One of the things this main program has to do is initialize program. One of the things this main program has to do is initialize
the Python interpreter. At the very least, you have to call the the Python interpreter. At the very least, you have to call the
function \code{initall()}. There are optional calls to pass command function \code{Py_Initialize()}. There are optional calls to pass command
line arguments to Python. Then later you can call the interpreter line arguments to Python. Then later you can call the interpreter
from any part of the application. from any part of the application.
There are several different ways to call the interpreter: you can pass There are several different ways to call the interpreter: you can pass
a string containing Python statements to \code{run_command()}, or you a string containing Python statements to \code{PyRun_SimpleString()},
can pass a stdio file pointer and a file name (for identification in or you can pass a stdio file pointer and a file name (for
error messages only) to \code{run_script()}. You can also call the identification in error messages only) to \code{PyRun_SimpleFile()}. You
lower-level operations described in the previous chapters to construct can also call the lower-level operations described in the previous
and use Python objects. chapters to construct and use Python objects.
A simple demo of embedding Python can be found in the directory A simple demo of embedding Python can be found in the directory
\file{Demo/embed}. \file{Demo/embed}.
...@@ -828,8 +1183,8 @@ dynamic loading of extension modules implemented in C. When shared ...@@ -828,8 +1183,8 @@ dynamic loading of extension modules implemented in C. When shared
libraries are used dynamic loading is configured automatically; libraries are used dynamic loading is configured automatically;
otherwise you have to select it as a build option (see below). Once otherwise you have to select it as a build option (see below). Once
configured, dynamic loading is trivial to use: when a Python program configured, dynamic loading is trivial to use: when a Python program
executes \code{import foo}, the search for modules tries to find a executes \code{import spam}, the search for modules tries to find a
file \file{foomodule.o} (\file{foomodule.so} when using shared file \file{spammodule.o} (\file{spammodule.so} when using shared
libraries) in the module search path, and if one is found, it is libraries) in the module search path, and if one is found, it is
loaded into the executing binary and executed. Once loaded, the loaded into the executing binary and executed. Once loaded, the
module acts just like a built-in extension module. module acts just like a built-in extension module.
...@@ -844,13 +1199,13 @@ loading a module that was compiled for a different version of Python ...@@ -844,13 +1199,13 @@ loading a module that was compiled for a different version of Python
(e.g. with a different representation of objects) may dump core. (e.g. with a different representation of objects) may dump core.
\section{Configuring and building the interpreter for dynamic loading} \section{Configuring and Building the Interpreter for Dynamic Loading}
There are three styles of dynamic loading: one using shared libraries, There are three styles of dynamic loading: one using shared libraries,
one using SGI IRIX 4 dynamic loading, and one using GNU dynamic one using SGI IRIX 4 dynamic loading, and one using GNU dynamic
loading. loading.
\subsection{Shared libraries} \subsection{Shared Libraries}
The following systems support dynamic loading using shared libraries: The following systems support dynamic loading using shared libraries:
SunOS 4; Solaris 2; SGI IRIX 5 (but not SGI IRIX 4!); and probably all SunOS 4; Solaris 2; SGI IRIX 5 (but not SGI IRIX 4!); and probably all
...@@ -862,7 +1217,7 @@ systems --- the \file{configure} detects the presence of the ...@@ -862,7 +1217,7 @@ systems --- the \file{configure} detects the presence of the
\file{<dlfcn.h>} header file and automatically configures dynamic \file{<dlfcn.h>} header file and automatically configures dynamic
loading. loading.
\subsection{SGI dynamic loading} \subsection{SGI IRIX 4 Dynamic Loading}
Only SGI IRIX 4 supports dynamic loading of modules using SGI dynamic Only SGI IRIX 4 supports dynamic loading of modules using SGI dynamic
loading. (SGI IRIX 5 might also support it but it is inferior to loading. (SGI IRIX 5 might also support it but it is inferior to
...@@ -883,7 +1238,7 @@ pathname of the \code{dl} directory. ...@@ -883,7 +1238,7 @@ pathname of the \code{dl} directory.
Now build and install Python as you normally would (see the Now build and install Python as you normally would (see the
\file{README} file in the toplevel Python directory.) \file{README} file in the toplevel Python directory.)
\subsection{GNU dynamic loading} \subsection{GNU Dynamic Loading}
GNU dynamic loading supports (according to its \file{README} file) the GNU dynamic loading supports (according to its \file{README} file) the
following hardware and software combinations: VAX (Ultrix), Sun 3 following hardware and software combinations: VAX (Ultrix), Sun 3
...@@ -909,13 +1264,13 @@ of the GNU DLD package. The Python interpreter you build hereafter ...@@ -909,13 +1264,13 @@ of the GNU DLD package. The Python interpreter you build hereafter
will support GNU dynamic loading. will support GNU dynamic loading.
\section{Building a dynamically loadable module} \section{Building a Dynamically Loadable Module}
Since there are three styles of dynamic loading, there are also three Since there are three styles of dynamic loading, there are also three
groups of instructions for building a dynamically loadable module. groups of instructions for building a dynamically loadable module.
Instructions common for all three styles are given first. Assuming Instructions common for all three styles are given first. Assuming
your module is called \code{foo}, the source filename must be your module is called \code{spam}, the source filename must be
\file{foomodule.c}, so the object name is \file{foomodule.o}. The \file{spammodule.c}, so the object name is \file{spammodule.o}. The
module must be written as a normal Python extension module (as module must be written as a normal Python extension module (as
described earlier). described earlier).
...@@ -931,7 +1286,7 @@ also add \samp{-DHAVE_CONFIG_H} to the definition of \var{CFLAGS} to ...@@ -931,7 +1286,7 @@ also add \samp{-DHAVE_CONFIG_H} to the definition of \var{CFLAGS} to
direct the Python headers to include \file{config.h}. direct the Python headers to include \file{config.h}.
\subsection{Shared libraries} \subsection{Shared Libraries}
You must link the \samp{.o} file to produce a shared library. This is You must link the \samp{.o} file to produce a shared library. This is
done using a special invocation of the \UNIX{} loader/linker, {\em done using a special invocation of the \UNIX{} loader/linker, {\em
...@@ -939,17 +1294,17 @@ ld}(1). Unfortunately the invocation differs slightly per system. ...@@ -939,17 +1294,17 @@ ld}(1). Unfortunately the invocation differs slightly per system.
On SunOS 4, use On SunOS 4, use
\begin{verbatim} \begin{verbatim}
ld foomodule.o -o foomodule.so ld spammodule.o -o spammodule.so
\end{verbatim} \end{verbatim}
On Solaris 2, use On Solaris 2, use
\begin{verbatim} \begin{verbatim}
ld -G foomodule.o -o foomodule.so ld -G spammodule.o -o spammodule.so
\end{verbatim} \end{verbatim}
On SGI IRIX 5, use On SGI IRIX 5, use
\begin{verbatim} \begin{verbatim}
ld -shared foomodule.o -o foomodule.so ld -shared spammodule.o -o spammodule.so
\end{verbatim} \end{verbatim}
On other systems, consult the manual page for {\em ld}(1) to find what On other systems, consult the manual page for {\em ld}(1) to find what
...@@ -960,46 +1315,46 @@ been linked with Python (e.g. a windowing system), these must be ...@@ -960,46 +1315,46 @@ been linked with Python (e.g. a windowing system), these must be
passed to the {\em ld} command as \samp{-l} options after the passed to the {\em ld} command as \samp{-l} options after the
\samp{.o} file. \samp{.o} file.
The resulting file \file{foomodule.so} must be copied into a directory The resulting file \file{spammodule.so} must be copied into a directory
along the Python module search path. along the Python module search path.
\subsection{SGI dynamic loading} \subsection{SGI IRIX 4 Dynamic Loading}
{bf IMPORTANT:} You must compile your extension module with the {bf IMPORTANT:} You must compile your extension module with the
additional C flag \samp{-G0} (or \samp{-G 0}). This instruct the additional C flag \samp{-G0} (or \samp{-G 0}). This instruct the
assembler to generate position-independent code. assembler to generate position-independent code.
You don't need to link the resulting \file{foomodule.o} file; just You don't need to link the resulting \file{spammodule.o} file; just
copy it into a directory along the Python module search path. copy it into a directory along the Python module search path.
The first time your extension is loaded, it takes some extra time and The first time your extension is loaded, it takes some extra time and
a few messages may be printed. This creates a file a few messages may be printed. This creates a file
\file{foomodule.ld} which is an image that can be loaded quickly into \file{spammodule.ld} which is an image that can be loaded quickly into
the Python interpreter process. When a new Python interpreter is the Python interpreter process. When a new Python interpreter is
installed, the \code{dl} package detects this and rebuilds installed, the \code{dl} package detects this and rebuilds
\file{foomodule.ld}. The file \file{foomodule.ld} is placed in the \file{spammodule.ld}. The file \file{spammodule.ld} is placed in the
directory where \file{foomodule.o} was found, unless this directory is directory where \file{spammodule.o} was found, unless this directory is
unwritable; in that case it is placed in a temporary unwritable; in that case it is placed in a temporary
directory.\footnote{Check the manual page of the \code{dl} package for directory.\footnote{Check the manual page of the \code{dl} package for
details.} details.}
If your extension modules uses additional system libraries, you must If your extension modules uses additional system libraries, you must
create a file \file{foomodule.libs} in the same directory as the create a file \file{spammodule.libs} in the same directory as the
\file{foomodule.o}. This file should contain one or more lines with \file{spammodule.o}. This file should contain one or more lines with
whitespace-separated options that will be passed to the linker --- whitespace-separated options that will be passed to the linker ---
normally only \samp{-l} options or absolute pathnames of libraries normally only \samp{-l} options or absolute pathnames of libraries
(\samp{.a} files) should be used. (\samp{.a} files) should be used.
\subsection{GNU dynamic loading} \subsection{GNU Dynamic Loading}
Just copy \file{foomodule.o} into a directory along the Python module Just copy \file{spammodule.o} into a directory along the Python module
search path. search path.
If your extension modules uses additional system libraries, you must If your extension modules uses additional system libraries, you must
create a file \file{foomodule.libs} in the same directory as the create a file \file{spammodule.libs} in the same directory as the
\file{foomodule.o}. This file should contain one or more lines with \file{spammodule.o}. This file should contain one or more lines with
whitespace-separated absolute pathnames of libraries (\samp{.a} whitespace-separated absolute pathnames of libraries (\samp{.a}
files). No \samp{-l} options can be used. files). No \samp{-l} options can be used.
......
