Branch data Line data Source code
1 : : /*-------------------------------------------------------------------------
2 : : *
3 : : * pg_wchar.h
4 : : * multibyte-character support
5 : : *
6 : : * Portions Copyright (c) 1996-2026, PostgreSQL Global Development Group
7 : : * Portions Copyright (c) 1994, Regents of the University of California
8 : : *
9 : : * src/include/mb/pg_wchar.h
10 : : *
11 : : * NOTES
12 : : * This is used both by the backend and by frontends, but should not be
13 : : * included by libpq client programs. In particular, a libpq client
14 : : * should not assume that the encoding IDs used by the version of libpq
15 : : * it's linked to match up with the IDs declared here.
16 : : * To help prevent mistakes, relevant functions that are exported by
17 : : * libpq have a physically different name when being referenced
18 : : * statically.
19 : : *
20 : : *-------------------------------------------------------------------------
21 : : */
22 : : #ifndef PG_WCHAR_H
23 : : #define PG_WCHAR_H
24 : :
25 : : /*
26 : : * The pg_wchar type
27 : : */
28 : : typedef unsigned int pg_wchar;
29 : :
30 : : /*
31 : : * Maximum byte length of multibyte characters in any backend encoding
32 : : */
33 : : #define MAX_MULTIBYTE_CHAR_LEN 4
34 : :
35 : : /*
36 : : * various definitions for EUC
37 : : */
38 : : #define SS2 0x8e /* single shift 2 (JIS0201) */
39 : : #define SS3 0x8f /* single shift 3 (JIS0212) */
40 : :
41 : : /*
42 : : * EUC_TW planes
43 : : */
44 : : #define LC_CNS11643_1 0x95 /* CNS 11643-1992 Plane 1 */
45 : : #define LC_CNS11643_2 0x96 /* CNS 11643-1992 Plane 2 */
46 : : #define LC_CNS11643_3 0xf6 /* CNS 11643-1992 Plane 3 */
47 : : #define LC_CNS11643_4 0xf7 /* CNS 11643-1992 Plane 4 */
48 : : #define LC_CNS11643_5 0xf8 /* CNS 11643-1992 Plane 5 */
49 : : #define LC_CNS11643_6 0xf9 /* CNS 11643-1992 Plane 6 */
50 : : #define LC_CNS11643_7 0xfa /* CNS 11643-1992 Plane 7 */
51 : :
52 : : /*
53 : : * SJIS validation macros
54 : : */
55 : : #define ISSJISHEAD(c) (((c) >= 0x81 && (c) <= 0x9f) || ((c) >= 0xe0 && (c) <= 0xfc))
56 : : #define ISSJISTAIL(c) (((c) >= 0x40 && (c) <= 0x7e) || ((c) >= 0x80 && (c) <= 0xfc))
57 : :
58 : : /*
59 : : * PostgreSQL encoding identifiers
60 : : *
61 : : * WARNING: If you add some encoding don't forget to update
62 : : * the pg_enc2name_tbl[] array (in src/common/encnames.c),
63 : : * the pg_enc2gettext_tbl[] array (in src/common/encnames.c) and
64 : : * the pg_wchar_table[] array (in src/common/wchar.c) and to check
65 : : * PG_ENCODING_BE_LAST macro.
66 : : *
67 : : * PG_SQL_ASCII is default encoding and must be = 0.
68 : : *
69 : : * XXX We must avoid renumbering any backend encoding until libpq's major
70 : : * version number is increased beyond 5; it turns out that the backend
71 : : * encoding IDs are effectively part of libpq's ABI as far as 8.2 initdb and
72 : : * psql are concerned.
