Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 1 | /* |
| 2 | * multibytecodec.h: Common Multibyte Codec Implementation |
| 3 | * |
| 4 | * Written by Hye-Shik Chang <perky@FreeBSD.org> |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 5 | */ |
| 6 | |
| 7 | #ifndef _PYTHON_MULTIBYTECODEC_H_ |
| 8 | #define _PYTHON_MULTIBYTECODEC_H_ |
| 9 | #ifdef __cplusplus |
| 10 | extern "C" { |
| 11 | #endif |
| 12 | |
Hye-Shik Chang | 2bb146f | 2004-07-18 03:06:29 +0000 | [diff] [blame] | 13 | #ifdef uint16_t |
| 14 | typedef uint16_t ucs2_t, DBCHAR; |
| 15 | #else |
| 16 | typedef unsigned short ucs2_t, DBCHAR; |
| 17 | #endif |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 18 | |
Christopher Thorne | ac22f6a | 2018-11-01 10:48:49 +0000 | [diff] [blame] | 19 | /* |
| 20 | * A struct that provides 8 bytes of state for multibyte |
| 21 | * codecs. Codecs are free to use this how they want. Note: if you |
| 22 | * need to add a new field to this struct, ensure that its byte order |
| 23 | * is independent of CPU endianness so that the return value of |
| 24 | * getstate doesn't differ between little and big endian CPUs. |
| 25 | */ |
| 26 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 27 | unsigned char c[8]; |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 28 | } MultibyteCodec_State; |
| 29 | |
Hye-Shik Chang | 2bb146f | 2004-07-18 03:06:29 +0000 | [diff] [blame] | 30 | typedef int (*mbcodec_init)(const void *config); |
Hye-Shik Chang | 4b96c13 | 2006-03-04 16:08:19 +0000 | [diff] [blame] | 31 | typedef Py_ssize_t (*mbencode_func)(MultibyteCodec_State *state, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 32 | const void *config, |
Victor Stinner | d949126 | 2013-04-14 02:06:32 +0200 | [diff] [blame] | 33 | int kind, void *data, |
| 34 | Py_ssize_t *inpos, Py_ssize_t inlen, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 35 | unsigned char **outbuf, Py_ssize_t outleft, |
| 36 | int flags); |
Hye-Shik Chang | 2bb146f | 2004-07-18 03:06:29 +0000 | [diff] [blame] | 37 | typedef int (*mbencodeinit_func)(MultibyteCodec_State *state, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 38 | const void *config); |
Hye-Shik Chang | 4b96c13 | 2006-03-04 16:08:19 +0000 | [diff] [blame] | 39 | typedef Py_ssize_t (*mbencodereset_func)(MultibyteCodec_State *state, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 40 | const void *config, |
| 41 | unsigned char **outbuf, Py_ssize_t outleft); |
Hye-Shik Chang | 4b96c13 | 2006-03-04 16:08:19 +0000 | [diff] [blame] | 42 | typedef Py_ssize_t (*mbdecode_func)(MultibyteCodec_State *state, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 43 | const void *config, |
| 44 | const unsigned char **inbuf, Py_ssize_t inleft, |
Victor Stinner | a0dd021 | 2013-04-11 22:09:04 +0200 | [diff] [blame] | 45 | _PyUnicodeWriter *writer); |
Hye-Shik Chang | 2bb146f | 2004-07-18 03:06:29 +0000 | [diff] [blame] | 46 | typedef int (*mbdecodeinit_func)(MultibyteCodec_State *state, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 47 | const void *config); |
Hye-Shik Chang | 4b96c13 | 2006-03-04 16:08:19 +0000 | [diff] [blame] | 48 | typedef Py_ssize_t (*mbdecodereset_func)(MultibyteCodec_State *state, |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 49 | const void *config); |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 50 | |
| 51 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 52 | const char *encoding; |
| 53 | const void *config; |
| 54 | mbcodec_init codecinit; |
| 55 | mbencode_func encode; |
| 56 | mbencodeinit_func encinit; |
| 57 | mbencodereset_func encreset; |
| 58 | mbdecode_func decode; |
| 59 | mbdecodeinit_func decinit; |
| 60 | mbdecodereset_func decreset; |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 61 | } MultibyteCodec; |
| 62 | |
| 63 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 64 | PyObject_HEAD |
| 65 | MultibyteCodec *codec; |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 66 | } MultibyteCodecObject; |
| 67 | |
