2 * conversion functions between pg_wchar and multibyte streams.
4 * $PostgreSQL: pgsql/src/backend/utils/mb/wchar.c,v 1.74 2010/01/04 20:38:31 adunstan Exp $
7 /* can be used in either frontend or backend */
9 #include "postgres_fe.h"
10 #define Assert(condition)
15 #include "mb/pg_wchar.h"
19 * conversion to pg_wchar is done by "table driven."
20 * to add an encoding support, define mb2wchar_with_len(), mblen(), dsplen()
21 * for the particular encoding. Note that if the encoding is only
22 * supported in the client, you don't need to define
23 * mb2wchar_with_len() function (SJIS is the case).
25 * These functions generally assume that their input is validly formed.
26 * The "verifier" functions, further down in the file, have to be more
27 * paranoid. We expect that mblen() does not need to examine more than
28 * the first byte of the character to discover the correct length.
30 * Note: for the display output of psql to work properly, the return values
31 * of the dsplen functions must conform to the Unicode standard. In particular
32 * the NUL character is zero width and control characters are generally
33 * width -1. It is recommended that non-ASCII encodings refer their ASCII
34 * subset to the ASCII routines to ensure consistency.
41 pg_ascii2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
45 while (len > 0 && *from)
56 pg_ascii_mblen(const unsigned char *s)
62 pg_ascii_dsplen(const unsigned char *s)
66 if (*s < 0x20 || *s == 0x7f)
76 pg_euc2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
80 while (len > 0 && *from)
82 if (*from == SS2 && len >= 2) /* JIS X 0201 (so called "1 byte
86 *to = (SS2 << 8) | *from++;
89 else if (*from == SS3 && len >= 3) /* JIS X 0212 KANJI */
92 *to = (SS3 << 16) | (*from++ << 8);
96 else if (IS_HIGHBIT_SET(*from) && len >= 2) /* JIS X 0208 KANJI */
116 pg_euc_mblen(const unsigned char *s)
124 else if (IS_HIGHBIT_SET(*s))
132 pg_euc_dsplen(const unsigned char *s)
140 else if (IS_HIGHBIT_SET(*s))
143 len = pg_ascii_dsplen(s);
151 pg_eucjp2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
153 return pg_euc2wchar_with_len(from, to, len);
157 pg_eucjp_mblen(const unsigned char *s)
159 return pg_euc_mblen(s);
163 pg_eucjp_dsplen(const unsigned char *s)
171 else if (IS_HIGHBIT_SET(*s))
174 len = pg_ascii_dsplen(s);
182 pg_euckr2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
184 return pg_euc2wchar_with_len(from, to, len);
188 pg_euckr_mblen(const unsigned char *s)
190 return pg_euc_mblen(s);
194 pg_euckr_dsplen(const unsigned char *s)
196 return pg_euc_dsplen(s);
204 pg_euccn2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
208 while (len > 0 && *from)
210 if (*from == SS2 && len >= 3) /* code set 2 (unused?) */
213 *to = (SS2 << 16) | (*from++ << 8);
217 else if (*from == SS3 && len >= 3) /* code set 3 (unsed ?) */
220 *to = (SS3 << 16) | (*from++ << 8);
224 else if (IS_HIGHBIT_SET(*from) && len >= 2) /* code set 1 */
243 pg_euccn_mblen(const unsigned char *s)
247 if (IS_HIGHBIT_SET(*s))
255 pg_euccn_dsplen(const unsigned char *s)
259 if (IS_HIGHBIT_SET(*s))
262 len = pg_ascii_dsplen(s);
271 pg_euctw2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
275 while (len > 0 && *from)
277 if (*from == SS2 && len >= 4) /* code set 2 */
280 *to = (((uint32) SS2) << 24) | (*from++ << 16);
285 else if (*from == SS3 && len >= 3) /* code set 3 (unused?) */
288 *to = (SS3 << 16) | (*from++ << 8);
292 else if (IS_HIGHBIT_SET(*from) && len >= 2) /* code set 2 */
311 pg_euctw_mblen(const unsigned char *s)
319 else if (IS_HIGHBIT_SET(*s))
327 pg_euctw_dsplen(const unsigned char *s)
335 else if (IS_HIGHBIT_SET(*s))
338 len = pg_ascii_dsplen(s);
346 pg_johab_mblen(const unsigned char *s)
348 return pg_euc_mblen(s);
352 pg_johab_dsplen(const unsigned char *s)
354 return pg_euc_dsplen(s);
358 * convert UTF8 string to pg_wchar (UCS-4)
359 * caller must allocate enough space for "to", including a trailing zero!
