blob: ff30704d657fdf78cbd3bf9a2ebd5bee1fc75572 [file] [log] [blame]
Guido van Rossumf70e43a1991-02-19 12:39:46 +00001
Guido van Rossum85a5fbb1990-10-14 12:07:46 +00002/* String object interface */
3
Fred Drake3cf4d2b2000-07-09 00:55:06 +00004#ifndef Py_STRINGOBJECT_H
5#define Py_STRINGOBJECT_H
6#ifdef __cplusplus
7extern "C" {
8#endif
9
Barry Warsawdadace02001-08-24 18:32:06 +000010#include <stdarg.h>
11
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000012/*
Guido van Rossumcaa63801995-01-12 11:45:45 +000013Type PyStringObject represents a character string. An extra zero byte is
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000014reserved at the end to ensure it is zero-terminated, but a size is
15present so strings with null bytes in them can be represented. This
16is an immutable object type.
17
18There are functions to create new string objects, to test
19an object for string-ness, and to get the
20string value. The latter function returns a null pointer
21if the object is not of the proper type.
22There is a variant that takes an explicit size as well as a
23variant that assumes a zero-terminated string. Note that none of the
24functions should be applied to nil objects.
25*/
26
Tim Peters1f7df352002-03-29 03:29:08 +000027/* Caching the hash (ob_shash) saves recalculation of a string's hash value.
Guido van Rossum45ec02a2002-08-19 21:43:18 +000028 Interning strings (ob_sstate) tries to ensure that only one string
Tim Peters1f7df352002-03-29 03:29:08 +000029 object with a given value exists, so equality tests can be one pointer
30 comparison. This is generally restricted to strings that "look like"
31 Python identifiers, although the intern() builtin can be used to force
32 interning of any string.
33 Together, these sped the interpreter by up to 20%. */
Guido van Rossumfdebf251996-07-30 16:42:03 +000034
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000035typedef struct {
Fred Drake3cf4d2b2000-07-09 00:55:06 +000036 PyObject_VAR_HEAD
Fred Drake3cf4d2b2000-07-09 00:55:06 +000037 long ob_shash;
Guido van Rossum45ec02a2002-08-19 21:43:18 +000038 int ob_sstate;
Fred Drake3cf4d2b2000-07-09 00:55:06 +000039 char ob_sval[1];
Guido van Rossumcaa63801995-01-12 11:45:45 +000040} PyStringObject;
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000041
Guido van Rossum45ec02a2002-08-19 21:43:18 +000042#define SSTATE_NOT_INTERNED 0
43#define SSTATE_INTERNED_MORTAL 1
44#define SSTATE_INTERNED_IMMORTAL 2
45
Mark Hammond91a681d2002-08-12 07:21:58 +000046PyAPI_DATA(PyTypeObject) PyBaseString_Type;
47PyAPI_DATA(PyTypeObject) PyString_Type;
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000048
Guido van Rossum5eef77a2001-08-30 03:08:07 +000049#define PyString_Check(op) PyObject_TypeCheck(op, &PyString_Type)
Tim Peters5a49ade2001-09-11 01:41:59 +000050#define PyString_CheckExact(op) ((op)->ob_type == &PyString_Type)
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000051
Mark Hammond91a681d2002-08-12 07:21:58 +000052PyAPI_FUNC(PyObject *) PyString_FromStringAndSize(const char *, int);
53PyAPI_FUNC(PyObject *) PyString_FromString(const char *);
54PyAPI_FUNC(PyObject *) PyString_FromFormatV(const char*, va_list)
Neil Schemenauer90b68902001-10-23 02:21:22 +000055 __attribute__((format(printf, 1, 0)));
Mark Hammond91a681d2002-08-12 07:21:58 +000056PyAPI_FUNC(PyObject *) PyString_FromFormat(const char*, ...)
Neil Schemenauer90b68902001-10-23 02:21:22 +000057 __attribute__((format(printf, 1, 2)));
Mark Hammond91a681d2002-08-12 07:21:58 +000058PyAPI_FUNC(int) PyString_Size(PyObject *);
59PyAPI_FUNC(char *) PyString_AsString(PyObject *);
Martin v. Löwis8a8da792002-08-14 07:46:28 +000060PyAPI_FUNC(PyObject *) PyString_Repr(PyObject *, int);
Mark Hammond91a681d2002-08-12 07:21:58 +000061PyAPI_FUNC(void) PyString_Concat(PyObject **, PyObject *);
62PyAPI_FUNC(void) PyString_ConcatAndDel(PyObject **, PyObject *);
63PyAPI_FUNC(int) _PyString_Resize(PyObject **, int);
64PyAPI_FUNC(int) _PyString_Eq(PyObject *, PyObject*);
65PyAPI_FUNC(PyObject *) PyString_Format(PyObject *, PyObject *);
66PyAPI_FUNC(PyObject *) _PyString_FormatLong(PyObject*, int, int,
Tim Peters38fd5b62000-09-21 05:43:11 +000067 int, char**, int*);
Martin v. Löwis8a8da792002-08-14 07:46:28 +000068extern DL_IMPORT(PyObject *) PyString_DecodeEscape(const char *, int,
69 const char *, int,
70 const char *);
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000071
Mark Hammond91a681d2002-08-12 07:21:58 +000072PyAPI_FUNC(void) PyString_InternInPlace(PyObject **);
Guido van Rossum45ec02a2002-08-19 21:43:18 +000073PyAPI_FUNC(void) PyString_InternImmortal(PyObject **);
Mark Hammond91a681d2002-08-12 07:21:58 +000074PyAPI_FUNC(PyObject *) PyString_InternFromString(const char *);
75PyAPI_FUNC(void) _Py_ReleaseInternedStrings(void);
Guido van Rossum1e6e9a21997-01-18 07:53:23 +000076