73 : : */
74 : : typedef enum pg_enc
75 : : {
76 : : PG_SQL_ASCII = 0, /* SQL/ASCII */
77 : : PG_EUC_JP, /* EUC for Japanese */
78 : : PG_EUC_CN, /* EUC for Chinese */
79 : : PG_EUC_KR, /* EUC for Korean */
80 : : PG_EUC_TW, /* EUC for Taiwan */
81 : : PG_EUC_JIS_2004, /* EUC-JIS-2004 */
82 : : PG_UTF8, /* Unicode UTF8 */
83 : : PG_UNUSED_1, /* (Was Mule internal code) */
84 : : PG_LATIN1, /* ISO-8859-1 Latin 1 */
85 : : PG_LATIN2, /* ISO-8859-2 Latin 2 */
86 : : PG_LATIN3, /* ISO-8859-3 Latin 3 */
87 : : PG_LATIN4, /* ISO-8859-4 Latin 4 */
88 : : PG_LATIN5, /* ISO-8859-9 Latin 5 */
89 : : PG_LATIN6, /* ISO-8859-10 Latin6 */
90 : : PG_LATIN7, /* ISO-8859-13 Latin7 */
91 : : PG_LATIN8, /* ISO-8859-14 Latin8 */
92 : : PG_LATIN9, /* ISO-8859-15 Latin9 */
93 : : PG_LATIN10, /* ISO-8859-16 Latin10 */
94 : : PG_WIN1256, /* windows-1256 */
95 : : PG_WIN1258, /* Windows-1258 */
96 : : PG_WIN866, /* (MS-DOS CP866) */
97 : : PG_WIN874, /* windows-874 */
98 : : PG_KOI8R, /* KOI8-R */
99 : : PG_WIN1251, /* windows-1251 */
100 : : PG_WIN1252, /* windows-1252 */
101 : : PG_ISO_8859_5, /* ISO-8859-5 */
102 : : PG_ISO_8859_6, /* ISO-8859-6 */
103 : : PG_ISO_8859_7, /* ISO-8859-7 */
104 : : PG_ISO_8859_8, /* ISO-8859-8 */
105 : : PG_WIN1250, /* windows-1250 */
106 : : PG_WIN1253, /* windows-1253 */
107 : : PG_WIN1254, /* windows-1254 */
108 : : PG_WIN1255, /* windows-1255 */
109 : : PG_WIN1257, /* windows-1257 */
110 : : PG_KOI8U, /* KOI8-U */
111 : : /* PG_ENCODING_BE_LAST points to the above entry */
112 : :
113 : : /* followings are for client encoding only */
114 : : PG_SJIS, /* Shift JIS (Windows-932) */
115 : : PG_BIG5, /* Big5 (Windows-950) */
116 : : PG_GBK, /* GBK (Windows-936) */
117 : : PG_UHC, /* UHC (Windows-949) */
118 : : PG_GB18030, /* GB18030 */
119 : : PG_JOHAB, /* EUC for Korean JOHAB */
120 : : PG_SHIFT_JIS_2004, /* Shift-JIS-2004 */
121 : : _PG_LAST_ENCODING_ /* mark only */
122 : :
123 : : } pg_enc;
124 : :
125 : : #define PG_ENCODING_BE_LAST PG_KOI8U
126 : :
127 : : #define PG_UNUSED_ENCODING(_enc) \
128 : : ((_enc) == PG_UNUSED_1)
129 : :
130 : : /*
131 : : * Please use these tests before access to pg_enc2name_tbl[]
132 : : * or to other places...
133 : : */
134 : : #define PG_VALID_BE_ENCODING(_enc) \
135 : : ((_enc) >= 0 && (_enc) <= PG_ENCODING_BE_LAST && !PG_UNUSED_ENCODING(_enc))
136 : :
137 : : #define PG_ENCODING_IS_CLIENT_ONLY(_enc) \
138 : : ((_enc) > PG_ENCODING_BE_LAST && (_enc) < _PG_LAST_ENCODING_)
139 : :
140 : : #define PG_VALID_ENCODING(_enc) \
141 : : ((_enc) >= 0 && (_enc) < _PG_LAST_ENCODING_ && !PG_UNUSED_ENCODING(_enc))
142 : :
143 : : /* On FE are possible all encodings */
144 : : #define PG_VALID_FE_ENCODING(_enc) PG_VALID_ENCODING(_enc)
145 : :
146 : : /*
147 : : * When converting strings between different encodings, we assume that space
148 : : * for converted result is 4-to-1 growth in the worst case. The rate for
149 : : * currently supported encoding pairs are within 3 (SJIS JIS X0201 half width
150 : : * kana -> UTF8 is the worst case). So "4" should be enough for the moment.