Doc/ext/ext.tex
Dosyayı görüntüle @ 5049bcb1
\documentstyle[twoside,11pt,myformat]{report} \documentstyle[twoside,11pt,myformat]{report}
% XXX PM Modulator
\title{Extending and Embedding the Python Interpreter} \title{Extending and Embedding the Python Interpreter}
\input{boilerplate} \input{boilerplate}
...@@ -45,294 +47,333 @@ system supports this feature. ...@@ -45,294 +47,333 @@ system supports this feature.
It is quite easy to add non-standard built-in modules to Python, if It is quite easy to add non-standard built-in modules to Python, if
you know how to program in C. A built-in module known to the Python you know how to program in C. A built-in module known to the Python
programmer as \code{foo} is generally implemented by a file called programmer as \code{spam} is generally implemented by a file called
\file{foomodule.c}. All but the two most essential standard built-in \file{spammodule.c} (if the module name is very long, like
modules also adhere to this convention, and in fact some of them form \samp{spammify}, you can drop the \samp{module}, leaving a file name
excellent examples of how to create an extension. like \file{spammify.c}). The standard built-in modules also adhere to
this convention, and in fact some of them are excellent examples of
how to create an extension.
Extension modules can do two things that can't be done directly in Extension modules can do two things that can't be done directly in
Python: they can implement new data types (which are different from Python: they can implement new data types (which are different from
classes, by the way), and they can make system calls or call C library classes, by the way), and they can make system calls or call C library
functions. We'll see how both types of extension are implemented by functions.
examining the code for a Python curses interface.
Note: unless otherwise mentioned, all file references in this To support extensions, the Python API (Application Programmers
document are relative to the toplevel directory of the Python Interface) defines many functions, macros and variables that provide
distribution --- i.e. the directory that contains the \file{configure} access to almost every aspect of the Python run-time system.
script. Most of the Python API is imported by including the single header file
\code{"Python.h"}. All user-visible symbols defined by including this
file have a prefix of \samp{Py} or \samp{PY}, except those defined in
standard header files --- for convenience, and since they are needed by
the Python interpreter, \file{"Python.h"} includes a few standard
header files: \file{<stdio.h>}, \file{<string.h>}, \file{<errno.h>},
and \file{<stdlib.h>}. If the latter header file does not exist on
your system, it declares the functions \code{malloc()}, \code{free()}
and \code{realloc()} itself.
The compilation of an extension module depends on your system setup The compilation of an extension module depends on your system setup
and the intended use of the module; details are given in a later and the intended use of the module; details are given in a later
section. section.
Note: unless otherwise mentioned, all file references in this
document are relative to the Python toplevel directory
(the directory that contains the \file{configure} script).
\section{A Simple Example}
\section{A first look at the code} Let's create an extension module called \samp{spam}. Create a file
\samp{spammodule.c}. The first line of this file can be:
It is important not to be impressed by the size and complexity of \begin{verbatim}
the average extension module; much of this is straightforward #include "Python.h"
`boilerplate' code (starting right with the copyright notice)! \end{verbatim}
which pulls in the Python API (you can add a comment describing the
purpose of the module and a copyright notice if you like).
Let's skip the boilerplate and have a look at an interesting function Let's create a Python interface to the C library function
in \file{posixmodule.c} first: \code{system()}.\footnote{An interface for this function already
exists in the \code{posix} module --- it was chosen as a simple and
straightfoward example.} This function takes a zero-terminated
character string as argument and returns an integer. We will want
this function to be callable from Python as follows:
\begin{verbatim} \begin{verbatim}
static object * >>> import spam
posix_system(self, args) >>> status = spam.system("ls -l")
object *self; \end{verbatim}
object *args;
The next thing we add to our module file is the C function that will
be called when the Python expression \samp{spam.system(\var{string})}
is evaluated (well see shortly how it ends up being called):
\begin{verbatim}
static PyObject *
spam_system(self, args)
PyObject *self;
PyObject *args;
{ {
char *command; char *command;
int sts; int sts;
if (!getargs(args, "s", &command)) if (!PyArg_ParseTuple(args, "s", &command))
return NULL; return NULL;
sts = system(command); sts = system(command);
return mkvalue("i", sts); return Py_BuildValue("i", sts);
} }
\end{verbatim} \end{verbatim}
This is the prototypical top-level function in an extension module. There is a straightforward translation from the argument list in
It will be called (we'll see later how) when the Python program Python (here the single expression \code{"ls -l"}) to the arguments
executes statements like that are passed to the C function. The C function always has two
arguments, conventionally named \var{self} and \var{args}.
\begin{verbatim}
>>> import posix The \var{self} argument is only used when the C function implements a
>>> sts = posix.system('ls -l') builtin method --- this will be discussed later. In the example,
\end{verbatim} \var{self} will always be a \code{NULL} pointer, since we are defining
a function, not a method. (This is done so that the interpreter
There is a straightforward translation from the arguments to the call doesn't have to understand two different types of C functions.)
in Python (here the single expression \code{'ls -l'}) to the arguments that
are passed to the C function. The C function always has two The \var{args} argument will be a pointer to a Python tuple object
parameters, conventionally named \var{self} and \var{args}. The containing the arguments --- the length of the tuple will be the
\var{self} argument is used when the C function implements a builtin number of arguments. It is necessary to do full argument type
method---this will be discussed later. checking in each call, since otherwise the Python user would be able
In the example, \var{self} will always be a \code{NULL} pointer, since to cause the Python interpreter to crash (rather than raising an
we are defining a function, not a method (this is done so that the exception) by passing invalid arguments to a function in an extension
interpreter doesn't have to understand two different types of C module. Because argument checking and converting arguments to C are
functions). such common tasks, there's a general function in the Python
interpreter that combines them: \code{PyArg_ParseTuple()}. It uses a
The \var{args} parameter will be a pointer to a Python object, or template string to determine the types of the Python argument and the
\code{NULL} if the Python function/method was called without types of the C variables into which it should store the converted
arguments. It is necessary to do full argument type checking on each values (more about this later).
call, since otherwise the Python user would be able to cause the
Python interpreter to `dump core' by passing invalid arguments to a \code{PyArg_ParseTuple()} returns nonzero if all arguments have the
function in an extension module. Because argument checking and right type and its components have been stored in the variables whose
converting arguments to C are such common tasks, there's a general addresses are passed. It returns zero if an invalid argument was
function in the Python interpreter that combines them: passed. In the latter case it also raises an appropriate exception by
\code{getargs()}. It uses a template string to determine both the so the calling function can return \code{NULL} immediately. Here's
types of the Python argument and the types of the C variables into why:
which it should store the converted values.\footnote{There are
convenience macros \code{getnoarg()}, \code{getstrarg()},
\code{getintarg()}, etc., for many common forms of \code{getargs()} \section{Intermezzo: Errors and Exceptions}
templates. These are relics from the past; the recommended practice
is to call \code{getargs()} directly.} (More about this later.)
If \code{getargs()} returns nonzero, the argument list has the right
type and its components have been stored in the variables whose
addresses are passed. If it returns zero, an error has occurred. In
the latter case it has already raised an appropriate exception by so
the calling function should return \code{NULL} immediately --- see the
next section.
\section{Intermezzo: errors and exceptions}
An important convention throughout the Python interpreter is the An important convention throughout the Python interpreter is the
following: when a function fails, it should set an exception condition following: when a function fails, it should set an exception condition
and return an error value (often a \code{NULL} pointer). Exceptions and return an error value (usually a \code{NULL} pointer). Exceptions
are stored in a static global variable in \file{Python/errors.c}; if are stored in a static global variable inside the interpreter; if
this variable is \code{NULL} no exception has occurred. A second this variable is \code{NULL} no exception has occurred. A second
static global variable stores the `associated value' of the exception global variable stores the `associated value' of the exception
--- the second argument to \code{raise}. --- the second argument to \code{raise}. A third variable contains
the stack traceback in case the error originated in Python code.
The file \file{errors.h} declares a host of functions to set various These three variables are the C equivalents of the Python variables
types of exceptions. The most common one is \code{err_setstr()} --- \code{sys.exc_type}, \code{sys.exc_value} and \code{sys.exc_traceback}
its arguments are an exception object (e.g. \code{RuntimeError} --- --- see the section on module \code{sys} in the Library Reference
actually it can be any string object) and a C string indicating the Manual. It is important to know about them to understand how errors
cause of the error (this is converted to a string object and stored as are passed around.
the `associated value' of the exception). Another useful function is
\code{err_errno()}, which only takes an exception argument and The Python API defines a host of functions to set various types of
constructs the associated value by inspection of the (UNIX) global exceptions. The most common one is \code{PyErr_SetString()} --- its
variable errno. The most general function is \code{err_set()}, which arguments are an exception object (e.g. \code{PyExc_RuntimeError} ---
takes two object arguments, the exception and its associated value. actually it can be any object that is a legal exception indicator),
You don't need to \code{INCREF()} the objects passed to any of these and a C string indicating the cause of the error (this is converted to
a string object and stored as the `associated value' of the
exception). Another useful function is \code{PyErr_SetFromErrno()},
which only takes an exception argument and constructs the associated
value by inspection of the (\UNIX{}) global variable \code{errno}. The
most general function is \code{PyErr_SetObject()}, which takes two
object arguments, the exception and its associated value. You don't
need to \code{Py_INCREF()} the objects passed to any of these
functions. functions.
You can test non-destructively whether an exception has been set with You can test non-destructively whether an exception has been set with
\code{err_occurred()}. However, most code never calls \code{PyErr_Occurred()} --- this returns the current exception object,
\code{err_occurred()} to see whether an error occurred or not, but or \code{NULL} if no exception has occurred. Most code never needs to
relies on error return values from the functions it calls instead. call \code{PyErr_Occurred()} to see whether an error occurred or not,
but relies on error return values from the functions it calls instead.
When a function that calls another function detects that the called When a function that calls another function detects that the called
function fails, it should return an error value (e.g. \code{NULL} or function fails, it should return an error value (e.g. \code{NULL} or
\code{-1}) but not call one of the \code{err_*} functions --- one has \code{-1}). It shouldn't call one of the \code{PyErr_*} functions ---
already been called. The caller is then supposed to also return an one has already been called. The caller is then supposed to also
error indication to {\em its} caller, again {\em without} calling return an error indication to {\em its} caller, again {\em without}
\code{err_*()}, and so on --- the most detailed cause of the error was calling \code{PyErr_*()}, and so on --- the most detailed cause of the
already reported by the function that first detected it. Once the error was already reported by the function that first detected it.
error has reached Python's interpreter main loop, this aborts the Once the error has reached Python's interpreter main loop, this aborts
currently executing Python code and tries to find an exception handler the currently executing Python code and tries to find an exception
specified by the Python programmer. handler specified by the Python programmer.
