| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 1 | #ifndef Py_CODECREGISTRY_H | 
|  | 2 | #define Py_CODECREGISTRY_H | 
|  | 3 | #ifdef __cplusplus | 
|  | 4 | extern "C" { | 
|  | 5 | #endif | 
|  | 6 |  | 
|  | 7 | /* ------------------------------------------------------------------------ | 
|  | 8 |  | 
|  | 9 | Python Codec Registry and support functions | 
|  | 10 |  | 
|  | 11 |  | 
|  | 12 | Written by Marc-Andre Lemburg (mal@lemburg.com). | 
|  | 13 |  | 
| Guido van Rossum | 16b1ad9 | 2000-08-03 16:24:25 +0000 | [diff] [blame] | 14 | Copyright (c) Corporation for National Research Initiatives. | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 15 |  | 
|  | 16 | ------------------------------------------------------------------------ */ | 
|  | 17 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 18 | /* Register a new codec search function. | 
|  | 19 |  | 
|  | 20 | As side effect, this tries to load the encodings package, if not | 
|  | 21 | yet done, to make sure that it is always first in the list of | 
|  | 22 | search functions. | 
|  | 23 |  | 
|  | 24 | The search_function's refcount is incremented by this function. */ | 
|  | 25 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 26 | PyAPI_FUNC(int) PyCodec_Register( | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 27 | PyObject *search_function | 
|  | 28 | ); | 
|  | 29 |  | 
| Marc-André Lemburg | b2750b5 | 2008-06-06 12:18:17 +0000 | [diff] [blame] | 30 | /* Codec registry lookup API. | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 31 |  | 
| Thomas Wouters | a977329 | 2006-04-21 09:43:23 +0000 | [diff] [blame] | 32 | Looks up the given encoding and returns a CodecInfo object with | 
|  | 33 | function attributes which implement the different aspects of | 
|  | 34 | processing the encoding. | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 35 |  | 
|  | 36 | The encoding string is looked up converted to all lower-case | 
|  | 37 | characters. This makes encodings looked up through this mechanism | 
|  | 38 | effectively case-insensitive. | 
|  | 39 |  | 
| Thomas Wouters | a977329 | 2006-04-21 09:43:23 +0000 | [diff] [blame] | 40 | If no codec is found, a KeyError is set and NULL returned. | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 41 |  | 
|  | 42 | As side effect, this tries to load the encodings package, if not | 
|  | 43 | yet done. This is part of the lazy load strategy for the encodings | 
|  | 44 | package. | 
|  | 45 |  | 
|  | 46 | */ | 
|  | 47 |  | 
| Martin v. Löwis | 4d0d471 | 2010-12-03 20:14:31 +0000 | [diff] [blame] | 48 | #ifndef Py_LIMITED_API | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 49 | PyAPI_FUNC(PyObject *) _PyCodec_Lookup( | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 50 | const char *encoding | 
|  | 51 | ); | 
| Nick Coghlan | 8fad167 | 2014-09-15 23:50:44 +1200 | [diff] [blame] | 52 |  | 
|  | 53 | PyAPI_FUNC(int) _PyCodec_Forget( | 
|  | 54 | const char *encoding | 
|  | 55 | ); | 
| Martin v. Löwis | 4d0d471 | 2010-12-03 20:14:31 +0000 | [diff] [blame] | 56 | #endif | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 57 |  | 
| Marc-André Lemburg | b2750b5 | 2008-06-06 12:18:17 +0000 | [diff] [blame] | 58 | /* Codec registry encoding check API. | 
|  | 59 |  | 
|  | 60 | Returns 1/0 depending on whether there is a registered codec for | 
|  | 61 | the given encoding. | 
|  | 62 |  | 
