1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
|
.. highlightlang:: c
.. _cporting-howto:
********************************
Porting Extension Modules to 3.0
********************************
:author: Benjamin Peterson
.. topic:: Abstract
Although changing the C-API was not one of Python 3.0's objectives, the many
Python level changes made leaving 2.x's API intact impossible. In fact, some
changes such as :func:`int` and :func:`long` unification are more obvious on
the C level. This document endeavors to document incompatibilities and how
they can be worked around.
Conditional compilation
=======================
The easiest way to compile only some code for 3.0 is to check if
:c:macro:`PY_MAJOR_VERSION` is greater than or equal to 3. ::
#if PY_MAJOR_VERSION >= 3
#define IS_PY3K
#endif
API functions that are not present can be aliased to their equivalents within
conditional blocks.
Changes to Object APIs
======================
Python 3.0 merged together some types with similar functions while cleanly
separating others.
str/unicode Unification
-----------------------
Python 3.0's :func:`str` (``PyString_*`` functions in C) type is equivalent to
2.x's :func:`unicode` (``PyUnicode_*``). The old 8-bit string type has become
:func:`bytes`. Python 2.6 and later provide a compatibility header,
:file:`bytesobject.h`, mapping ``PyBytes`` names to ``PyString`` ones. For best
compatibility with 3.0, :c:type:`PyUnicode` should be used for textual data and
:c:type:`PyBytes` for binary data. It's also important to remember that
:c:type:`PyBytes` and :c:type:`PyUnicode` in 3.0 are not interchangeable like
:c:type:`PyString` and :c:type:`PyUnicode` are in 2.x. The following example
shows best practices with regards to :c:type:`PyUnicode`, :c:type:`PyString`,
and :c:type:`PyBytes`. ::
#include "stdlib.h"
#include "Python.h"
#include "bytesobject.h"
/* text example */
static PyObject *
say_hello(PyObject *self, PyObject *args) {
PyObject *name, *result;
if (!PyArg_ParseTuple(args, "U:say_hello", &name))
return NULL;
result = PyUnicode_FromFormat("Hello, %S!", name);
return result;
}
/* just a forward */
static char * do_encode(PyObject *);
/* bytes example */
static PyObject *
encode_object(PyObject *self, PyObject *args) {
char *encoded;
PyObject *result, *myobj;
if (!PyArg_ParseTuple(args, "O:encode_object", &myobj))
return NULL;
encoded = do_encode(myobj);
if (encoded == NULL)
return NULL;
result = PyBytes_FromString(encoded);
free(encoded);
return result;
}
long/int Unification
--------------------
In Python 3.0, there is only one integer type. It is called :func:`int` on the
Python level, but actually corresponds to 2.x's :func:`long` type. In the
C-API, ``PyInt_*`` functions are replaced by their ``PyLong_*`` neighbors. The
best course of action here is using the ``PyInt_*`` functions aliased to
``PyLong_*`` found in :file:`intobject.h`. The abstract ``PyNumber_*`` APIs
can also be used in some cases. ::
#include "Python.h"
#include "intobject.h"
static PyObject *
add_ints(PyObject *self, PyObject *args) {
int one, two;
PyObject *result;
if (!PyArg_ParseTuple(args, "ii:add_ints", &one, &two))
return NULL;
return PyInt_FromLong(one + two);
}
Module initialization and state
===============================
Python 3.0 has a revamped extension module initialization system. (See
:pep:`3121`.) Instead of storing module state in globals, they should be stored
in an interpreter specific structure. Creating modules that act correctly in
both 2.x and 3.0 is tricky. The following simple example demonstrates how. ::
#include "Python.h"
struct module_state {
PyObject *error;
};
#if PY_MAJOR_VERSION >= 3
#define GETSTATE(m) ((struct module_state*)PyModule_GetState(m))
#else
#define GETSTATE(m) (&_state)
static struct module_state _state;
#endif
static PyObject *
error_out(PyObject *m) {
struct module_state *st = GETSTATE(m);
PyErr_SetString(st->error, "something bad happened");
return NULL;
}
static PyMethodDef myextension_methods[] = {
{"error_out", (PyCFunction)error_out, METH_NOARGS, NULL},
{NULL, NULL}
};
#if PY_MAJOR_VERSION >= 3
static int myextension_traverse(PyObject *m, visitproc visit, void *arg) {
Py_VISIT(GETSTATE(m)->error);
return 0;
}
static int myextension_clear(PyObject *m) {
Py_CLEAR(GETSTATE(m)->error);
return 0;
}
static struct PyModuleDef moduledef = {
PyModuleDef_HEAD_INIT,
"myextension",
NULL,
sizeof(struct module_state),
myextension_methods,
NULL,
myextension_traverse,
myextension_clear,
NULL
};
#define INITERROR return NULL
PyObject *
PyInit_myextension(void)
#else
#define INITERROR return
void
initmyextension(void)
#endif
{
#if PY_MAJOR_VERSION >= 3
PyObject *module = PyModule_Create(&moduledef);
#else
PyObject *module = Py_InitModule("myextension", myextension_methods);
#endif
if (module == NULL)
INITERROR;
struct module_state *st = GETSTATE(module);
st->error = PyErr_NewException("myextension.Error", NULL, NULL);
if (st->error == NULL) {
Py_DECREF(module);
INITERROR;
}
#if PY_MAJOR_VERSION >= 3
return module;
#endif
}
Other options
=============
If you are writing a new extension module, you might consider `Cython
<http://www.cython.org>`_. It translates a Python-like language to C. The
extension modules it creates are compatible with Python 3.x and 2.x.
|