(There are situations where a module can actually give a more detailed (There are situations where a module can actually give a more detailed
error message by calling another \code{err_*} function, and in such error message by calling another \code{PyErr_*} function, and in such
cases it is fine to do so. As a general rule, however, this is not cases it is fine to do so. As a general rule, however, this is not
necessary, and can cause information about the cause of the error to necessary, and can cause information about the cause of the error to
be lost: most operations can fail for a variety of reasons.) be lost: most operations can fail for a variety of reasons.)
To ignore an exception set by a function call that failed, the To ignore an exception set by a function call that failed, the exception
exception condition must be cleared explicitly by calling condition must be cleared explicitly by calling \code{PyErr_Clear()}.
\code{err_clear()}. The only time C code should call The only time C code should call \code{PyErr_Clear()} is if it doesn't
\code{err_clear()} is if it doesn't want to pass the error on to the want to pass the error on to the interpreter but wants to handle it
interpreter but wants to handle it completely by itself (e.g. by completely by itself (e.g. by trying something else or pretending
trying something else or pretending nothing happened). nothing happened).
Finally, the function \code{err_get()} gives you both error variables
{\em and clears them}. Note that even if an error occurred the second
one may be \code{NULL}. You have to \code{XDECREF()} both when you
are finished with them. I doubt you will need to use this function.
Note that a failing \code{malloc()} call must also be turned into an Note that a failing \code{malloc()} call must also be turned into an
exception --- the direct caller of \code{malloc()} (or exception --- the direct caller of \code{malloc()} (or
\code{realloc()}) must call \code{err_nomem()} and return a failure \code{realloc()}) must call \code{PyErr_NoMemory()} and return a
indicator itself. All the object-creating functions failure indicator itself. All the object-creating functions
(\code{newintobject()} etc.) already do this, so only if you call (\code{PyInt_FromLong()} etc.) already do this, so only if you call
\code{malloc()} directly this note is of importance. \code{malloc()} directly this note is of importance.
Also note that, with the important exception of \code{getargs()}, Also note that, with the important exception of
functions that return an integer status usually return \code{0} or a \code{PyArg_ParseTuple()}, functions that return an integer status
positive value for success and \code{-1} for failure. usually return \code{0} or a positive value for success and \code{-1}
for failure (like \UNIX{} system calls).
Finally, be careful about cleaning up garbage (making \code{XDECREF()} Finally, be careful about cleaning up garbage (making \code{Py_XDECREF()}
or \code{DECREF()} calls for objects you have already created) when or \code{Py_DECREF()} calls for objects you have already created) when
you return an error! you return an error!
The choice of which exception to raise is entirely yours. There are The choice of which exception to raise is entirely yours. There are
predeclared C objects corresponding to all built-in Python exceptions, predeclared C objects corresponding to all built-in Python exceptions,
e.g. \code{ZeroDevisionError} which you can use directly. Of course, e.g. \code{PyExc_ZeroDevisionError} which you can use directly. Of
you should chose exceptions wisely --- don't use \code{TypeError} to course, you should chose exceptions wisely --- don't use
mean that a file couldn't be opened (that should probably be \code{PyExc_TypeError} to mean that a file couldn't be opened (that
\code{IOError}). If anything's wrong with the argument list the should probably be \code{PyExc_IOError}). If something's wrong with
\code{getargs()} function raises \code{TypeError}. If you have an the argument list, the \code{PyArg_ParseTuple()} function usually
argument whose value which must be in a particular range or must raises \code{PyExc_TypeError}. If you have an argument whose value
satisfy other conditions, \code{ValueError} is appropriate. which must be in a particular range or must satisfy other conditions,
\code{PyExc_ValueError} is appropriate.
You can also define a new exception that is unique to your module. You can also define a new exception that is unique to your module.
For this, you usually declare a static object variable at the For this, you usually declare a static object variable at the
beginning of your file, e.g. beginning of your file, e.g.
\begin{verbatim} \begin{verbatim}
static object *FooError; static PyObject *SpamError;
\end{verbatim} \end{verbatim}
and initialize it in your module's initialization function and initialize it in your module's initialization function
(\code{initfoo()}) with a string object, e.g. (leaving out the error (\code{initspam()}) with a string object, e.g. (leaving out the error
checking for simplicity): checking for simplicity):
\begin{verbatim} \begin{verbatim}
void void
initfoo() initspam()
{ {
object *m, *d; PyObject *m, *d;
m = initmodule("foo", foo_methods); m = Py_InitModule("spam", spam_methods);
d = getmoduledict(m); d = PyModule_GetDict(m);
FooError = newstringobject("foo.error"); SpamError = PyString_FromString("spam.error");
dictinsert(d, "error", FooError); PyDict_SetItemString(d, "error", SpamError);
} }
\end{verbatim} \end{verbatim}
Note that the Python name for the exception object is \code{spam.error}
--- it is conventional for module and exception names to be spelled in
lower case. It is also conventional that the \emph{value} of the
exception object is the same as its name, e.g.\ the string
\code{"spam.error"}.
\section{Back to the example}
Going back to \code{posix_system()}, you should now be able to \section{Back to the Example}
understand this bit:
Going back to our example function, you should now be able to
understand this statement:
\begin{verbatim} \begin{verbatim}
if (!getargs(args, "s", &command)) if (!PyArg_ParseTuple(args, "s", &command))
return NULL; return NULL;
\end{verbatim} \end{verbatim}
It returns \code{NULL} (the error indicator for functions of this It returns \code{NULL} (the error indicator for functions returning
kind) if an error is detected in the argument list, relying on the object pointers) if an error is detected in the argument list, relying
exception set by \code{getargs()}. Otherwise the string value of the on the exception set by \code{PyArg_ParseTuple()}. Otherwise the
argument has been copied to the local variable \code{command} --- this string value of the argument has been copied to the local variable
is in fact just a pointer assignment and you are not supposed to \code{command}. This is a pointer assignment and you are not supposed
modify the string to which it points. to modify the string to which it points (so in ANSI C, the variable
\code{command} should properly be declared as \code{const char
If a function is called with multiple arguments, the argument list *command}).
(the argument \code{args}) is turned into a tuple. If it is called
without arguments, \code{args} is \code{NULL}. \code{getargs()} knows
about this; see later.
The next statement in \code{posix_system()} is a call to the C library The next statement is a call to the \UNIX{} function \code{system()},
function \code{system()}, passing it the string we just got from passing it the string we just got from \code{PyArg_ParseTuple()}:
\code{getargs()}:
\begin{verbatim} \begin{verbatim}
sts = system(command); sts = system(command);
\end{verbatim} \end{verbatim}
Finally, \code{posix.system()} must return a value: the integer status Our \code{spam.system()} function must return a value: the integer
returned by the C library \code{system()} function. This is done \code{sts} which contains the return value of the \UNIX{}
using the function \code{mkvalue()}, which is something like the \code{system()} function. This is done using the function
inverse of \code{getargs()}: it takes a format string and a variable \code{Py_BuildValue()}, which is something like the inverse of
number of C values and returns a new Python object. \code{PyArg_ParseTuple()}: it takes a format string and an arbitrary
number of C values, and returns a new Python object. More info on
\code{Py_BuildValue()} is given later.
\begin{verbatim} \begin{verbatim}
return mkvalue("i", sts); return Py_BuildValue("i", sts);
\end{verbatim} \end{verbatim}
In this case, it returns an integer object (yes, even integers are In this case, it will return an integer object. (Yes, even integers
objects on the heap in Python!). More info on \code{mkvalue()} is are objects on the heap in Python!)
given later.
If you had a function that returned no useful argument (a.k.a. a If you have a C function that returns no useful argument (a function
procedure), you would need this idiom: returning \code{void}), the corresponding Python function must return
\code{None}. You need this idiom to do so:
\begin{verbatim} \begin{verbatim}
INCREF(None); Py_INCREF(Py_None);
return None; return Py_None;
\end{verbatim} \end{verbatim}
\code{None} is a unique Python object representing `no value'. It \code{Py_None} is the C name for the special Python object
differs from \code{NULL}, which means `error' in most contexts. \code{None}. It is a genuine Python object (not a \code{NULL}
pointer, which means `error' in most contexts, as we have seen).
\section{The module's function table} \section{The Module's Method Table and Initialization Function}
I promised to show how I made the function \code{posix_system()} I promised to show how \code{spam_system()} is called from Python
callable from Python programs. This is shown later in programs. First, we need to list its name and address in a ``method
\file{Modules/posixmodule.c}: table'':
\begin{verbatim} \begin{verbatim}
static struct methodlist posix_methods[] = { static PyMethodDef spam_methods[] = {
... ...
{"system", posix_system}, {"system", spam_system, 1},
... ...
{NULL, NULL} /* Sentinel */ {NULL, NULL} /* Sentinel */
}; };
\end{verbatim}
Note the third entry (\samp{1}). This is a flag telling the
interpreter the calling convention to be used for the C function. It
should normally always be \samp{1}; a value of \samp{0} means that an
obsolete variant of \code{PyArg_ParseTuple()} is used.
The method table must be passed to the interpreter in the module's
initialization function (which should be the only non-\code{static}
item defined in the module file):
\begin{verbatim}
void void
initposix() initspam()
{ {
(void) initmodule("posix", posix_methods); (void) Py_InitModule("spam", spam_methods);
} }
\end{verbatim} \end{verbatim}
(The actual \code{initposix()} is somewhat more complicated, but many When the Python program imports module \code{spam} for the first time,
extension modules can be as simple as shown here.) When the Python \code{initspam()} is called. It calls \code{Py_InitModule()}, which
program first imports module \code{posix}, \code{initposix()} is creates a ``module object'' (which is inserted in the dictionary
called, which calls \code{initmodule()} with specific parameters. \code{sys.modules} under the key \code{"spam"}), and inserts built-in
This creates a `module object' (which is inserted in the table function objects into the newly created module based upon the table
\code{sys.modules} under the key \code{'posix'}), and adds (an array of \code{PyMethodDef} structures) that was passed as its
built-in-function objects to the newly created module based upon the second argument. \code{Py_InitModule()} returns a pointer to the
table (of type struct methodlist) that was passed as its second
parameter. The function \code{initmodule()} returns a pointer to the
module object that it creates (which is unused here). It aborts with module object that it creates (which is unused here). It aborts with
a fatal error if the module could not be initialized satisfactorily, a fatal error if the module could not be initialized satisfactorily,
so you don't need to check for errors. so the caller doesn't need to check for errors.
\section{Compilation and linkage} \section{Compilation and Linkage}
There are two more things to do before you can use your new extension There are two more things to do before you can use your new extension
module: compiling and linking it with the Python system. If you use module: compiling and linking it with the Python system. If you use
...@@ -342,13 +383,13 @@ about this. ...@@ -342,13 +383,13 @@ about this.