360 * len: length of from.
361 * "from" not necessarily null terminated.
364 pg_utf2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
372 while (len > 0 && *from)
374 if ((*from & 0x80) == 0)
379 else if ((*from & 0xe0) == 0xc0)
382 break; /* drop trailing incomplete char */
385 *to = (c1 << 6) | c2;
388 else if ((*from & 0xf0) == 0xe0)
391 break; /* drop trailing incomplete char */
395 *to = (c1 << 12) | (c2 << 6) | c3;
398 else if ((*from & 0xf8) == 0xf0)
401 break; /* drop trailing incomplete char */
406 *to = (c1 << 18) | (c2 << 12) | (c3 << 6) | c4;
411 /* treat a bogus char as length 1; not ours to raise error */
424 * Map a Unicode code point to UTF-8. utf8string must have 4 bytes of
428 unicode_to_utf8(pg_wchar c, unsigned char *utf8string)
436 utf8string[0] = 0xC0 | ((c >> 6) & 0x1F);
437 utf8string[1] = 0x80 | (c & 0x3F);
439 else if (c <= 0xFFFF)
441 utf8string[0] = 0xE0 | ((c >> 12) & 0x0F);
442 utf8string[1] = 0x80 | ((c >> 6) & 0x3F);
443 utf8string[2] = 0x80 | (c & 0x3F);
447 utf8string[0] = 0xF0 | ((c >> 18) & 0x07);
448 utf8string[1] = 0x80 | ((c >> 12) & 0x3F);
449 utf8string[2] = 0x80 | ((c >> 6) & 0x3F);
450 utf8string[3] = 0x80 | (c & 0x3F);
458 * Return the byte length of a UTF8 character pointed to by s
460 * Note: in the current implementation we do not support UTF8 sequences
461 * of more than 4 bytes; hence do NOT return a value larger than 4.
462 * We return "1" for any leading byte that is either flat-out illegal or
463 * indicates a length larger than we support.
465 * pg_utf2wchar_with_len(), utf2ucs(), pg_utf8_islegal(), and perhaps
466 * other places would need to be fixed to change this.
469 pg_utf_mblen(const unsigned char *s)
473 if ((*s & 0x80) == 0)
475 else if ((*s & 0xe0) == 0xc0)
477 else if ((*s & 0xf0) == 0xe0)
479 else if ((*s & 0xf8) == 0xf0)
482 else if ((*s & 0xfc) == 0xf8)
484 else if ((*s & 0xfe) == 0xfc)
493 * This is an implementation of wcwidth() and wcswidth() as defined in
494 * "The Single UNIX Specification, Version 2, The Open Group, 1997"
495 * <http://www.UNIX-systems.org/online.html>
497 * Markus Kuhn -- 2001-09-08 -- public domain
499 * customised for PostgreSQL
501 * original available at : http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
506 unsigned short first;
510 /* auxiliary function for binary search in interval table */
512 mbbisearch(pg_wchar ucs, const struct mbinterval * table, int max)
517 if (ucs < table[0].first || ucs > table[max].last)
521 mid = (min + max) / 2;
522 if (ucs > table[mid].last)
524 else if (ucs < table[mid].first)
534 /* The following functions define the column width of an ISO 10646
535 * character as follows:
537 * - The null character (U+0000) has a column width of 0.
539 * - Other C0/C1 control characters and DEL will lead to a return
542 * - Non-spacing and enclosing combining characters (general
543 * category code Mn or Me in the Unicode database) have a
546 * - Other format characters (general category code Cf in the Unicode
547 * database) and ZERO WIDTH SPACE (U+200B) have a column width of 0.
549 * - Hangul Jamo medial vowels and final consonants (U+1160-U+11FF)
550 * have a column width of 0.
552 * - Spacing characters in the East Asian Wide (W) or East Asian
553 * FullWidth (F) category as defined in Unicode Technical
554 * Report #11 have a column width of 2.
556 * - All remaining characters (including all printable
557 * ISO 8859-1 and WGL4 characters, Unicode control characters,
558 * etc.) have a column width of 1.