Guido van Rossum45ec02a2002-08-19 21:43:18 +000077/* Use only if you know it's a string */
78#define PyString_CHECK_INTERNED(op) (((PyStringObject *)(op))->ob_sstate)
79
Guido van Rossum85a5fbb1990-10-14 12:07:46 +000080/* Macro, trading safety for speed */
Barry Warsawaccfb841997-01-06 22:42:50 +000081#define PyString_AS_STRING(op) (((PyStringObject *)(op))->ob_sval)
82#define PyString_GET_SIZE(op) (((PyStringObject *)(op))->ob_size)
Guido van Rossuma3309961993-07-28 09:05:47 +000083
Tim Petersa7259592001-06-16 05:11:17 +000084/* _PyString_Join(sep, x) is like sep.join(x). sep must be PyStringObject*,
85 x must be an iterable object. */
Mark Hammond91a681d2002-08-12 07:21:58 +000086PyAPI_FUNC(PyObject *) _PyString_Join(PyObject *sep, PyObject *x);
Tim Petersa7259592001-06-16 05:11:17 +000087
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +000088/* --- Generic Codecs ----------------------------------------------------- */
89
Marc-André Lemburg2d920412001-05-15 12:00:02 +000090/* Create an object by decoding the encoded string s of the
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +000091 given size. */
92
Mark Hammond91a681d2002-08-12 07:21:58 +000093PyAPI_FUNC(PyObject*) PyString_Decode(
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +000094 const char *s, /* encoded string */
95 int size, /* size of buffer */
96 const char *encoding, /* encoding */
97 const char *errors /* error handling */
98 );
99
100/* Encodes a char buffer of the given size and returns a
Marc-André Lemburg2d920412001-05-15 12:00:02 +0000101 Python object. */
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +0000102
Mark Hammond91a681d2002-08-12 07:21:58 +0000103PyAPI_FUNC(PyObject*) PyString_Encode(
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +0000104 const char *s, /* string char buffer */
105 int size, /* number of chars to encode */
106 const char *encoding, /* encoding */
107 const char *errors /* error handling */
108 );
109
Marc-André Lemburg2d920412001-05-15 12:00:02 +0000110/* Encodes a string object and returns the result as Python
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +0000111 object. */
112
Mark Hammond91a681d2002-08-12 07:21:58 +0000113PyAPI_FUNC(PyObject*) PyString_AsEncodedObject(
Marc-André Lemburg2d920412001-05-15 12:00:02 +0000114 PyObject *str, /* string object */
115 const char *encoding, /* encoding */
116 const char *errors /* error handling */
117 );
118
119/* Encodes a string object and returns the result as Python string
120 object.
121
122 If the codec returns an Unicode object, the object is converted
123 back to a string using the default encoding.
124
125 DEPRECATED - use PyString_AsEncodedObject() instead. */
126
Mark Hammond91a681d2002-08-12 07:21:58 +0000127PyAPI_FUNC(PyObject*) PyString_AsEncodedString(
Marc-André Lemburg3d1a1d72000-07-06 11:25:40 +0000128 PyObject *str, /* string object */
129 const char *encoding, /* encoding */
130 const char *errors /* error handling */
131 );
132
Marc-André Lemburg2d920412001-05-15 12:00:02 +0000133/* Decodes a string object and returns the result as Python
134 object. */
135
Mark Hammond91a681d2002-08-12 07:21:58 +0000136PyAPI_FUNC(PyObject*) PyString_AsDecodedObject(
Marc-André Lemburg2d920412001-05-15 12:00:02 +0000137 PyObject *str, /* string object */
138 const char *encoding, /* encoding */
139 const char *errors /* error handling */
140 );
141
142/* Decodes a string object and returns the result as Python string
143 object.
144
145 If the codec returns an Unicode object, the object is converted
146 back to a string using the default encoding.
147
148 DEPRECATED - use PyString_AsDecodedObject() instead. */
149
Mark Hammond91a681d2002-08-12 07:21:58 +0000150PyAPI_FUNC(PyObject*) PyString_AsDecodedString(
Marc-André Lemburg2d920412001-05-15 12:00:02 +0000151 PyObject *str, /* string object */
152 const char *encoding, /* encoding */
153 const char *errors /* error handling */
154 );
155
Marc-André Lemburgd1ba4432000-09-19 21:04:18 +0000156/* Provides access to the internal data buffer and size of a string
157 object or the default encoded version of an Unicode object. Passing
158 NULL as *len parameter will force the string buffer to be
159 0-terminated (passing a string with embedded NULL characters will
160 cause an exception). */
161
Mark Hammond91a681d2002-08-12 07:21:58 +0000162PyAPI_FUNC(int) PyString_AsStringAndSize(
Marc-André Lemburgd1ba4432000-09-19 21:04:18 +0000163 register PyObject *obj, /* string or Unicode object */
164 register char **s, /* pointer to buffer variable */
165 register int *len /* pointer to length variable or NULL
166 (only possible for 0-terminated
167 strings) */
168 );
169
170
Guido van Rossuma3309961993-07-28 09:05:47 +0000171#ifdef __cplusplus
172}
173#endif
174#endif /* !Py_STRINGOBJECT_H */