|  | 63 | */ | 
|  | 64 |  | 
|  | 65 | PyAPI_FUNC(int) PyCodec_KnownEncoding( | 
|  | 66 | const char *encoding | 
|  | 67 | ); | 
|  | 68 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 69 | /* Generic codec based encoding API. | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 70 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 71 | object is passed through the encoder function found for the given | 
|  | 72 | encoding using the error handling method defined by errors. errors | 
|  | 73 | may be NULL to use the default method defined for the codec. | 
| Serhiy Storchaka | 009b811 | 2015-03-18 21:53:15 +0200 | [diff] [blame] | 74 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 75 | Raises a LookupError in case no encoder can be found. | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 76 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 77 | */ | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 78 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 79 | PyAPI_FUNC(PyObject *) PyCodec_Encode( | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 80 | PyObject *object, | 
|  | 81 | const char *encoding, | 
|  | 82 | const char *errors | 
|  | 83 | ); | 
|  | 84 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 85 | /* Generic codec based decoding API. | 
|  | 86 |  | 
|  | 87 | object is passed through the decoder function found for the given | 
|  | 88 | encoding using the error handling method defined by errors. errors | 
|  | 89 | may be NULL to use the default method defined for the codec. | 
| Serhiy Storchaka | 009b811 | 2015-03-18 21:53:15 +0200 | [diff] [blame] | 90 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 91 | Raises a LookupError in case no encoder can be found. | 
|  | 92 |  | 
|  | 93 | */ | 
|  | 94 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 95 | PyAPI_FUNC(PyObject *) PyCodec_Decode( | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 96 | PyObject *object, | 
|  | 97 | const char *encoding, | 
|  | 98 | const char *errors | 
|  | 99 | ); | 
|  | 100 |  | 
| Martin v. Löwis | 1c0689c | 2014-01-03 21:36:49 +0100 | [diff] [blame] | 101 | #ifndef Py_LIMITED_API | 
| Nick Coghlan | c72e4e6 | 2013-11-22 22:39:36 +1000 | [diff] [blame] | 102 | /* Text codec specific encoding and decoding API. | 
|  | 103 |  | 
|  | 104 | Checks the encoding against a list of codecs which do not | 
|  | 105 | implement a str<->bytes encoding before attempting the | 
|  | 106 | operation. | 
|  | 107 |  | 
|  | 108 | Please note that these APIs are internal and should not | 
|  | 109 | be used in Python C extensions. | 
|  | 110 |  | 
| Nick Coghlan | a9b1524 | 2014-02-04 22:11:18 +1000 | [diff] [blame] | 111 | XXX (ncoghlan): should we make these, or something like them, public | 
|  | 112 | in Python 3.5+? | 
|  | 113 |  | 
| Nick Coghlan | c72e4e6 | 2013-11-22 22:39:36 +1000 | [diff] [blame] | 114 | */ | 
| Nick Coghlan | a9b1524 | 2014-02-04 22:11:18 +1000 | [diff] [blame] | 115 | PyAPI_FUNC(PyObject *) _PyCodec_LookupTextEncoding( | 
|  | 116 | const char *encoding, | 
|  | 117 | const char *alternate_command | 
|  | 118 | ); | 
| Nick Coghlan | c72e4e6 | 2013-11-22 22:39:36 +1000 | [diff] [blame] | 119 |  | 
|  | 120 | PyAPI_FUNC(PyObject *) _PyCodec_EncodeText( | 
|  | 121 | PyObject *object, | 
|  | 122 | const char *encoding, | 
|  | 123 | const char *errors | 
|  | 124 | ); | 
|  | 125 |  | 