If you can't use dynamic loading, or if you want to make your module a If you can't use dynamic loading, or if you want to make your module a
permanent part of the Python interpreter, you will have to change the permanent part of the Python interpreter, you will have to change the
configuration setup and rebuild the interpreter. Luckily, in the 1.0 configuration setup and rebuild the interpreter. Luckily, this is
release this is very simple: just place your file (named very simple: just place your file (\file{spammodule.c} for example) in
\file{foomodule.c} for example) in the \file{Modules} directory, add a the \file{Modules} directory, add a line to the file
line to the file \file{Modules/Setup} describing your file: \file{Modules/Setup} describing your file:
\begin{verbatim} \begin{verbatim}
foo foomodule.o spam spammodule.o
\end{verbatim} \end{verbatim}
and rebuild the interpreter by running \code{make} in the toplevel and rebuild the interpreter by running \code{make} in the toplevel
...@@ -357,8 +398,15 @@ subdirectory, but then you must first rebuilt the \file{Makefile} ...@@ -357,8 +398,15 @@ subdirectory, but then you must first rebuilt the \file{Makefile}
there by running \code{make Makefile}. (This is necessary each time there by running \code{make Makefile}. (This is necessary each time
you change the \file{Setup} file.) you change the \file{Setup} file.)
If your module requires additional libraries to link with, these can
be listed on the line in the \file{Setup} file as well, for instance:
\begin{verbatim}
spam spammodule.o -lX11
\end{verbatim}
\section{Calling Python functions from C} \section{Calling Python Functions From C}
So far we have concentrated on making C functions callable from So far we have concentrated on making C functions callable from
Python. The reverse is also useful: calling Python functions from C. Python. The reverse is also useful: calling Python functions from C.
...@@ -378,211 +426,259 @@ Calling a Python function is easy. First, the Python program must ...@@ -378,211 +426,259 @@ Calling a Python function is easy. First, the Python program must
somehow pass you the Python function object. You should provide a somehow pass you the Python function object. You should provide a
function (or some other interface) to do this. When this function is function (or some other interface) to do this. When this function is
called, save a pointer to the Python function object (be careful to called, save a pointer to the Python function object (be careful to
\code{INCREF()} it!) in a global variable --- or whereever you see fit. \code{Py_INCREF()} it!) in a global variable --- or whereever you see fit.
For example, the following function might be part of a module For example, the following function might be part of a module
definition: definition:
\begin{verbatim} \begin{verbatim}
static object *my_callback = NULL; static PyObject *my_callback = NULL;
static object * static PyObject *
my_set_callback(dummy, arg) my_set_callback(dummy, arg)
object *dummy, *arg; PyObject *dummy, *arg;
{ {
XDECREF(my_callback); /* Dispose of previous callback */ Py_XDECREF(my_callback); /* Dispose of previous callback */
my_callback = arg; Py_XINCREF(arg); /* Add a reference to new callback */
XINCREF(my_callback); /* Remember new callback */ my_callback = arg; /* Remember new callback */
/* Boilerplate for "void" return */ /* Boilerplate to return "None" */
INCREF(None); Py_INCREF(Py_None);
return None; return Py_None;
} }
\end{verbatim} \end{verbatim}
This particular function doesn't do any typechecking on its argument The macros \code{Py_XINCREF()} and \code{Py_XDECREF()} increment/decrement
--- that will be done by \code{call_object()}, which is a bit late but
at least protects the Python interpreter from shooting itself in its
foot. (The problem with typechecking functions is that there are at
least five different Python object types that can be called, so the
test would be somewhat cumbersome.)
The macros \code{XINCREF()} and \code{XDECREF()} increment/decrement
the reference count of an object and are safe in the presence of the reference count of an object and are safe in the presence of
\code{NULL} pointers. More info on them in the section on Reference \code{NULL} pointers. More info on them in the section on Reference
Counts below. Counts below.
Later, when it is time to call the function, you call the C function Later, when it is time to call the function, you call the C function
\code{call_object()}. This function has two arguments, both pointers \code{PyEval_CallObject()}. This function has two arguments, both
to arbitrary Python objects: the Python function, and the argument pointers to arbitrary Python objects: the Python function, and the
list. The argument list must always be a tuple object, whose length argument list. The argument list must always be a tuple object, whose
is the number of arguments. To call the Python function with no length is the number of arguments. To call the Python function with
arguments, you must pass an empty tuple. For example: no arguments, pass an empty tuple; to call it with one argument, pass
a singleton tuple. \code{Py_BuildValue()} returns a tuple when its
format string consists of zero or more format codes between
parentheses. For example:
\begin{verbatim} \begin{verbatim}
object *arglist; int arg;
object *result; PyObject *arglist;
PyObject *result;
...
arg = 123;
... ...
/* Time to call the callback */ /* Time to call the callback */
arglist = newtupleobject(0); arglist = Py_BuildValue("(i)", arg);
result = call_object(my_callback, arglist); result = PyEval_CallObject(my_callback, arglist);
DECREF(arglist); Py_DECREF(arglist);
\end{verbatim} \end{verbatim}
\code{call_object()} returns a Python object pointer: this is \code{PyEval_CallObject()} returns a Python object pointer: this is
the return value of the Python function. \code{call_object()} is the return value of the Python function. \code{PyEval_CallObject()} is
`reference-count-neutral' with respect to its arguments. In the `reference-count-neutral' with respect to its arguments. In the
example a new tuple was created to serve as the argument list, which example a new tuple was created to serve as the argument list, which
is \code{DECREF()}-ed immediately after the call. is \code{Py_DECREF()}-ed immediately after the call.
The return value of \code{call_object()} is `new': either it is a The return value of \code{PyEval_CallObject()} is ``new'': either it
brand new object, or it is an existing object whose reference count is a brand new object, or it is an existing object whose reference
has been incremented. So, unless you want to save it in a global count has been incremented. So, unless you want to save it in a
variable, you should somehow \code{DECREF()} the result, even global variable, you should somehow \code{Py_DECREF()} the result,
(especially!) if you are not interested in its value. even (especially!) if you are not interested in its value.
Before you do this, however, it is important to check that the return Before you do this, however, it is important to check that the return
value isn't \code{NULL}. If it is, the Python function terminated by raising value isn't \code{NULL}. If it is, the Python function terminated by raising
an exception. If the C code that called \code{call_object()} is an exception. If the C code that called \code{PyEval_CallObject()} is
called from Python, it should now return an error indication to its called from Python, it should now return an error indication to its
Python caller, so the interpreter can print a stack trace, or the Python caller, so the interpreter can print a stack trace, or the
calling Python code can handle the exception. If this is not possible calling Python code can handle the exception. If this is not possible
or desirable, the exception should be cleared by calling or desirable, the exception should be cleared by calling
\code{err_clear()}. For example: \code{PyErr_Clear()}. For example:
\begin{verbatim} \begin{verbatim}
if (result == NULL) if (result == NULL)
return NULL; /* Pass error back */ return NULL; /* Pass error back */
/* Here maybe use the result */ ...use result...
DECREF(result); Py_DECREF(result);
\end{verbatim} \end{verbatim}
Depending on the desired interface to the Python callback function, Depending on the desired interface to the Python callback function,
you may also have to provide an argument list to \code{call_object()}. you may also have to provide an argument list to \code{PyEval_CallObject()}.
In some cases the argument list is also provided by the Python In some cases the argument list is also provided by the Python
program, through the same interface that specified the callback program, through the same interface that specified the callback
function. It can then be saved and used in the same manner as the function. It can then be saved and used in the same manner as the
function object. In other cases, you may have to construct a new function object. In other cases, you may have to construct a new
tuple to pass as the argument list. The simplest way to do this is to tuple to pass as the argument list. The simplest way to do this is to
call \code{mkvalue()}. For example, if you want to pass an integral call \code{Py_BuildValue()}. For example, if you want to pass an integral
event code, you might use the following code: event code, you might use the following code:
\begin{verbatim} \begin{verbatim}
object *arglist; PyObject *arglist;
... ...
arglist = mkvalue("(l)", eventcode); arglist = Py_BuildValue("(l)", eventcode);
result = call_object(my_callback, arglist); result = PyEval_CallObject(my_callback, arglist);
DECREF(arglist); Py_DECREF(arglist);
if (result == NULL) if (result == NULL)
return NULL; /* Pass error back */ return NULL; /* Pass error back */
/* Here maybe use the result */ /* Here maybe use the result */
DECREF(result); Py_DECREF(result);
\end{verbatim} \end{verbatim}
Note the placement of DECREF(argument) immediately after the call, Note the placement of \code{Py_DECREF(argument)} immediately after the call,
before the error check! Also note that strictly spoken this code is before the error check! Also note that strictly spoken this code is
not complete: \code{mkvalue()} may run out of memory, and this should not complete: \code{Py_BuildValue()} may run out of memory, and this should
be checked. be checked.
\section{Format strings for {\tt getargs()}} \section{Format Strings for {\tt PyArg_ParseTuple()}}
The \code{getargs()} function is declared in \file{modsupport.h} as The \code{PyArg_ParseTuple()} function is declared as follows:
follows:
\begin{verbatim} \begin{verbatim}
int getargs(object *arg, char *format, ...); int PyArg_ParseTuple(PyObject *arg, char *format, ...);
\end{verbatim} \end{verbatim}
The remaining arguments must be addresses of variables whose type is The \var{arg} argument must be a tuple object containing an argument
list passed from Python to a C function. The \var{format} argument
must be a format string, whose syntax is explained below. The
remaining arguments must be addresses of variables whose type is
determined by the format string. For the conversion to succeed, the determined by the format string. For the conversion to succeed, the
\var{arg} object must match the format and the format must be exhausted. \var{arg} object must match the format and the format must be
Note that while \code{getargs()} checks that the Python object really exhausted.
is of the specified type, it cannot check the validity of the
addresses of C variables provided in the call: if you make mistakes Note that while \code{PyArg_ParseTuple()} checks that the Python
there, your code will probably dump core. arguments have the required types, it cannot check the validity of the
addresses of C variables passed to the call: if you make mistakes
A non-empty format string consists of a single `format unit'. A there, your code will probably crash or at least overwrite random bits
format unit describes one Python object; it is usually a single in memory. So be careful!
character or a parenthesized sequence of format units. The type of a
format units is determined from its first character, the `format A format string consists of zero or more ``format units''. A format
letter': unit describes one Python object; it is usually a single character or
a parenthesized sequence of format units. With a few exceptions, a
format unit that is not a parenthesized sequence normally corresponds
to a single address argument to \code{PyArg_ParseTuple()}. In the
following description, the quoted form is the format unit; the entry
in (round) parentheses is the Python object type that matches the
format unit; and the entry in [square] brackets is the type of the C
variable(s) whose address should be passed. (Use the \samp{\&}
operator to pass a variable's address.)
\begin{description} \begin{description}
\item[\samp{s} (string)] \item[\samp{s} (string) [char *]]
The Python object must be a string object. The C argument must be a Convert a Python string to a C pointer to a character string. You
\code{(char**)} (i.e. the address of a character pointer), and a pointer must not provide storage for the string itself; a pointer to an
to the C string contained in the Python object is stored into it. You existing string is stored into the character pointer variable whose
must not provide storage to store the string; a pointer to an existing address you pass. The C string is null-terminated. The Python string
string is stored into the character pointer variable whose address you must not contain embedded null bytes; if it does, a \code{TypeError}
pass. If the next character in the format string is \samp{\#}, exception is raised.
another C argument of type \code{(int*)} must be present, and the
length of the Python string (not counting the trailing zero byte) is \item[\samp{s\#} (string) {[char *, int]}]
stored into it. This variant on \code{'s'} stores into two C variables, the first one
a pointer to a character string, the second one its length. In this
\item[\samp{z} (string or zero, i.e. \code{NULL})] case the Python string may contain embedded null bytes.