151 : : *
152 : : * Note that this is not the same as the maximum character width in any
153 : : * particular encoding.
154 : : */
155 : : #define MAX_CONVERSION_GROWTH 4
156 : :
157 : : /*
158 : : * Maximum byte length of a string that's required in any encoding to convert
159 : : * at least one character to any other encoding. In other words, if you feed
160 : : * MAX_CONVERSION_INPUT_LENGTH bytes to any encoding conversion function, it
161 : : * is guaranteed to be able to convert something without needing more input
162 : : * (assuming the input is valid).
163 : : *
164 : : * Currently, the maximum case is the conversion UTF8 -> SJIS JIS X0201 half
165 : : * width kana, where a pair of UTF-8 characters is converted into a single
166 : : * SHIFT_JIS_2004 character (the reverse of the worst case for
167 : : * MAX_CONVERSION_GROWTH). It needs 6 bytes of input. In theory, a
168 : : * user-defined conversion function might have more complicated cases, although
169 : : * for the reverse mapping you would probably also need to bump up
170 : : * MAX_CONVERSION_GROWTH. But there is no need to be stingy here, so make it
171 : : * generous.
172 : : */
173 : : #define MAX_CONVERSION_INPUT_LENGTH 16
174 : :
175 : : /*
176 : : * Maximum byte length of the string equivalent to any one Unicode code point,
177 : : * in any backend encoding. The current value assumes that a 4-byte UTF-8
178 : : * character might expand by MAX_CONVERSION_GROWTH, which is a huge
179 : : * overestimate. But in current usage we don't allocate large multiples of
180 : : * this, so there's little point in being stingy.
181 : : */
182 : : #define MAX_UNICODE_EQUIVALENT_STRING 16
183 : :
184 : : /*
185 : : * Table for mapping an encoding number to official encoding name and
186 : : * possibly other subsidiary data. Be careful to check encoding number
187 : : * before accessing a table entry!
188 : : *
189 : : * if (PG_VALID_ENCODING(encoding))
190 : : * pg_enc2name_tbl[ encoding ];
191 : : */
192 : : typedef struct pg_enc2name
193 : : {
194 : : const char *name;
195 : : pg_enc encoding;
196 : : #ifdef WIN32
197 : : unsigned codepage; /* codepage for WIN32 */
198 : : #endif
199 : : } pg_enc2name;
200 : :
201 : : extern PGDLLIMPORT const pg_enc2name pg_enc2name_tbl[];
202 : :
203 : : /*
204 : : * Encoding names for gettext
205 : : */
206 : : extern PGDLLIMPORT const char *pg_enc2gettext_tbl[];
207 : :
208 : : /*
209 : : * pg_wchar stuff
210 : : */
211 : : typedef int (*mb2wchar_with_len_converter) (const unsigned char *from,
212 : : pg_wchar *to,
213 : : int len);
214 : :
215 : : typedef int (*wchar2mb_with_len_converter) (const pg_wchar *from,
216 : : unsigned char *to,
217 : : int len);
218 : :
219 : : typedef int (*mblen_converter) (const unsigned char *mbstr);
220 : :
221 : : typedef int (*mbdisplaylen_converter) (const unsigned char *mbstr);
222 : :
223 : : typedef bool (*mbcharacter_incrementer) (unsigned char *mbstr, int len);
224 : :
225 : : typedef int (*mbchar_verifier) (const unsigned char *mbstr, int len);
226 : :
227 : : typedef int (*mbstr_verifier) (const unsigned char *mbstr, int len);
228 : :
229 : : typedef struct
230 : : {
231 : : mb2wchar_with_len_converter mb2wchar_with_len; /* convert a multibyte
232 : : * string to a wchar */
233 : : wchar2mb_with_len_converter wchar2mb_with_len; /* convert a wchar string
234 : : * to a multibyte */
235 : : mblen_converter mblen; /* get byte length of a char */
236 : : mbdisplaylen_converter dsplen; /* get display width of a char */
237 : : mbchar_verifier mbverifychar; /* verify multibyte character */
238 : : mbstr_verifier mbverifystr; /* verify multibyte string */
239 : : int maxmblen; /* max bytes for a char in this encoding */
240 : : } pg_wchar_tbl;
241 : :
242 : : extern PGDLLIMPORT const pg_wchar_tbl pg_wchar_table[];
243 : :
244 : : /*
245 : : * Data structures for conversions between UTF-8 and other encodings
246 : : * (UtfToLocal() and LocalToUtf()). In these data structures, characters of
247 : : * either encoding are represented by uint32 words; hence we can only support
248 : : * characters up to 4 bytes long. For example, the byte sequence 0xC2 0x89
249 : : * would be represented by 0x0000C289, and 0xE8 0xA2 0xB4 by 0x00E8A2B4.