560 * This implementation assumes that wchar_t characters are encoded
565 ucs_wcwidth(pg_wchar ucs)
567 /* sorted list of non-overlapping intervals of non-spacing characters */
568 static const struct mbinterval combining[] = {
569 {0x0300, 0x034E}, {0x0360, 0x0362}, {0x0483, 0x0486},
570 {0x0488, 0x0489}, {0x0591, 0x05A1}, {0x05A3, 0x05B9},
571 {0x05BB, 0x05BD}, {0x05BF, 0x05BF}, {0x05C1, 0x05C2},
572 {0x05C4, 0x05C4}, {0x064B, 0x0655}, {0x0670, 0x0670},
573 {0x06D6, 0x06E4}, {0x06E7, 0x06E8}, {0x06EA, 0x06ED},
574 {0x070F, 0x070F}, {0x0711, 0x0711}, {0x0730, 0x074A},
575 {0x07A6, 0x07B0}, {0x0901, 0x0902}, {0x093C, 0x093C},
576 {0x0941, 0x0948}, {0x094D, 0x094D}, {0x0951, 0x0954},
577 {0x0962, 0x0963}, {0x0981, 0x0981}, {0x09BC, 0x09BC},
578 {0x09C1, 0x09C4}, {0x09CD, 0x09CD}, {0x09E2, 0x09E3},
579 {0x0A02, 0x0A02}, {0x0A3C, 0x0A3C}, {0x0A41, 0x0A42},
580 {0x0A47, 0x0A48}, {0x0A4B, 0x0A4D}, {0x0A70, 0x0A71},
581 {0x0A81, 0x0A82}, {0x0ABC, 0x0ABC}, {0x0AC1, 0x0AC5},
582 {0x0AC7, 0x0AC8}, {0x0ACD, 0x0ACD}, {0x0B01, 0x0B01},
583 {0x0B3C, 0x0B3C}, {0x0B3F, 0x0B3F}, {0x0B41, 0x0B43},
584 {0x0B4D, 0x0B4D}, {0x0B56, 0x0B56}, {0x0B82, 0x0B82},
585 {0x0BC0, 0x0BC0}, {0x0BCD, 0x0BCD}, {0x0C3E, 0x0C40},
586 {0x0C46, 0x0C48}, {0x0C4A, 0x0C4D}, {0x0C55, 0x0C56},
587 {0x0CBF, 0x0CBF}, {0x0CC6, 0x0CC6}, {0x0CCC, 0x0CCD},
588 {0x0D41, 0x0D43}, {0x0D4D, 0x0D4D}, {0x0DCA, 0x0DCA},
589 {0x0DD2, 0x0DD4}, {0x0DD6, 0x0DD6}, {0x0E31, 0x0E31},
590 {0x0E34, 0x0E3A}, {0x0E47, 0x0E4E}, {0x0EB1, 0x0EB1},
591 {0x0EB4, 0x0EB9}, {0x0EBB, 0x0EBC}, {0x0EC8, 0x0ECD},
592 {0x0F18, 0x0F19}, {0x0F35, 0x0F35}, {0x0F37, 0x0F37},
593 {0x0F39, 0x0F39}, {0x0F71, 0x0F7E}, {0x0F80, 0x0F84},
594 {0x0F86, 0x0F87}, {0x0F90, 0x0F97}, {0x0F99, 0x0FBC},
595 {0x0FC6, 0x0FC6}, {0x102D, 0x1030}, {0x1032, 0x1032},
596 {0x1036, 0x1037}, {0x1039, 0x1039}, {0x1058, 0x1059},
597 {0x1160, 0x11FF}, {0x17B7, 0x17BD}, {0x17C6, 0x17C6},
598 {0x17C9, 0x17D3}, {0x180B, 0x180E}, {0x18A9, 0x18A9},
599 {0x200B, 0x200F}, {0x202A, 0x202E}, {0x206A, 0x206F},
600 {0x20D0, 0x20E3}, {0x302A, 0x302F}, {0x3099, 0x309A},
601 {0xFB1E, 0xFB1E}, {0xFE20, 0xFE23}, {0xFEFF, 0xFEFF},
605 /* test for 8-bit control characters */
609 if (ucs < 0x20 || (ucs >= 0x7f && ucs < 0xa0) || ucs > 0x0010ffff)
612 /* binary search in table of non-spacing characters */
613 if (mbbisearch(ucs, combining,
614 sizeof(combining) / sizeof(struct mbinterval) - 1))
618 * if we arrive here, ucs is not a combining or C0/C1 control character
623 (ucs <= 0x115f || /* Hangul Jamo init. consonants */
624 (ucs >= 0x2e80 && ucs <= 0xa4cf && (ucs & ~0x0011) != 0x300a &&
625 ucs != 0x303f) || /* CJK ... Yi */
626 (ucs >= 0xac00 && ucs <= 0xd7a3) || /* Hangul Syllables */
627 (ucs >= 0xf900 && ucs <= 0xfaff) || /* CJK Compatibility
629 (ucs >= 0xfe30 && ucs <= 0xfe6f) || /* CJK Compatibility Forms */
630 (ucs >= 0xff00 && ucs <= 0xff5f) || /* Fullwidth Forms */