Like \samp{s}, but the object may also be None. In this case the
string pointer is set to \code{NULL} and if a \samp{\#} is present the \item[\samp{z} (string or \code{None}) {[char *]}]
size is set to 0. Like \samp{s}, but the Python object may also be \code{None}, in which
case the C pointer is set to \code{NULL}.
\item[\samp{b} (byte, i.e. char interpreted as tiny int)]
The object must be a Python integer. The C argument must be a \item[\samp{z\#} (string or \code{None}) {[char *, int]}]
\code{(char*)}. This is to \code{'s\#'} as \code{'z'} is to \code{'s'}.
\item[\samp{h} (half, i.e. short)] \item[\samp{b} (integer) {[char]}]
The object must be a Python integer. The C argument must be a Convert a Python integer to a tiny int, stored in a C \code{char}.
\code{(short*)}.
\item[\samp{h} (integer) {[short int]}]
\item[\samp{i} (int)] Convert a Python integer to a C \code{short int}.
The object must be a Python integer. The C argument must be an
\code{(int*)}. \item[\samp{i} (integer) {[int]}]
Convert a Python integer to a plain C \code{int}.
\item[\samp{l} (long)]
The object must be a (plain!) Python integer. The C argument must be \item[\samp{l} (integer) {[long int]}]
a \code{(long*)}. Convert a Python integer to a C \code{long int}.
\item[\samp{c} (char)] \item[\samp{c} (string of length 1) {[char]}]
The Python object must be a string of length 1. The C argument must Convert a Python character, represented as a string of length 1, to a
be a \code{(char*)}. (Don't pass an \code{(int*)}!) C \code{char}.
\item[\samp{f} (float)] \item[\samp{f} (float) {[float]}]
The object must be a Python int or float. The C argument must be a Convert a Python floating point number to a C \code{float}.
\code{(float*)}.
\item[\samp{d} (float) {[double]}]
\item[\samp{d} (double)] Convert a Python floating point number to a C \code{double}.
The object must be a Python int or float. The C argument must be a
\code{(double*)}. \item[\samp{O} (object) {[PyObject *]}]
Store a Python object (without any conversion) in a C object pointer.
\item[\samp{S} (string object)] The C program thus receives the actual object that was passed. The
The object must be a Python string. The C argument must be an object's reference count is not increased. The pointer stored is not
\code{(object**)} (i.e. the address of an object pointer). The C \code{NULL}.
program thus gets back the actual string object that was passed, not
just a pointer to its array of characters and its size as for format \item[\samp{O!} (object) {[\var{typeobject}, PyObject *]}]
character \samp{s}. The reference count of the object has not been Store a Python object in a C object pointer. This is similar to
increased. \samp{O}, but takes two C arguments: the first is the address of a
Python type object, the second is the address of the C variable (of
\item[\samp{O} (object)] type \code{PyObject *}) into which the object pointer is stored.
The object can be any Python object, including None, but not If the Python object does not have the required type, a
\code{NULL}. The C argument must be an \code{(object**)}. This can be \code{TypeError} exception is raised.
used if an argument list must contain objects of a type for which no
format letter exist: the caller must then check that it has the right \item[\samp{O\&} (object) {[\var{converter}, \var{anything}]}]
type. The reference count of the object has not been increased. Convert a Python object to a C variable through a \var{converter}
function. This takes two arguments: the first is a function, the
\item[\samp{(} (tuple)] second is the address of a C variable (of arbitrary type), converted
The object must be a Python tuple. Following the \samp{(} character to \code{void *}. The \var{converter} function in turn is called as
in the format string must come a number of format units describing the follows:
elements of the tuple, followed by a \samp{)} character. Tuple
format units may be nested. (There are no exceptions for empty and \code{\var{status} = \var{converter}(\var{object}, \var{address});}
singleton tuples; \samp{()} specifies an empty tuple and \samp{(i)} a
singleton of one integer. Normally you don't want to use the latter, where \var{object} is the Python object to be converted and
since it is hard for the Python user to specify. \var{address} is the \code{void *} argument that was passed to
\code{PyArg_ConvertTuple()}. The returned \var{status} should be
\code{1} for a successful conversion and \code{0} if the conversion
has failed. When the conversion fails, the \var{converter} function
should raise an exception.
\item[\samp{S} (string) {[PyStringObject *]}]
Like \samp{O} but raises a \code{TypeError} exception that the object
is a string object. The C variable may also be declared as
\code{PyObject *}.
\item[\samp{(\var{items})} (tuple) {[\var{matching-items}]}]
The object must be a Python tuple whose length is the number of format
units in \var{items}. The C arguments must correspond to the
individual format units in \var{items}. Format units for tuples may
be nested.
\end{description} \end{description}
More format characters will probably be added as the need arises. It It is possible to pass Python long integers where integers are
should (but currently isn't) be allowed to use Python long integers requested; however no proper range checking is done -- the most
whereever integers are expected, and perform a range check. (A range significant bits are silently truncated when the receiving field is
check is in fact always necessary for the \samp{b}, \samp{h} and too small to receive the value (actually, the semantics are inherited
\samp{i} format letters, but this is currently not implemented.) from downcasts in C --- your milage may vary).
A few other characters have a meaning in a format string. These may
not occur inside nested parentheses. They are:
\begin{description}
\item[\samp{|}]
Indicates that the remaining arguments in the Python argument list are
optional. The C variables corresponding to optional arguments should
be initialized to their default value --- when an optional argument is
not specified, the \code{PyArg_ParseTuple} does not touch the contents
of the corresponding C variable(s).
\item[\samp{:}]
The list of format units ends here; the string after the colon is used
as the function name in error messages (the ``associated value'' of
the exceptions that \code{PyArg_ParseTuple} raises).
\item[\samp{;}]
The list of format units ends here; the string after the colon is used
as the error message \emph{instead} of the default error message.
Clearly, \samp{:} and \samp{;} mutually exclude each other.
\end{description}
Some example calls: Some example calls:
...@@ -593,186 +689,444 @@ Some example calls: ...@@ -593,186 +689,444 @@ Some example calls:
char *s; char *s;
int size; int size;
ok = getargs(args, ""); /* No arguments */ ok = PyArg_ParseTuple(args, ""); /* No arguments */
/* Python call: f() */ /* Python call: f() */
ok = getargs(args, "s", &s); /* A string */ ok = PyArg_ParseTuple(args, "s", &s); /* A string */
/* Possible Python call: f('whoops!') */ /* Possible Python call: f('whoops!') */
ok = getargs(args, "(lls)", &k, &l, &s); /* Two longs and a string */ ok = PyArg_ParseTuple(args, "lls", &k, &l, &s); /* Two longs and a string */
/* Possible Python call: f(1, 2, 'three') */ /* Possible Python call: f(1, 2, 'three') */
ok = getargs(args, "((ii)s#)", &i, &j, &s, &size); ok = PyArg_ParseTuple(args, "(ii)s#", &i, &j, &s, &size);
/* A pair of ints and a string, whose size is also returned */ /* A pair of ints and a string, whose size is also returned */
/* Possible Python call: f(1, 2, 'three') */ /* Possible Python call: f(1, 2, 'three') */
{
char *file;
char *mode = "r";
int bufsize = 0;
ok = PyArg_ParseTuple(args, "s|si", &file, &mode, &bufsize);
/* A string, and optionally another string and an integer */
/* Possible Python calls:
f('spam')
f('spam', 'w')
f('spam', 'wb', 100000) */
}
{ {
int left, top, right, bottom, h, v; int left, top, right, bottom, h, v;
ok = getargs(args, "(((ii)(ii))(ii))", ok = PyArg_ParseTuple(args, "((ii)(ii))(ii)",
&left, &top, &right, &bottom, &h, &v); &left, &top, &right, &bottom, &h, &v);
/* A rectangle and a point */ /* A rectangle and a point */
/* Possible Python call: /* Possible Python call:
f( ((0, 0), (400, 300)), (10, 10)) */ f(((0, 0), (400, 300)), (10, 10)) */
} }
\end{verbatim} \end{verbatim}
Note that the `top level' of a non-empty format string must consist of
a single unit; strings like \samp{is} and \samp{(ii)s\#} are not valid
format strings. (But \samp{s\#} is.) If you have multiple arguments,
the format must therefore always be enclosed in parentheses, as in the
examples \samp{((ii)s\#)} and \samp{(((ii)(ii))(ii)}. (The current
implementation does not complain when more than one unparenthesized
format unit is given. Sorry.)
The \code{getargs()} function does not support variable-length \section{The {\tt Py_BuildValue()} Function}
argument lists. In simple cases you can fake these by trying several
calls to This function is the counterpart to \code{PyArg_ParseTuple()}. It is
\code{getargs()} until one succeeds, but you must take care to call declared as follows:
\code{err_clear()} before each retry. For example:
\begin{verbatim} \begin{verbatim}
static object *my_method(self, args) object *self, *args; { PyObject *Py_BuildValue(char *format, ...);
int i, j, k;
if (getargs(args, "(ii)", &i, &j)) {
k = 0; /* Use default third argument */
}
else {
err_clear();
if (!getargs(args, "(iii)", &i, &j, &k))
return NULL;
}
/* ... use i, j and k here ... */
INCREF(None);
return None;
}
\end{verbatim} \end{verbatim}
(It is possible to think of an extension to the definition of format It recognizes a set of format units similar to the ones recognized by
strings to accommodate this directly, e.g. placing a \samp{|} in a \code{PyArg_ParseTuple()}, but the arguments (which are input to the
tuple might specify that the remaining arguments are optional. function, not output) must not be pointers, just values. It returns a
\code{getargs()} should then return one more than the number of new Python object, suitable for returning from a C function called
variables stored into.) from Python.