250 : : *
251 : : * There are three possible ways a character can be mapped:
252 : : *
253 : : * 1. Using a radix tree, from source to destination code.
254 : : * 2. Using a sorted array of source -> destination code pairs. This
255 : : * method is used for "combining" characters. There are so few of
256 : : * them that building a radix tree would be wasteful.
257 : : * 3. Using a conversion function.
258 : : */
259 : :
260 : : /*
261 : : * Radix tree for character conversion.
262 : : *
263 : : * Logically, this is actually four different radix trees, for 1-byte,
264 : : * 2-byte, 3-byte and 4-byte inputs. The 1-byte tree is a simple lookup
265 : : * table from source to target code. The 2-byte tree consists of two levels:
266 : : * one lookup table for the first byte, where the value in the lookup table
267 : : * points to a lookup table for the second byte. And so on.
268 : : *
269 : : * Physically, all the trees are stored in one big array, in 'chars16' or
270 : : * 'chars32', depending on the maximum value that needs to be represented. For
271 : : * each level in each tree, we also store lower and upper bound of allowed
272 : : * values - values outside those bounds are considered invalid, and are left
273 : : * out of the tables.
274 : : *
275 : : * In the intermediate levels of the trees, the values stored are offsets
276 : : * into the chars[16|32] array.
277 : : *
278 : : * In the beginning of the chars[16|32] array, there is always a number of
279 : : * zeros, so that you safely follow an index from an intermediate table
280 : : * without explicitly checking for a zero. Following a zero any number of
281 : : * times will always bring you to the dummy, all-zeros table in the
282 : : * beginning. This helps to shave some cycles when looking up values.
283 : : */
284 : : typedef struct
285 : : {
286 : : /*
287 : : * Array containing all the values. Only one of chars16 or chars32 is
288 : : * used, depending on how wide the values we need to represent are.