|  | 126 | PyAPI_FUNC(PyObject *) _PyCodec_DecodeText( | 
|  | 127 | PyObject *object, | 
|  | 128 | const char *encoding, | 
|  | 129 | const char *errors | 
|  | 130 | ); | 
| Nick Coghlan | a9b1524 | 2014-02-04 22:11:18 +1000 | [diff] [blame] | 131 |  | 
|  | 132 | /* These two aren't actually text encoding specific, but _io.TextIOWrapper | 
|  | 133 | * is the only current API consumer. | 
|  | 134 | */ | 
|  | 135 | PyAPI_FUNC(PyObject *) _PyCodecInfo_GetIncrementalDecoder( | 
|  | 136 | PyObject *codec_info, | 
|  | 137 | const char *errors | 
|  | 138 | ); | 
|  | 139 |  | 
|  | 140 | PyAPI_FUNC(PyObject *) _PyCodecInfo_GetIncrementalEncoder( | 
|  | 141 | PyObject *codec_info, | 
|  | 142 | const char *errors | 
|  | 143 | ); | 
| Nick Coghlan | c72e4e6 | 2013-11-22 22:39:36 +1000 | [diff] [blame] | 144 | #endif | 
|  | 145 |  | 
|  | 146 |  | 
|  | 147 |  | 
| Serhiy Storchaka | 009b811 | 2015-03-18 21:53:15 +0200 | [diff] [blame] | 148 | /* --- Codec Lookup APIs -------------------------------------------------- | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 149 |  | 
|  | 150 | All APIs return a codec object with incremented refcount and are | 
|  | 151 | based on _PyCodec_Lookup().  The same comments w/r to the encoding | 
|  | 152 | name also apply to these APIs. | 
|  | 153 |  | 
|  | 154 | */ | 
|  | 155 |  | 
|  | 156 | /* Get an encoder function for the given encoding. */ | 
|  | 157 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 158 | PyAPI_FUNC(PyObject *) PyCodec_Encoder( | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 159 | const char *encoding | 
|  | 160 | ); | 
|  | 161 |  | 
|  | 162 | /* Get a decoder function for the given encoding. */ | 
|  | 163 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 164 | PyAPI_FUNC(PyObject *) PyCodec_Decoder( | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 165 | const char *encoding | 
|  | 166 | ); | 
|  | 167 |  | 
| Martin Panter | 7462b649 | 2015-11-02 03:37:02 +0000 | [diff] [blame] | 168 | /* Get an IncrementalEncoder object for the given encoding. */ | 
| Thomas Wouters | a977329 | 2006-04-21 09:43:23 +0000 | [diff] [blame] | 169 |  | 
|  | 170 | PyAPI_FUNC(PyObject *) PyCodec_IncrementalEncoder( | 
|  | 171 | const char *encoding, | 
|  | 172 | const char *errors | 
|  | 173 | ); | 
|  | 174 |  | 
| Martin Panter | 7462b649 | 2015-11-02 03:37:02 +0000 | [diff] [blame] | 175 | /* Get an IncrementalDecoder object function for the given encoding. */ | 
| Thomas Wouters | a977329 | 2006-04-21 09:43:23 +0000 | [diff] [blame] | 176 |  | 
|  | 177 | PyAPI_FUNC(PyObject *) PyCodec_IncrementalDecoder( | 
|  | 178 | const char *encoding, | 
|  | 179 | const char *errors | 
|  | 180 | ); | 
|  | 181 |  | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 182 | /* Get a StreamReader factory function for the given encoding. */ | 
|  | 183 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 184 | PyAPI_FUNC(PyObject *) PyCodec_StreamReader( | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 185 | const char *encoding, | 
|  | 186 | PyObject *stream, | 
|  | 187 | const char *errors | 
|  | 188 | ); | 
|  | 189 |  | 
|  | 190 | /* Get a StreamWriter factory function for the given encoding. */ | 
|  | 191 |  | 
| Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 192 | PyAPI_FUNC(PyObject *) PyCodec_StreamWriter( | 
| Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 193 | const char *encoding, | 