One difference with \code{PyArg_ParseTuple()}: while the latter
requires its first argument to be a tuple (since Python argument lists
are always represented as tuples internally), \code{BuildValue()} does
not always build a tuple. It builds a tuple only if its format string
contains two or more format units. If the format string is empty, it
returns \code{None}; if it contains exactly one format unit, it
returns whatever object is described by that format unit. To force it
to return a tuple of size 0 or one, parenthesize the format string.
In the following description, the quoted form is the format unit; the
entry in (round) parentheses is the Python object type that the format
unit will return; and the entry in [square] brackets is the type of
the C value(s) to be passed.
The characters space, tab, colon and comma are ignored in format
strings (but not within format units such as \samp{s\#}). This can be
used to make long format strings a tad more readable.
\begin{description}
\item[\samp{s} (string) {[char *]}]
Convert a null-terminated C string to a Python object. If the C
string pointer is \code{NULL}, \code{None} is returned.
\item[\samp{s\#} (string) {[char *, int]}]
Convert a C string and its length to a Python object. If the C string
pointer is \code{NULL}, the length is ignored and \code{None} is
returned.
\item[\samp{z} (string or \code{None}) {[char *]}]
Same as \samp{s}.
\item[\samp{z\#} (string or \code{None}) {[char *, int]}]
Same as \samp{s\#}.
\item[\samp{i} (integer) {[int]}]
Convert a plain C \code{int} to a Python integer object.
Advanced users note: If you set the `varargs' flag in the method list \item[\samp{b} (integer) {[char]}]
for a function, the argument will always be a tuple (the `raw argument Same as \samp{i}.
list'). In this case you must enclose single and empty argument lists
in parentheses, e.g. \samp{(s)} and \samp{()}.
\item[\samp{h} (integer) {[short int]}]
Same as \samp{i}.
\section{The {\tt mkvalue()} function} \item[\samp{l} (integer) {[long int]}]
Convert a C \code{long int} to a Python integer object.
\item[\samp{c} (string of length 1) {[char]}]
Convert a C \code{int} representing a character to a Python string of
length 1.
\item[\samp{d} (float) {[double]}]
Convert a C \code{double} to a Python floating point number.
\item[\samp{f} (float) {[float]}]
Same as \samp{d}.
\item[\samp{O} (object) {[PyObject *]}]
Pass a Python object untouched (except for its reference count, which
is incremented by one). If the object passed in is a \code{NULL}
pointer, it is assumed that this was caused because the call producing
the argument found an error and set an exception. Therefore,
\code{Py_BuildValue()} will return \code{NULL} but won't raise an
exception. If no exception has been raised yet,
\code{PyExc_SystemError} is set.
\item[\samp{S} (object) {[PyObject *]}]
Same as \samp{O}.
\item[\samp{O\&} (object) {[\var{converter}, \var{anything}]}]
Convert \var{anything} to a Python object through a \var{converter}
function. The function is called with \var{anything} (which should be
compatible with \code{void *}) as its argument and should return a
``new'' Python object, or \code{NULL} if an error occurred.
\item[\samp{(\var{items})} (tuple) {[\var{matching-items}]}]
Convert a sequence of C values to a Python tuple with the same number
of items.
\item[\samp{[\var{items}]} (list) {[\var{matching-items}]}]
Convert a sequence of C values to a Python list with the same number
of items.
\item[\samp{\{\var{items}\}} (dictionary) {[\var{matching-items}]}]
Convert a sequence of C values to a Python dictionary. Each pair of
consecutive C values adds one item to the dictionary, serving as key
and value, respectively.
\end{description}
This function is the counterpart to \code{getargs()}. It is declared If there is an error in the format string, the
in \file{Include/modsupport.h} as follows: \code{PyExc_SystemError} exception is raised and \code{NULL} returned.
Examples (to the left the call, to the right the resulting Python value):
\begin{verbatim}
Py_BuildValue("") None
Py_BuildValue("i", 123) 123
Py_BuildValue("ii", 123, 456) (123, 456)
Py_BuildValue("s", "hello") 'hello'
Py_BuildValue("ss", "hello", "world") ('hello', 'world')
Py_BuildValue("s#", "hello", 4) 'hell'
Py_BuildValue("()") ()
Py_BuildValue("(i)", 123) (123,)
Py_BuildValue("(ii)", 123, 456) (123, 456)
Py_BuildValue("(i,i)", 123, 456) (123, 456)
Py_BuildValue("[i,i]", 123, 456) [123, 456]
Py_BuildValue("{s:i,s:i}", "abc", 123, "def", 456)
{'abc': 123, 'def': 456}
Py_BuildValue("((ii)(ii)) (ii)", 1, 2, 3, 4, 5, 6)
(((1, 2), (3, 4)), (5, 6))
\end{verbatim}
\section{Reference Counts}
\subsection{Introduction}
In languages like C or \Cpp{}, the programmer is responsible for
dynamic allocation and deallocation of memory on the heap. In C, this
is done using the functions \code{malloc()} and \code{free()}. In
\Cpp{}, the operators \code{new} and \code{delete} are used with
essentially the same meaning; they are actually implemented using
\code{malloc()} and \code{free()}, so we'll restrict the following
discussion to the latter.
Every block of memory allocated with \code{malloc()} should eventually
be returned to the pool of available memory by exactly one call to
\code{free()}. It is important to call \code{free()} at the right
time. If a block's address is forgotten but \code{free()} is not
called for it, the memory it occupies cannot be reused until the
program terminates. This is called a \dfn{memory leak}. On the other
hand, if a program calls \code{free()} for a block and then continues
to use the block, it creates a conflict with re-use of the block
through another \code{malloc()} call. This is called \dfn{using freed
memory} has the same bad consequences as referencing uninitialized
data --- core dumps, wrong results, mysterious crashes.
Common causes of memory leaks are unusual paths through the code. For
instance, a function may allocate a block of memory, do some
calculation, and then free the block again. Now a change in the
requirements for the function may add a test to the calculation that
detects an error condition and can return prematurely from the
function. It's easy to forget to free the allocated memory block when
taking this premature exit, especially when it is added later to the
code. Such leaks, once introduced, often go undetected for a long
time: the error exit is taken only in a small fraction of all calls,
and most modern machines have plenty of virtual memory, so the leak
only becomes apparent in a long-running process that uses the leaking
function frequently. Therefore, it's important to prevent leaks from
happening by having a coding convention or strategy that minimizes
this kind of errors.
Since Python makes heavy use of \code{malloc()} and \code{free()}, it
needs a strategy to avoid memory leaks as well as the use of freed
memory. The chosen method is called \dfn{reference counting}. The
principle is simple: every object contains a counter, which is
incremented when a reference to the object is stored somewhere, and
which is decremented when a reference to it is deleted. When the
counter reaches zero, the last reference to the object has been
deleted and the object is freed.
An alternative strategy is called \dfn{automatic garbage collection}.
(Sometimes, reference counting is also referred to as a garbage
collection strategy, hence my use of ``automatic'' to distinguish the
two.) The big advantage of automatic garbage collection is that the
user doesn't need to call \code{free()} explicitly. (Another claimed
advantage is an improvement in speed or memory usage --- this is no
hard fact however.) The disadvantage is that for C, there is no
truly portable automatic garbage collector, while reference counting
can be implemented portably (as long as the functions \code{malloc()}
and \code{free()} are available --- which the C Standard guarantees).
Maybe some day a sufficiently portable automatic garbage collector
will be available for C. Until then, we'll have to live with
reference counts.
\subsection{Reference Counting in Python}
There are two macros, \code{Py_INCREF(x)} and \code{Py_DECREF(x)},
which handle the incrementing and decrementing of the reference count.
\code{Py_DECREF()} also frees the object when the count reaches zero.
For flexibility, it doesn't call \code{free()} directly --- rather, it
makes a call through a function pointer in the object's \dfn{type
object}. For this purpose (and others), every object also contains a
pointer to its type object.
The big question now remains: when to use \code{Py_INCREF(x)} and
\code{Py_DECREF(x)}? Let's first introduce some terms. Nobody
``owns'' an object; however, you can \dfn{own a reference} to an
object. An object's reference count is now defined as the number of
owned references to it. The owner of a reference is responsible for
calling \code{Py_DECREF()} when the reference is no longer needed.
Ownership of a reference can be transferred. There are three ways to
dispose of an owned reference: pass it on, store it, or call
\code{Py_DECREF()}. Forgetting to dispose of an owned reference creates
a memory leak.
It is also possible to \dfn{borrow}\footnote{The metaphor of
``borrowing'' a reference is not completely correct: the owner still
has a copy of the reference.} a reference to an object. The borrower
of a reference should not call \code{Py_DECREF()}. The borrower must
not hold on to the object longer than the owner from which it was
borrowed. Using a borrowed reference after the owner has disposed of
it risks using freed memory and should be avoided
completely.\footnote{Checking that the reference count is at least 1
\strong{does not work} --- the reference count itself could be in
freed memory and may thus be reused for another object!}
The advantage of borrowing over owning a reference is that you don't
need to take care of disposing of the reference on all possible paths
through the code --- in other words, with a borrowed reference you
don't run the risk of leaking when a premature exit is taken. The
disadvantage of borrowing over leaking is that there are some subtle
situations where in seemingly correct code a borrowed reference can be
used after the owner from which it was borrowed has in fact disposed
of it.
A borrowed reference can be changed into an owned reference by calling
\code{Py_INCREF()}. This does not affect the status of the owner from
which the reference was borrowed --- it creates a new owned reference,
and gives full owner responsibilities (i.e., the new owner must
dispose of the reference properly, as well as the previous owner).
\subsection{Ownership Rules}
Whenever an object reference is passed into or out of a function, it
is part of the function's interface specification whether ownership is
transferred with the reference or not.
Most functions that return a reference to an object pass on ownership
with the reference. In particular, all functions whose function it is
to create a new object, e.g.\ \code{PyInt_FromLong()} and
\code{Py_BuildValue()}, pass ownership to the receiver. Even if in
fact, in some cases, you don't receive a reference to a brand new
object, you still receive ownership of the reference. For instance,
\code{PyInt_FromLong()} maintains a cache of popular values and can
return a reference to a cached item.
Many functions that extract objects from other objects also transfer
ownership with the reference, for instance
\code{PyObject_GetAttrString()}. The picture is less clear, here,
however, since a few common routines are exceptions:
\code{PyTuple_GetItem()}, \code{PyList_GetItem()} and
\code{PyDict_GetItem()} (and \code{PyDict_GetItemString()}) all return
references that you borrow from the tuple, list or dictionary.
The function \code{PyImport_AddModule()} also returns a borrowed
reference, even though it may actually create the object it returns:
this is possible because an owned reference to the object is stored in
\code{sys.modules}.