289 : : */
290 : : const uint16 *chars16;
291 : : const uint32 *chars32;
292 : :
293 : : /* Radix tree for 1-byte inputs */
294 : : uint32 b1root; /* offset of table in the chars[16|32] array */
295 : : uint8 b1_lower; /* min allowed value for a single byte input */
296 : : uint8 b1_upper; /* max allowed value for a single byte input */
297 : :
298 : : /* Radix tree for 2-byte inputs */
299 : : uint32 b2root; /* offset of 1st byte's table */
300 : : uint8 b2_1_lower; /* min/max allowed value for 1st input byte */
301 : : uint8 b2_1_upper;
302 : : uint8 b2_2_lower; /* min/max allowed value for 2nd input byte */
303 : : uint8 b2_2_upper;
304 : :
305 : : /* Radix tree for 3-byte inputs */
306 : : uint32 b3root; /* offset of 1st byte's table */
307 : : uint8 b3_1_lower; /* min/max allowed value for 1st input byte */
308 : : uint8 b3_1_upper;
309 : : uint8 b3_2_lower; /* min/max allowed value for 2nd input byte */
310 : : uint8 b3_2_upper;
311 : : uint8 b3_3_lower; /* min/max allowed value for 3rd input byte */
312 : : uint8 b3_3_upper;
313 : :
314 : : /* Radix tree for 4-byte inputs */
315 : : uint32 b4root; /* offset of 1st byte's table */
316 : : uint8 b4_1_lower; /* min/max allowed value for 1st input byte */
317 : : uint8 b4_1_upper;
318 : : uint8 b4_2_lower; /* min/max allowed value for 2nd input byte */
319 : : uint8 b4_2_upper;
320 : : uint8 b4_3_lower; /* min/max allowed value for 3rd input byte */
321 : : uint8 b4_3_upper;
322 : : uint8 b4_4_lower; /* min/max allowed value for 4th input byte */
323 : : uint8 b4_4_upper;
324 : :
325 : : } pg_mb_radix_tree;
326 : :
327 : : /*
328 : : * UTF-8 to local code conversion map (for combined characters)
329 : : */
330 : : typedef struct
331 : : {
332 : : uint32 utf1; /* UTF-8 code 1 */
333 : : uint32 utf2; /* UTF-8 code 2 */
334 : : uint32 code; /* local code */
335 : : } pg_utf_to_local_combined;
336 : :
337 : : /*
338 : : * local code to UTF-8 conversion map (for combined characters)
339 : : */
340 : : typedef struct
341 : : {
342 : : uint32 code; /* local code */
343 : : uint32 utf1; /* UTF-8 code 1 */
344 : : uint32 utf2; /* UTF-8 code 2 */
345 : : } pg_local_to_utf_combined;
346 : :
347 : : /*
348 : : * callback function for algorithmic encoding conversions (in either direction)
349 : : *
350 : : * if function returns zero, it does not know how to convert the code
351 : : */
352 : : typedef uint32 (*utf_local_conversion_func) (uint32 code);
353 : :
354 : : /*
355 : : * Support macro for encoding conversion functions to validate their
356 : : * arguments. (This could be made more compact if we included fmgr.h
357 : : * here, but we don't want to do that because this header file is also
358 : : * used by frontends.)
359 : : */
360 : : #define CHECK_ENCODING_CONVERSION_ARGS(srcencoding,destencoding) \
361 : : check_encoding_conversion_args(PG_GETARG_INT32(0), \
362 : : PG_GETARG_INT32(1), \
363 : : PG_GETARG_INT32(4), \
364 : : (srcencoding), \
365 : : (destencoding))
366 : :
367 : :
368 : : /*
369 : : * Some handy functions for Unicode-specific tests.
370 : : */
371 : : static inline bool
372 : 1793 : is_valid_unicode_codepoint(char32_t c)
373 : : {
374 [ + - + + ]: 1793 : return (c > 0 && c <= 0x10FFFF);
375 : : }
376 : :
377 : : static inline bool
378 : 1224 : is_utf16_surrogate_first(char32_t c)
379 : : {
380 [ + + + + ]: 1224 : return (c >= 0xD800 && c <= 0xDBFF);
381 : : }
382 : :
383 : : static inline bool
384 : 1082 : is_utf16_surrogate_second(char32_t c)
385 : : {
386 [ + + + + ]: 1082 : return (c >= 0xDC00 && c <= 0xDFFF);
387 : : }
388 : :
389 : : static inline char32_t
390 : 42 : surrogate_pair_to_codepoint(char16_t first, char16_t second)
391 : : {
392 : 42 : return ((first & 0x3FF) << 10) + 0x10000 + (second & 0x3FF);
393 : : }
394 : :
395 : : /*
396 : : * Convert a UTF-8 character to a Unicode code point.
397 : : * This is a one-character version of pg_utf2wchar_with_len.
398 : : *
399 : : * No error checks here, c must point to a long-enough string.
