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 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 48 | PyAPI_FUNC(PyObject *) _PyCodec_Lookup( |
Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 49 | const char *encoding |
| 50 | ); |
| 51 | |
Marc-André Lemburg | b2750b5 | 2008-06-06 12:18:17 +0000 | [diff] [blame] | 52 | /* Codec registry encoding check API. |
| 53 | |
| 54 | Returns 1/0 depending on whether there is a registered codec for |
| 55 | the given encoding. |
| 56 | |
| 57 | */ |
| 58 | |
| 59 | PyAPI_FUNC(int) PyCodec_KnownEncoding( |
| 60 | const char *encoding |
| 61 | ); |
| 62 | |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 63 | /* Generic codec based encoding API. |
Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 64 | |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 65 | object is passed through the encoder function found for the given |
| 66 | encoding using the error handling method defined by errors. errors |
| 67 | may be NULL to use the default method defined for the codec. |
| 68 | |
| 69 | Raises a LookupError in case no encoder can be found. |
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 | */ |
Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 72 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 73 | PyAPI_FUNC(PyObject *) PyCodec_Encode( |
Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 74 | PyObject *object, |
| 75 | const char *encoding, |
| 76 | const char *errors |
| 77 | ); |
| 78 | |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 79 | /* Generic codec based decoding API. |
| 80 | |
| 81 | object is passed through the decoder function found for the given |
| 82 | encoding using the error handling method defined by errors. errors |
| 83 | may be NULL to use the default method defined for the codec. |
| 84 | |
| 85 | Raises a LookupError in case no encoder can be found. |
| 86 | |
| 87 | */ |
| 88 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 89 | PyAPI_FUNC(PyObject *) PyCodec_Decode( |
Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 90 | PyObject *object, |
| 91 | const char *encoding, |
| 92 | const char *errors |
| 93 | ); |
| 94 | |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 95 | /* --- Codec Lookup APIs -------------------------------------------------- |
| 96 | |
| 97 | All APIs return a codec object with incremented refcount and are |
| 98 | based on _PyCodec_Lookup(). The same comments w/r to the encoding |
| 99 | name also apply to these APIs. |
| 100 | |
| 101 | */ |
| 102 | |
| 103 | /* Get an encoder function for the given encoding. */ |
| 104 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 105 | PyAPI_FUNC(PyObject *) PyCodec_Encoder( |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 106 | const char *encoding |
| 107 | ); |
| 108 | |
| 109 | /* Get a decoder function for the given encoding. */ |
| 110 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 111 | PyAPI_FUNC(PyObject *) PyCodec_Decoder( |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 112 | const char *encoding |
| 113 | ); |
| 114 | |
Thomas Wouters | a977329 | 2006-04-21 09:43:23 +0000 | [diff] [blame] | 115 | /* Get a IncrementalEncoder object for the given encoding. */ |
| 116 | |
| 117 | PyAPI_FUNC(PyObject *) PyCodec_IncrementalEncoder( |
| 118 | const char *encoding, |
| 119 | const char *errors |
| 120 | ); |
| 121 | |
| 122 | /* Get a IncrementalDecoder object function for the given encoding. */ |
| 123 | |
| 124 | PyAPI_FUNC(PyObject *) PyCodec_IncrementalDecoder( |
| 125 | const char *encoding, |
| 126 | const char *errors |
| 127 | ); |
| 128 | |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 129 | /* Get a StreamReader factory function for the given encoding. */ |
| 130 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 131 | PyAPI_FUNC(PyObject *) PyCodec_StreamReader( |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 132 | const char *encoding, |
| 133 | PyObject *stream, |
| 134 | const char *errors |
| 135 | ); |
| 136 | |
| 137 | /* Get a StreamWriter factory function for the given encoding. */ |
| 138 | |
Mark Hammond | 91a681d | 2002-08-12 07:21:58 +0000 | [diff] [blame] | 139 | PyAPI_FUNC(PyObject *) PyCodec_StreamWriter( |
Fred Drake | 3ac3edc | 2000-05-09 19:51:10 +0000 | [diff] [blame] | 140 | const char *encoding, |
| 141 | PyObject *stream, |
| 142 | const char *errors |
| 143 | ); |
| 144 | |
Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 145 | /* Unicode encoding error handling callback registry API */ |
| 146 | |
Georg Brandl | f65e25b | 2010-11-26 09:05:43 +0000 | [diff] [blame] | 147 | /* Register the error handling callback function error under the given |
Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 148 | name. This function will be called by the codec when it encounters |
| 149 | unencodable characters/undecodable bytes and doesn't know the |
| 150 | callback name, when name is specified as the error parameter |
| 151 | in the call to the encode/decode function. |
| 152 | Return 0 on success, -1 on error */ |
| 153 | PyAPI_FUNC(int) PyCodec_RegisterError(const char *name, PyObject *error); |
| 154 | |
Georg Brandl | f65e25b | 2010-11-26 09:05:43 +0000 | [diff] [blame] | 155 | /* Lookup the error handling callback function registered under the given |
| 156 | name. As a special case NULL can be passed, in which case |
Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 157 | the error handling callback for "strict" will be returned. */ |
| 158 | PyAPI_FUNC(PyObject *) PyCodec_LookupError(const char *name); |
| 159 | |
| 160 | /* raise exc as an exception */ |
| 161 | PyAPI_FUNC(PyObject *) PyCodec_StrictErrors(PyObject *exc); |
| 162 | |
| 163 | /* ignore the unicode error, skipping the faulty input */ |
| 164 | PyAPI_FUNC(PyObject *) PyCodec_IgnoreErrors(PyObject *exc); |
| 165 | |
Georg Brandl | f65e25b | 2010-11-26 09:05:43 +0000 | [diff] [blame] | 166 | /* replace the unicode encode error with ? or U+FFFD */ |
Walter Dörwald | 3aeb632 | 2002-09-02 13:14:32 +0000 | [diff] [blame] | 167 | PyAPI_FUNC(PyObject *) PyCodec_ReplaceErrors(PyObject *exc); |
| 168 | |
| 169 | /* replace the unicode encode error with XML character references */ |
| 170 | PyAPI_FUNC(PyObject *) PyCodec_XMLCharRefReplaceErrors(PyObject *exc); |
| 171 | |
| 172 | /* replace the unicode encode error with backslash escapes (\x, \u and \U) */ |
| 173 | PyAPI_FUNC(PyObject *) PyCodec_BackslashReplaceErrors(PyObject *exc); |
| 174 | |
Guido van Rossum | 3094484 | 2000-03-10 22:32:23 +0000 | [diff] [blame] | 175 | #ifdef __cplusplus |
| 176 | } |
| 177 | #endif |
| 178 | #endif /* !Py_CODECREGISTRY_H */ |