631 (ucs >= 0xffe0 && ucs <= 0xffe6) ||
632 (ucs >= 0x20000 && ucs <= 0x2ffff)));
636 utf2ucs(const unsigned char *c)
639 * one char version of pg_utf2wchar_with_len. no control here, c must
640 * point to a large enough string
642 if ((*c & 0x80) == 0)
643 return (pg_wchar) c[0];
644 else if ((*c & 0xe0) == 0xc0)
645 return (pg_wchar) (((c[0] & 0x1f) << 6) |
647 else if ((*c & 0xf0) == 0xe0)
648 return (pg_wchar) (((c[0] & 0x0f) << 12) |
649 ((c[1] & 0x3f) << 6) |
651 else if ((*c & 0xf8) == 0xf0)
652 return (pg_wchar) (((c[0] & 0x07) << 18) |
653 ((c[1] & 0x3f) << 12) |
654 ((c[2] & 0x3f) << 6) |
657 /* that is an invalid code on purpose */
662 pg_utf_dsplen(const unsigned char *s)
664 return ucs_wcwidth(utf2ucs(s));
668 * convert mule internal code to pg_wchar
669 * caller should allocate enough space for "to"
670 * len: length of from.
671 * "from" not necessarily null terminated.
674 pg_mule2wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
678 while (len > 0 && *from)
680 if (IS_LC1(*from) && len >= 2)
686 else if (IS_LCPRV1(*from) && len >= 3)
693 else if (IS_LC2(*from) && len >= 3)
700 else if (IS_LCPRV2(*from) && len >= 4)
710 *to = (unsigned char) *from++;
721 pg_mule_mblen(const unsigned char *s)
727 else if (IS_LCPRV1(*s))
731 else if (IS_LCPRV2(*s))
734 len = 1; /* assume ASCII */
739 pg_mule_dsplen(const unsigned char *s)
745 else if (IS_LCPRV1(*s))
749 else if (IS_LCPRV2(*s))
752 len = 1; /* assume ASCII */
761 pg_latin12wchar_with_len(const unsigned char *from, pg_wchar *to, int len)
765 while (len > 0 && *from)
776 pg_latin1_mblen(const unsigned char *s)
782 pg_latin1_dsplen(const unsigned char *s)
784 return pg_ascii_dsplen(s);
791 pg_sjis_mblen(const unsigned char *s)
795 if (*s >= 0xa1 && *s <= 0xdf)
796 len = 1; /* 1 byte kana? */
797 else if (IS_HIGHBIT_SET(*s))
798 len = 2; /* kanji? */
800 len = 1; /* should be ASCII */
805 pg_sjis_dsplen(const unsigned char *s)
809 if (*s >= 0xa1 && *s <= 0xdf)
810 len = 1; /* 1 byte kana? */
811 else if (IS_HIGHBIT_SET(*s))
812 len = 2; /* kanji? */
814 len = pg_ascii_dsplen(s); /* should be ASCII */
822 pg_big5_mblen(const unsigned char *s)
826 if (IS_HIGHBIT_SET(*s))
827 len = 2; /* kanji? */
829 len = 1; /* should be ASCII */
834 pg_big5_dsplen(const unsigned char *s)
838 if (IS_HIGHBIT_SET(*s))
839 len = 2; /* kanji? */
841 len = pg_ascii_dsplen(s); /* should be ASCII */
849 pg_gbk_mblen(const unsigned char *s)
853 if (IS_HIGHBIT_SET(*s))
854 len = 2; /* kanji? */
856 len = 1; /* should be ASCII */
861 pg_gbk_dsplen(const unsigned char *s)
865 if (IS_HIGHBIT_SET(*s))
866 len = 2; /* kanji? */
868 len = pg_ascii_dsplen(s); /* should be ASCII */
876 pg_uhc_mblen(const unsigned char *s)
880 if (IS_HIGHBIT_SET(*s))
881 len = 2; /* 2byte? */
883 len = 1; /* should be ASCII */
888 pg_uhc_dsplen(const unsigned char *s)
892 if (IS_HIGHBIT_SET(*s))
893 len = 2; /* 2byte? */
895 len = pg_ascii_dsplen(s); /* should be ASCII */