400 : : */
401 : : static inline char32_t
402 : 30613575 : utf8_to_unicode(const unsigned char *c)
403 : : {
404 [ + + ]: 30613575 : if ((*c & 0x80) == 0)
405 : 30606262 : return (char32_t) c[0];
406 [ + + ]: 7313 : else if ((*c & 0xe0) == 0xc0)
407 : 6093 : return (char32_t) (((c[0] & 0x1f) << 6) |
408 : 6093 : (c[1] & 0x3f));
409 [ + + ]: 1220 : else if ((*c & 0xf0) == 0xe0)
410 : 1155 : return (char32_t) (((c[0] & 0x0f) << 12) |
411 : 1155 : ((c[1] & 0x3f) << 6) |
412 : 1155 : (c[2] & 0x3f));
413 [ + - ]: 65 : else if ((*c & 0xf8) == 0xf0)
414 : 65 : return (char32_t) (((c[0] & 0x07) << 18) |
415 : 65 : ((c[1] & 0x3f) << 12) |
416 : 65 : ((c[2] & 0x3f) << 6) |
417 : 65 : (c[3] & 0x3f));
418 : : else
419 : : /* that is an invalid code on purpose */
420 : 0 : return 0xffffffff;
421 : : }
422 : :
423 : : /*
424 : : * Map a Unicode code point to UTF-8. utf8string must have at least
425 : : * unicode_utf8len(c) bytes available.
426 : : */
427 : : static inline unsigned char *
428 : 8299259 : unicode_to_utf8(char32_t c, unsigned char *utf8string)
429 : : {
430 [ + + ]: 8299259 : if (c <= 0x7F)
431 : : {
432 : 8296794 : utf8string[0] = c;
433 : : }
434 [ + + ]: 2465 : else if (c <= 0x7FF)
435 : : {
436 : 1572 : utf8string[0] = 0xC0 | ((c >> 6) & 0x1F);
437 : 1572 : utf8string[1] = 0x80 | (c & 0x3F);
438 : : }
439 [ + + ]: 893 : else if (c <= 0xFFFF)
440 : : {
441 : 839 : utf8string[0] = 0xE0 | ((c >> 12) & 0x0F);
442 : 839 : utf8string[1] = 0x80 | ((c >> 6) & 0x3F);
443 : 839 : utf8string[2] = 0x80 | (c & 0x3F);
444 : : }
445 : : else
446 : : {
447 : 54 : utf8string[0] = 0xF0 | ((c >> 18) & 0x07);
448 : 54 : utf8string[1] = 0x80 | ((c >> 12) & 0x3F);
449 : 54 : utf8string[2] = 0x80 | ((c >> 6) & 0x3F);
450 : 54 : utf8string[3] = 0x80 | (c & 0x3F);
451 : : }
452 : :
453 : 8299259 : return utf8string;
454 : : }
455 : :
456 : : /*
457 : : * Number of bytes needed to represent the given char in UTF8.
458 : : */
459 : : static inline int
460 : 329105 : unicode_utf8len(char32_t c)
461 : : {
462 [ + + ]: 329105 : if (c <= 0x7F)
463 : 327801 : return 1;
464 [ + + ]: 1304 : else if (c <= 0x7FF)
465 : 984 : return 2;
466 [ + - ]: 320 : else if (c <= 0xFFFF)
467 : 320 : return 3;
468 : : else
469 : 0 : return 4;
470 : : }
471 : :
472 : : /*
473 : : * The functions in this list are exported by libpq, and we need to be sure
474 : : * that we know which calls are satisfied by libpq and which are satisfied
475 : : * by static linkage to libpgcommon. (This is because we might be using a
476 : : * libpq.so that's of a different major version and has encoding IDs that
477 : : * differ from the current version's.) The nominal function names are what
478 : : * are actually used in and exported by libpq, while the names exported by
479 : : * libpgcommon.a and libpgcommon_srv.a end in "_private".