When you pass an object reference into another function, in general,
the function borrows the reference from you --- if it needs to store
it, it will use \code{Py_INCREF()} to become an independent owner.
There are exactly two important exceptions to this rule:
\code{PyTuple_SetItem()} and \code{PyList_SetItem()}. These functions
take over ownership of the item passed to them --- even if they fail!
(Note that \code{PyDict_SetItem()} and friends don't take over
ownership --- they are ``normal''.)
When a C function is called from Python, it borrows references to its
arguments from the caller. The caller owns a reference to the object,
so the borrowed reference's lifetime is guaranteed until the function
returns. Only when such a borrowed reference must be stored or passed
on, it must be turned into an owned reference by calling
\code{Py_INCREF()}.
The object reference returned from a C function that is called from
Python must be an owned reference --- ownership is tranferred from the
function to its caller.
\subsection{Thin Ice}
There are a few situations where seemingly harmless use of a borrowed
reference can lead to problems. These all have to do with implicit
invocations of the interpreter, which can cause the owner of a
reference to dispose of it.
The first and most important case to know about is using
\code{Py_DECREF()} on an unrelated object while borrowing a reference
to a list item. For instance:
\begin{verbatim}
bug(PyObject *list) {
PyObject *item = PyList_GetItem(list, 0);
PyList_SetItem(list, 1, PyInt_FromLong(0L));
PyObject_Print(item, stdout, 0); /* BUG! */
}
\end{verbatim}
This function first borrows a reference to \code{list[0]}, then
replaces \code{list[1]} with the value \code{0}, and finally prints
the borrowed reference. Looks harmless, right? But it's not!
Let's follow the control flow into \code{PyList_SetItem()}. The list
owns references to all its items, so when item 1 is replaced, it has
to dispose of the original item 1. Now let's suppose the original
item 1 was an instance of a user-defined class, and let's further
suppose that the class defined a \code{__del__()} method. If this
class instance has a reference count of 1, disposing of it will call
its \code{__del__()} method.
Since it is written in Python, the \code{__del__()} method can execute
arbitrary Python code. Could it perhaps do something to invalidate
the reference to \code{item} in \code{bug()}? You bet! Assuming that
the list passed into \code{bug()} is accessible to the
\code{__del__()} method, it could execute a statement to the effect of
\code{del list[0]}, and assuming this was the last reference to that
object, it would free the memory associated with it, thereby
invalidating \code{item}.
The solution, once you know the source of the problem, is easy:
temporarily increment the reference count. The correct version of the
function reads:
\begin{verbatim} \begin{verbatim}
object *mkvalue(char *format, ...); no_bug(PyObject *list) {
PyObject *item = PyList_GetItem(list, 0);
Py_INCREF(item);
PyList_SetItem(list, 1, PyInt_FromLong(0L));
PyObject_Print(item, stdout, 0);
Py_DECREF(item);
}
\end{verbatim} \end{verbatim}
It supports exactly the same format letters as \code{getargs()}, but This is a true story. An older version of Python contained variants
the arguments (which are input to the function, not output) must not of this bug and someone spent a considerable amount of time in a C
be pointers, just values. If a byte, short or float is passed to a debugger to figure out why his \code{__del__()} methods would fail...
varargs function, it is widened by the compiler to int or double, so
\samp{b} and \samp{h} are treated as \samp{i} and \samp{f} is The second case of problems with a borrowed reference is a variant
treated as \samp{d}. \samp{S} is treated as \samp{O}, \samp{s} is involving threads. Normally, multiple threads in the Python
treated as \samp{z}. \samp{z\#} and \samp{s\#} are supported: a interpreter can't get in each other's way, because there is a global
second argument specifies the length of the data (negative means use lock protecting Python's entire object space. However, it is possible
\code{strlen()}). \samp{S} and \samp{O} add a reference to their to temporarily release this lock using the macro
argument (so you should \code{DECREF()} it if you've just created it \code{Py_BEGIN_ALLOW_THREADS}, and to re-acquire it using
and aren't going to use it again). \code{Py_END_ALLOW_THREADS}. This is common around blocking I/O
calls, to let other threads use the CPU while waiting for the I/O to
If the argument for \samp{O} or \samp{S} is a \code{NULL} pointer, it is complete. Obviously, the following function has the same problem as
assumed that this was caused because the call producing the argument the previous one:
found an error and set an exception. Therefore, \code{mkvalue()} will
return \code{NULL} but won't set an exception if one is already set.
If no exception is set, \code{SystemError} is set.
If there is an error in the format string, the \code{SystemError}
exception is set, since it is the calling C code's fault, not that of
the Python user who sees the exception.
Example:
\begin{verbatim} \begin{verbatim}
return mkvalue("(ii)", 0, 0); bug(PyObject *list) {
PyObject *item = PyList_GetItem(list, 0);
Py_BEGIN_ALLOW_THREADS
...some blocking I/O call...
Py_END_ALLOW_THREADS
PyObject_Print(item, stdout, 0); /* BUG! */
}
\end{verbatim} \end{verbatim}
returns a tuple containing two zeros. (Outer parentheses in the \subsection{NULL Pointers}
format string are actually superfluous, but you can use them for
compatibility with \code{getargs()}, which requires them if more than In general, functions that take object references as arguments don't
one argument is expected.) expect you to pass them \code{NULL} pointers, and will dump core (or
cause later core dumps) if you do so. Functions that return object
references generally return \code{NULL} only to indicate that an
\section{Reference counts} exception occurred. The reason for not testing for \code{NULL}
arguments is that functions often pass the objects they receive on to
Here's a useful explanation of \code{INCREF()} and \code{DECREF()} other function --- if each function were to test for \code{NULL},
(after an original by Sjoerd Mullender). there would be a lot of redundant tests and the code would run slower.
Use \code{XINCREF()} or \code{XDECREF()} instead of \code{INCREF()} or It is better to test for \code{NULL} only at the ``source'', i.e.\
\code{DECREF()} when the argument may be \code{NULL} --- the versions when a pointer that may be \code{NULL} is received, e.g.\ from
without \samp{X} are faster but wull dump core when they encounter a \code{malloc()} or from a function that may raise an exception.
\code{NULL} pointer.
The macros \code{Py_INCREF()} and \code{Py_DECREF()}
The basic idea is, if you create an extra reference to an object, you don't check for \code{NULL} pointers --- however, their variants
must \code{INCREF()} it, if you throw away a reference to an object, \code{Py_XINCREF()} and \code{Py_XDECREF()} do.
you must \code{DECREF()} it. Functions such as
\code{newstringobject()}, \code{newsizedstringobject()}, The macros for checking for a particular object type
\code{newintobject()}, etc. create a reference to an object. If you (\code{Py\var{type}_Check()}) don't check for \code{NULL} pointers ---
want to throw away the object thus created, you must use again, there is much code that calls several of these in a row to test
\code{DECREF()}. an object against various different expected types, and this would
generate redundant tests. There are no variants with \code{NULL}
If you put an object into a tuple or list using \code{settupleitem()} checking.
or \code{setlistitem()}, the idea is that you usually don't want to
keep a reference of your own around, so Python does not The C function calling mechanism guarantees that the argument list
\code{INCREF()} the elements. It does \code{DECREF()} the old value. passed to C functions (\code{args} in the examples) is never
This means that if you put something into such an object using the \code{NULL} --- in fact it guarantees that it is always a tuple.%
functions Python provides for this, you must \code{INCREF()} the \footnote{These guarantees don't hold when you use the ``old'' style
object if you also want to keep a separate reference to the object around. calling convention --- this is still found in much existing code.}
Also, if you replace an element, you should \code{INCREF()} the old
element first if you want to keep it. If you didn't \code{INCREF()} It is a severe error to ever let a \code{NULL} pointer ``escape'' to
it before you replaced it, you are not allowed to look at it anymore, the Python user.
since it may have been freed.
Returning an object to Python (i.e. when your C function returns) \section{Writing Extensions in \Cpp{}}
creates a reference to an object, but it does not change the reference
count. When your code does not keep another reference to the object,
you should not \code{INCREF()} or \code{DECREF()} it (assuming it is a
newly created object). When you do keep a reference around, you
should \code{INCREF()} the object. Also, when you return a global
object such as \code{None}, you should \code{INCREF()} it.
If you want to return a tuple, you should consider using
\code{mkvalue()}. This function creates a new tuple with a reference
count of 1 which you can return. If any of the elements you put into
the tuple are objects (format codes \samp{O} or \samp{S}), they
are \code{INCREF()}'ed by \code{mkvalue()}. If you don't want to keep
references to those elements around, you should \code{DECREF()} them
after having called \code{mkvalue()}.
Usually you don't have to worry about arguments. They are
\code{INCREF()}'ed before your function is called and
\code{DECREF()}'ed after your function returns. When you keep a
reference to an argument, you should \code{INCREF()} it and
\code{DECREF()} when you throw it away. Also, when you return an
argument, you should \code{INCREF()} it, because returning the
argument creates an extra reference to it.
If you use \code{getargs()} to parse the arguments, you can get a
reference to an object (by using \samp{O} in the format string). This
object was not \code{INCREF()}'ed, so you should not \code{DECREF()}
it. If you want to keep the object, you must \code{INCREF()} it
yourself.
If you create your own type of objects, you should use \code{NEWOBJ()}
to create the object. This sets the reference count to 1. If you
want to throw away the object, you should use \code{DECREF()}. When
the reference count reaches zero, your type's \code{dealloc()}
function is called. In it, you should \code{DECREF()} all object to
which you keep references in your object, but you should not use
\code{DECREF()} on your object. You should use \code{DEL()} instead.
\section{Writing extensions in \Cpp{}}
It is possible to write extension modules in \Cpp{}. Some restrictions It is possible to write extension modules in \Cpp{}. Some restrictions
apply: since the main program (the Python interpreter) is compiled and apply: since the main program (the Python interpreter) is compiled and
...@@ -782,8 +1136,9 @@ indirectly (i.e. via function pointers) by the Python interpreter will ...@@ -782,8 +1136,9 @@ indirectly (i.e. via function pointers) by the Python interpreter will
have to be declared using \code{extern "C"}; this applies to all have to be declared using \code{extern "C"}; this applies to all
`methods' as well as to the module's initialization function. `methods' as well as to the module's initialization function.
It is unnecessary to enclose the Python header files in It is unnecessary to enclose the Python header files in
\code{extern "C" \{...\}} --- they do this already. \code{extern "C" \{...\}} --- they use this form already if the symbol
\samp{__cplusplus} is defined (all recent C++ compilers define this
symbol).
\chapter{Embedding Python in another application} \chapter{Embedding Python in another application}
...@@ -797,16 +1152,16 @@ interpreter to run some Python code. ...@@ -797,16 +1152,16 @@ interpreter to run some Python code.