|  | 194 | PyObject *stream, | 
|  | 195 | const char *errors | 
|  | 196 | ); | 
|  | 197 |  | 
| Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 198 | /* Unicode encoding error handling callback registry API */ | 
|  | 199 |  | 
| Georg Brandl | bab3378 | 2010-11-20 13:44:41 +0000 | [diff] [blame] | 200 | /* Register the error handling callback function error under the given | 
| Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 201 | name. This function will be called by the codec when it encounters | 
|  | 202 | unencodable characters/undecodable bytes and doesn't know the | 
|  | 203 | callback name, when name is specified as the error parameter | 
|  | 204 | in the call to the encode/decode function. | 
|  | 205 | Return 0 on success, -1 on error */ | 
|  | 206 | PyAPI_FUNC(int) PyCodec_RegisterError(const char *name, PyObject *error); | 
|  | 207 |  | 
| Georg Brandl | bab3378 | 2010-11-20 13:44:41 +0000 | [diff] [blame] | 208 | /* Lookup the error handling callback function registered under the given | 
|  | 209 | name. As a special case NULL can be passed, in which case | 
| Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 210 | the error handling callback for "strict" will be returned. */ | 
|  | 211 | PyAPI_FUNC(PyObject *) PyCodec_LookupError(const char *name); | 
|  | 212 |  | 
|  | 213 | /* raise exc as an exception */ | 
|  | 214 | PyAPI_FUNC(PyObject *) PyCodec_StrictErrors(PyObject *exc); | 
|  | 215 |  | 
|  | 216 | /* ignore the unicode error, skipping the faulty input */ | 
|  | 217 | PyAPI_FUNC(PyObject *) PyCodec_IgnoreErrors(PyObject *exc); | 
|  | 218 |  | 
| Georg Brandl | bab3378 | 2010-11-20 13:44:41 +0000 | [diff] [blame] | 219 | /* replace the unicode encode error with ? or U+FFFD */ | 
| Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 220 | PyAPI_FUNC(PyObject *) PyCodec_ReplaceErrors(PyObject *exc); | 
|  | 221 |  | 
|  | 222 | /* replace the unicode encode error with XML character references */ | 
|  | 223 | PyAPI_FUNC(PyObject *) PyCodec_XMLCharRefReplaceErrors(PyObject *exc); | 
|  | 224 |  | 
|  | 225 | /* replace the unicode encode error with backslash escapes (\x, \u and \U) */ | 
|  | 226 | PyAPI_FUNC(PyObject *) PyCodec_BackslashReplaceErrors(PyObject *exc); | 
|  | 227 |  | 
| Serhiy Storchaka | 34d0ac8 | 2016-12-27 14:57:39 +0200 | [diff] [blame] | 228 | #if !defined(Py_LIMITED_API) || Py_LIMITED_API+0 >= 0x03050000 | 
| Serhiy Storchaka | 166ebc4 | 2014-11-25 13:57:17 +0200 | [diff] [blame] | 229 | /* replace the unicode encode error with backslash escapes (\N, \x, \u and \U) */ | 
|  | 230 | PyAPI_FUNC(PyObject *) PyCodec_NameReplaceErrors(PyObject *exc); | 
| Serhiy Storchaka | 34d0ac8 | 2016-12-27 14:57:39 +0200 | [diff] [blame] | 231 | #endif | 
| Serhiy Storchaka | 166ebc4 | 2014-11-25 13:57:17 +0200 | [diff] [blame] | 232 |  | 
| Serhiy Storchaka | 34d0ac8 | 2016-12-27 14:57:39 +0200 | [diff] [blame] | 233 | #ifndef Py_LIMITED_API | 
| Antoine Pitrou | e606983 | 2011-10-15 16:38:20 +0200 | [diff] [blame] | 234 | PyAPI_DATA(const char *) Py_hexdigits; | 
| Serhiy Storchaka | 34d0ac8 | 2016-12-27 14:57:39 +0200 | [diff] [blame] | 235 | #endif | 
| Victor Stinner | f5cff56 | 2011-10-14 02:13:11 +0200 | [diff] [blame] | 236 |  | 
| Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 237 | #ifdef __cplusplus | 
|  | 238 | } | 
|  | 239 | #endif | 
|  | 240 | #endif /* !Py_CODECREGISTRY_H */ |