480 : : */
481 : : #if defined(USE_PRIVATE_ENCODING_FUNCS) || !defined(FRONTEND)
482 : : #define pg_char_to_encoding pg_char_to_encoding_private
483 : : #define pg_encoding_to_char pg_encoding_to_char_private
484 : : #define pg_valid_server_encoding pg_valid_server_encoding_private
485 : : #define pg_valid_server_encoding_id pg_valid_server_encoding_id_private
486 : : #define pg_utf_mblen pg_utf_mblen_private
487 : : #endif
488 : :
489 : : /*
490 : : * These functions are considered part of libpq's exported API and
491 : : * are also declared in libpq-fe.h.
492 : : */
493 : : extern int pg_char_to_encoding(const char *name);
494 : : extern const char *pg_encoding_to_char(int encoding);
495 : : extern int pg_valid_server_encoding_id(int encoding);
496 : :
497 : : /*
498 : : * These functions are available to frontend code that links with libpgcommon
499 : : * (in addition to the ones just above). The constant tables declared
500 : : * earlier in this file are also available from libpgcommon.
501 : : */
502 : : extern void pg_encoding_set_invalid(int encoding, char *dst);
503 : : extern int pg_encoding_mblen(int encoding, const char *mbstr);
504 : : extern int pg_encoding_mblen_or_incomplete(int encoding, const char *mbstr,
505 : : size_t remaining);
506 : : extern int pg_encoding_mblen_bounded(int encoding, const char *mbstr);
507 : : extern int pg_encoding_dsplen(int encoding, const char *mbstr);
508 : : extern int pg_encoding_verifymbchar(int encoding, const char *mbstr, int len);
509 : : extern int pg_encoding_verifymbstr(int encoding, const char *mbstr, int len);
510 : : extern int pg_encoding_max_length(int encoding);
511 : : extern int pg_valid_client_encoding(const char *name);
512 : : extern int pg_valid_server_encoding(const char *name);
513 : : extern bool is_encoding_supported_by_icu(int encoding);
514 : : extern const char *get_encoding_name_for_icu(int encoding);
515 : :
516 : : extern bool pg_utf8_islegal(const unsigned char *source, int length);
517 : : extern int pg_utf_mblen(const unsigned char *s);
518 : :
519 : : /*
520 : : * The remaining functions are backend-only.
521 : : */
522 : : extern int pg_mb2wchar(const char *from, pg_wchar *to);
523 : : extern int pg_mb2wchar_with_len(const char *from, pg_wchar *to, int len);
524 : : extern int pg_encoding_mb2wchar_with_len(int encoding,
525 : : const char *from, pg_wchar *to, int len);
526 : : extern int pg_wchar2mb(const pg_wchar *from, char *to);
527 : : extern int pg_wchar2mb_with_len(const pg_wchar *from, char *to, int len);
528 : : extern int pg_encoding_wchar2mb_with_len(int encoding,
529 : : const pg_wchar *from, char *to, int len);
530 : : extern int pg_char_and_wchar_strcmp(const char *s1, const pg_wchar *s2);
531 : : extern int pg_wchar_strncmp(const pg_wchar *s1, const pg_wchar *s2, size_t n);
532 : : extern int pg_char_and_wchar_strncmp(const char *s1, const pg_wchar *s2, size_t n);
533 : : extern size_t pg_wchar_strlen(const pg_wchar *str);
534 : : extern int pg_mblen_cstr(const char *mbstr);
535 : : extern int pg_mblen_range(const char *mbstr, const char *end);
536 : : extern int pg_mblen_with_len(const char *mbstr, int limit);
537 : : extern int pg_mblen_unbounded(const char *mbstr);
538 : :
539 : : /* deprecated */
540 : : extern int pg_mblen(const char *mbstr);
541 : :
542 : : extern int pg_dsplen(const char *mbstr);
543 : : extern int pg_mbstrlen(const char *mbstr);
544 : : extern int pg_mbstrlen_with_len(const char *mbstr, int limit);
545 : : extern int pg_mbcliplen(const char *mbstr, int len, int limit);