So if you are embedding Python, you are providing your own main So if you are embedding Python, you are providing your own main
program. One of the things this main program has to do is initialize program. One of the things this main program has to do is initialize
the Python interpreter. At the very least, you have to call the the Python interpreter. At the very least, you have to call the
function \code{initall()}. There are optional calls to pass command function \code{Py_Initialize()}. There are optional calls to pass command
line arguments to Python. Then later you can call the interpreter line arguments to Python. Then later you can call the interpreter
from any part of the application. from any part of the application.
There are several different ways to call the interpreter: you can pass There are several different ways to call the interpreter: you can pass
a string containing Python statements to \code{run_command()}, or you a string containing Python statements to \code{PyRun_SimpleString()},
can pass a stdio file pointer and a file name (for identification in or you can pass a stdio file pointer and a file name (for
error messages only) to \code{run_script()}. You can also call the identification in error messages only) to \code{PyRun_SimpleFile()}. You
lower-level operations described in the previous chapters to construct can also call the lower-level operations described in the previous
and use Python objects. chapters to construct and use Python objects.
A simple demo of embedding Python can be found in the directory A simple demo of embedding Python can be found in the directory
\file{Demo/embed}. \file{Demo/embed}.
...@@ -828,8 +1183,8 @@ dynamic loading of extension modules implemented in C. When shared ...@@ -828,8 +1183,8 @@ dynamic loading of extension modules implemented in C. When shared
libraries are used dynamic loading is configured automatically; libraries are used dynamic loading is configured automatically;
otherwise you have to select it as a build option (see below). Once otherwise you have to select it as a build option (see below). Once
configured, dynamic loading is trivial to use: when a Python program configured, dynamic loading is trivial to use: when a Python program
executes \code{import foo}, the search for modules tries to find a executes \code{import spam}, the search for modules tries to find a
file \file{foomodule.o} (\file{foomodule.so} when using shared file \file{spammodule.o} (\file{spammodule.so} when using shared
libraries) in the module search path, and if one is found, it is libraries) in the module search path, and if one is found, it is
loaded into the executing binary and executed. Once loaded, the loaded into the executing binary and executed. Once loaded, the
module acts just like a built-in extension module. module acts just like a built-in extension module.
...@@ -844,13 +1199,13 @@ loading a module that was compiled for a different version of Python ...@@ -844,13 +1199,13 @@ loading a module that was compiled for a different version of Python
(e.g. with a different representation of objects) may dump core. (e.g. with a different representation of objects) may dump core.
\section{Configuring and building the interpreter for dynamic loading} \section{Configuring and Building the Interpreter for Dynamic Loading}
There are three styles of dynamic loading: one using shared libraries, There are three styles of dynamic loading: one using shared libraries,
one using SGI IRIX 4 dynamic loading, and one using GNU dynamic one using SGI IRIX 4 dynamic loading, and one using GNU dynamic
loading. loading.
\subsection{Shared libraries} \subsection{Shared Libraries}
The following systems support dynamic loading using shared libraries: The following systems support dynamic loading using shared libraries:
SunOS 4; Solaris 2; SGI IRIX 5 (but not SGI IRIX 4!); and probably all SunOS 4; Solaris 2; SGI IRIX 5 (but not SGI IRIX 4!); and probably all
...@@ -862,7 +1217,7 @@ systems --- the \file{configure} detects the presence of the ...@@ -862,7 +1217,7 @@ systems --- the \file{configure} detects the presence of the
\file{<dlfcn.h>} header file and automatically configures dynamic \file{<dlfcn.h>} header file and automatically configures dynamic
loading. loading.
\subsection{SGI dynamic loading} \subsection{SGI IRIX 4 Dynamic Loading}
Only SGI IRIX 4 supports dynamic loading of modules using SGI dynamic Only SGI IRIX 4 supports dynamic loading of modules using SGI dynamic
loading. (SGI IRIX 5 might also support it but it is inferior to loading. (SGI IRIX 5 might also support it but it is inferior to
...@@ -883,7 +1238,7 @@ pathname of the \code{dl} directory. ...@@ -883,7 +1238,7 @@ pathname of the \code{dl} directory.
Now build and install Python as you normally would (see the Now build and install Python as you normally would (see the
\file{README} file in the toplevel Python directory.) \file{README} file in the toplevel Python directory.)
\subsection{GNU dynamic loading} \subsection{GNU Dynamic Loading}
GNU dynamic loading supports (according to its \file{README} file) the GNU dynamic loading supports (according to its \file{README} file) the
following hardware and software combinations: VAX (Ultrix), Sun 3 following hardware and software combinations: VAX (Ultrix), Sun 3
...@@ -909,13 +1264,13 @@ of the GNU DLD package. The Python interpreter you build hereafter ...@@ -909,13 +1264,13 @@ of the GNU DLD package. The Python interpreter you build hereafter
will support GNU dynamic loading. will support GNU dynamic loading.
\section{Building a dynamically loadable module} \section{Building a Dynamically Loadable Module}
Since there are three styles of dynamic loading, there are also three Since there are three styles of dynamic loading, there are also three
groups of instructions for building a dynamically loadable module. groups of instructions for building a dynamically loadable module.
Instructions common for all three styles are given first. Assuming Instructions common for all three styles are given first. Assuming
your module is called \code{foo}, the source filename must be your module is called \code{spam}, the source filename must be
\file{foomodule.c}, so the object name is \file{foomodule.o}. The \file{spammodule.c}, so the object name is \file{spammodule.o}. The
module must be written as a normal Python extension module (as module must be written as a normal Python extension module (as
described earlier). described earlier).
...@@ -931,7 +1286,7 @@ also add \samp{-DHAVE_CONFIG_H} to the definition of \var{CFLAGS} to ...@@ -931,7 +1286,7 @@ also add \samp{-DHAVE_CONFIG_H} to the definition of \var{CFLAGS} to
direct the Python headers to include \file{config.h}. direct the Python headers to include \file{config.h}.
\subsection{Shared libraries} \subsection{Shared Libraries}
You must link the \samp{.o} file to produce a shared library. This is You must link the \samp{.o} file to produce a shared library. This is
done using a special invocation of the \UNIX{} loader/linker, {\em done using a special invocation of the \UNIX{} loader/linker, {\em
...@@ -939,17 +1294,17 @@ ld}(1). Unfortunately the invocation differs slightly per system. ...@@ -939,17 +1294,17 @@ ld}(1). Unfortunately the invocation differs slightly per system.
On SunOS 4, use On SunOS 4, use
\begin{verbatim} \begin{verbatim}
ld foomodule.o -o foomodule.so ld spammodule.o -o spammodule.so
\end{verbatim} \end{verbatim}
On Solaris 2, use On Solaris 2, use
\begin{verbatim} \begin{verbatim}
ld -G foomodule.o -o foomodule.so ld -G spammodule.o -o spammodule.so
\end{verbatim} \end{verbatim}
On SGI IRIX 5, use On SGI IRIX 5, use
\begin{verbatim} \begin{verbatim}
ld -shared foomodule.o -o foomodule.so ld -shared spammodule.o -o spammodule.so
\end{verbatim} \end{verbatim}
On other systems, consult the manual page for {\em ld}(1) to find what On other systems, consult the manual page for {\em ld}(1) to find what
...@@ -960,46 +1315,46 @@ been linked with Python (e.g. a windowing system), these must be ...@@ -960,46 +1315,46 @@ been linked with Python (e.g. a windowing system), these must be
passed to the {\em ld} command as \samp{-l} options after the passed to the {\em ld} command as \samp{-l} options after the
\samp{.o} file. \samp{.o} file.
The resulting file \file{foomodule.so} must be copied into a directory The resulting file \file{spammodule.so} must be copied into a directory
along the Python module search path. along the Python module search path.
\subsection{SGI dynamic loading} \subsection{SGI IRIX 4 Dynamic Loading}
{bf IMPORTANT:} You must compile your extension module with the {bf IMPORTANT:} You must compile your extension module with the
additional C flag \samp{-G0} (or \samp{-G 0}). This instruct the additional C flag \samp{-G0} (or \samp{-G 0}). This instruct the
assembler to generate position-independent code. assembler to generate position-independent code.
You don't need to link the resulting \file{foomodule.o} file; just You don't need to link the resulting \file{spammodule.o} file; just
copy it into a directory along the Python module search path. copy it into a directory along the Python module search path.
The first time your extension is loaded, it takes some extra time and The first time your extension is loaded, it takes some extra time and
a few messages may be printed. This creates a file a few messages may be printed. This creates a file
\file{foomodule.ld} which is an image that can be loaded quickly into \file{spammodule.ld} which is an image that can be loaded quickly into
the Python interpreter process. When a new Python interpreter is the Python interpreter process. When a new Python interpreter is
installed, the \code{dl} package detects this and rebuilds installed, the \code{dl} package detects this and rebuilds
\file{foomodule.ld}. The file \file{foomodule.ld} is placed in the \file{spammodule.ld}. The file \file{spammodule.ld} is placed in the
directory where \file{foomodule.o} was found, unless this directory is directory where \file{spammodule.o} was found, unless this directory is
unwritable; in that case it is placed in a temporary unwritable; in that case it is placed in a temporary
directory.\footnote{Check the manual page of the \code{dl} package for directory.\footnote{Check the manual page of the \code{dl} package for
details.} details.}
If your extension modules uses additional system libraries, you must If your extension modules uses additional system libraries, you must
create a file \file{foomodule.libs} in the same directory as the create a file \file{spammodule.libs} in the same directory as the
\file{foomodule.o}. This file should contain one or more lines with \file{spammodule.o}. This file should contain one or more lines with
whitespace-separated options that will be passed to the linker --- whitespace-separated options that will be passed to the linker ---
normally only \samp{-l} options or absolute pathnames of libraries normally only \samp{-l} options or absolute pathnames of libraries
(\samp{.a} files) should be used. (\samp{.a} files) should be used.
\subsection{GNU dynamic loading} \subsection{GNU Dynamic Loading}
Just copy \file{foomodule.o} into a directory along the Python module Just copy \file{spammodule.o} into a directory along the Python module
search path. search path.
If your extension modules uses additional system libraries, you must If your extension modules uses additional system libraries, you must
create a file \file{foomodule.libs} in the same directory as the create a file \file{spammodule.libs} in the same directory as the
\file{foomodule.o}. This file should contain one or more lines with \file{spammodule.o}. This file should contain one or more lines with
whitespace-separated absolute pathnames of libraries (\samp{.a} whitespace-separated absolute pathnames of libraries (\samp{.a}
files). No \samp{-l} options can be used. files). No \samp{-l} options can be used.
......
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