Andy Lester | dffe4c0 | 2020-03-04 07:15:20 -0600 | [diff] [blame] | 68 | #define MultibyteCodec_Check(op) Py_IS_TYPE((op), &MultibyteCodec_Type) |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 69 | |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 70 | #define _MultibyteStatefulCodec_HEAD \ |
| 71 | PyObject_HEAD \ |
| 72 | MultibyteCodec *codec; \ |
| 73 | MultibyteCodec_State state; \ |
| 74 | PyObject *errors; |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 75 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 76 | _MultibyteStatefulCodec_HEAD |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 77 | } MultibyteStatefulCodecContext; |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 78 | |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 79 | #define MAXENCPENDING 2 |
| 80 | #define _MultibyteStatefulEncoder_HEAD \ |
| 81 | _MultibyteStatefulCodec_HEAD \ |
Victor Stinner | d949126 | 2013-04-14 02:06:32 +0200 | [diff] [blame] | 82 | PyObject *pending; |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 83 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 84 | _MultibyteStatefulEncoder_HEAD |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 85 | } MultibyteStatefulEncoderContext; |
| 86 | |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 87 | #define MAXDECPENDING 8 |
| 88 | #define _MultibyteStatefulDecoder_HEAD \ |
| 89 | _MultibyteStatefulCodec_HEAD \ |
| 90 | unsigned char pending[MAXDECPENDING]; \ |
| 91 | Py_ssize_t pendingsize; |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 92 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 93 | _MultibyteStatefulDecoder_HEAD |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 94 | } MultibyteStatefulDecoderContext; |
| 95 | |
| 96 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 97 | _MultibyteStatefulEncoder_HEAD |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 98 | } MultibyteIncrementalEncoderObject; |
| 99 | |
| 100 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 101 | _MultibyteStatefulDecoder_HEAD |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 102 | } MultibyteIncrementalDecoderObject; |
| 103 | |
| 104 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 105 | _MultibyteStatefulDecoder_HEAD |
| 106 | PyObject *stream; |
Thomas Wouters | 49fd7fa | 2006-04-21 10:40:58 +0000 | [diff] [blame] | 107 | } MultibyteStreamReaderObject; |
| 108 | |
| 109 | typedef struct { |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 110 | _MultibyteStatefulEncoder_HEAD |
| 111 | PyObject *stream; |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 112 | } MultibyteStreamWriterObject; |
| 113 | |
| 114 | /* positive values for illegal sequences */ |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 115 | #define MBERR_TOOSMALL (-1) /* insufficient output buffer space */ |
| 116 | #define MBERR_TOOFEW (-2) /* incomplete input buffer */ |
| 117 | #define MBERR_INTERNAL (-3) /* internal runtime error */ |
Victor Stinner | d1f9942 | 2013-07-16 21:41:43 +0200 | [diff] [blame] | 118 | #define MBERR_EXCEPTION (-4) /* an exception has been raised */ |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 119 | |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 120 | #define ERROR_STRICT (PyObject *)(1) |
| 121 | #define ERROR_IGNORE (PyObject *)(2) |
| 122 | #define ERROR_REPLACE (PyObject *)(3) |
| 123 | #define ERROR_ISCUSTOM(p) ((p) < ERROR_STRICT || ERROR_REPLACE < (p)) |
Victor Stinner | 14c9fea | 2013-10-29 00:19:27 +0100 | [diff] [blame] | 124 | #define ERROR_DECREF(p) \ |
| 125 | do { \ |
| 126 | if (p != NULL && ERROR_ISCUSTOM(p)) \ |
| 127 | Py_DECREF(p); \ |
| 128 | } while (0); |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 129 | |
Antoine Pitrou | f95a1b3 | 2010-05-09 15:52:27 +0000 | [diff] [blame] | 130 | #define MBENC_FLUSH 0x0001 /* encode all characters encodable */ |
| 131 | #define MBENC_MAX MBENC_FLUSH |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 132 | |
Benjamin Peterson | b173f78 | 2009-05-05 22:31:58 +0000 | [diff] [blame] | 133 | #define PyMultibyteCodec_CAPSULE_NAME "multibytecodec.__map_*" |
| 134 | |
| 135 | |
Hye-Shik Chang | 3e2a306 | 2004-01-17 14:29:29 +0000 | [diff] [blame] | 136 | #ifdef __cplusplus |
| 137 | } |
| 138 | #endif |
| 139 | #endif |