546 : : extern int pg_encoding_mbcliplen(int encoding, const char *mbstr,
547 : : int len, int limit);
548 : : extern int pg_mbcharcliplen(const char *mbstr, int len, int limit);
549 : : extern int pg_database_encoding_max_length(void);
550 : : extern mbcharacter_incrementer pg_database_encoding_character_incrementer(void);
551 : :
552 : : extern int PrepareClientEncoding(int encoding);
553 : : extern int SetClientEncoding(int encoding);
554 : : extern void InitializeClientEncoding(void);
555 : : extern int pg_get_client_encoding(void);
556 : : extern const char *pg_get_client_encoding_name(void);
557 : :
558 : : extern void SetDatabaseEncoding(int encoding);
559 : : extern int GetDatabaseEncoding(void);
560 : : extern const char *GetDatabaseEncodingName(void);
561 : : extern void SetMessageEncoding(int encoding);
562 : : extern int GetMessageEncoding(void);
563 : :
564 : : #ifdef ENABLE_NLS
565 : : extern int pg_bind_textdomain_codeset(const char *domainname);
566 : : #endif
567 : :
568 : : extern unsigned char *pg_do_encoding_conversion(unsigned char *src, int len,
569 : : int src_encoding,
570 : : int dest_encoding);
571 : : extern int pg_do_encoding_conversion_buf(Oid proc,
572 : : int src_encoding,
573 : : int dest_encoding,
574 : : unsigned char *src, int srclen,
575 : : unsigned char *dest, int destlen,
576 : : bool noError);
577 : :
578 : : extern char *pg_client_to_server(const char *s, int len);
579 : : extern char *pg_server_to_client(const char *s, int len);
580 : : extern char *pg_any_to_server(const char *s, int len, int encoding);
581 : : extern char *pg_server_to_any(const char *s, int len, int encoding);
582 : :
583 : : extern void pg_unicode_to_server(char32_t c, unsigned char *s);
584 : : extern bool pg_unicode_to_server_noerror(char32_t c, unsigned char *s);
585 : :
586 : : extern unsigned short BIG5toCNS(unsigned short big5, unsigned char *lc);
587 : : extern unsigned short CNStoBIG5(unsigned short cns, unsigned char lc);
588 : :
589 : : extern int UtfToLocal(const unsigned char *utf, int len,
590 : : unsigned char *iso,
591 : : const pg_mb_radix_tree *map,
592 : : const pg_utf_to_local_combined *cmap, int cmapsize,
593 : : utf_local_conversion_func conv_func,
594 : : int encoding, bool noError);
595 : : extern int LocalToUtf(const unsigned char *iso, int len,
596 : : unsigned char *utf,
597 : : const pg_mb_radix_tree *map,
598 : : const pg_local_to_utf_combined *cmap, int cmapsize,
599 : : utf_local_conversion_func conv_func,
600 : : int encoding, bool noError);
601 : :
602 : : extern bool pg_verifymbstr(const char *mbstr, int len, bool noError);
603 : : extern bool pg_verify_mbstr(int encoding, const char *mbstr, int len,
604 : : bool noError);
605 : : extern int pg_verify_mbstr_len(int encoding, const char *mbstr, int len,
606 : : bool noError);
607 : :
608 : : extern void check_encoding_conversion_args(int src_encoding,
609 : : int dest_encoding,
610 : : int len,
611 : : int expected_src_encoding,
612 : : int expected_dest_encoding);
613 : :
614 : : pg_noreturn extern void report_invalid_encoding(int encoding, const char *mbstr, int len);
615 : : pg_noreturn extern void report_untranslatable_char(int src_encoding, int dest_encoding,
616 : : const char *mbstr, int len);
617 : :
618 : : extern int local2local(const unsigned char *l, unsigned char *p, int len,
619 : : int src_encoding, int dest_encoding,
620 : : const unsigned char *tab, bool noError);
621 : :
622 : : #ifdef WIN32
623 : : extern WCHAR *pgwin32_message_to_UTF16(const char *str, int len, int *utf16len);
624 : : #endif
625 : :
626 : : #endif /* PG_WCHAR_H */
|