901 * * Added by Bill Huang <bhuang@redhat.com>,<bill_huanghb@ybb.ne.jp>
904 pg_gb18030_mblen(const unsigned char *s)
908 if (!IS_HIGHBIT_SET(*s))
912 if ((*(s + 1) >= 0x40 && *(s + 1) <= 0x7e) || (*(s + 1) >= 0x80 && *(s + 1) <= 0xfe))
914 else if (*(s + 1) >= 0x30 && *(s + 1) <= 0x39)
923 pg_gb18030_dsplen(const unsigned char *s)
927 if (IS_HIGHBIT_SET(*s))
930 len = pg_ascii_dsplen(s); /* ASCII */
935 *-------------------------------------------------------------------
936 * multibyte sequence validators
938 * These functions accept "s", a pointer to the first byte of a string,
939 * and "len", the remaining length of the string. If there is a validly
940 * encoded character beginning at *s, return its length in bytes; else
943 * The functions can assume that len > 0 and that *s != '\0', but they must
944 * test for and reject zeroes in any additional bytes of a multibyte character.
946 * Note that this definition allows the function for a single-byte
947 * encoding to be just "return 1".
948 *-------------------------------------------------------------------
952 pg_ascii_verifier(const unsigned char *s, int len)
957 #define IS_EUC_RANGE_VALID(c) ((c) >= 0xa1 && (c) <= 0xfe)
960 pg_eucjp_verifier(const unsigned char *s, int len)
970 case SS2: /* JIS X 0201 */
975 if (c2 < 0xa1 || c2 > 0xdf)
979 case SS3: /* JIS X 0212 */
984 if (!IS_EUC_RANGE_VALID(c2))
987 if (!IS_EUC_RANGE_VALID(c2))
992 if (IS_HIGHBIT_SET(c1)) /* JIS X 0208? */
997 if (!IS_EUC_RANGE_VALID(c1))
1000 if (!IS_EUC_RANGE_VALID(c2))
1015 pg_euckr_verifier(const unsigned char *s, int len)
1023 if (IS_HIGHBIT_SET(c1))
1028 if (!IS_EUC_RANGE_VALID(c1))
1031 if (!IS_EUC_RANGE_VALID(c2))
1043 /* EUC-CN byte sequences are exactly same as EUC-KR */
1044 #define pg_euccn_verifier pg_euckr_verifier
1047 pg_euctw_verifier(const unsigned char *s, int len)
1057 case SS2: /* CNS 11643 Plane 1-7 */
1062 if (c2 < 0xa1 || c2 > 0xa7)
1065 if (!IS_EUC_RANGE_VALID(c2))
1068 if (!IS_EUC_RANGE_VALID(c2))
1072 case SS3: /* unused */
1076 if (IS_HIGHBIT_SET(c1)) /* CNS 11643 Plane 1 */
1081 /* no further range check on c1? */
1083 if (!IS_EUC_RANGE_VALID(c2))
1097 pg_johab_verifier(const unsigned char *s, int len)
1103 l = mbl = pg_johab_mblen(s);
1108 if (!IS_HIGHBIT_SET(*s))
1114 if (!IS_EUC_RANGE_VALID(c))
1121 pg_mule_verifier(const unsigned char *s, int len)
1127 l = mbl = pg_mule_mblen(s);
1135 if (!IS_HIGHBIT_SET(c))
1142 pg_latin1_verifier(const unsigned char *s, int len)
1148 pg_sjis_verifier(const unsigned char *s, int len)
1155 l = mbl = pg_sjis_mblen(s);
1160 if (l == 1) /* pg_sjis_mblen already verified it */
1165 if (!ISSJISHEAD(c1) || !ISSJISTAIL(c2))
1171 pg_big5_verifier(const unsigned char *s, int len)
1176 l = mbl = pg_big5_mblen(s);
1191 pg_gbk_verifier(const unsigned char *s, int len)
1196 l = mbl = pg_gbk_mblen(s);
1211 pg_uhc_verifier(const unsigned char *s, int len)
1216 l = mbl = pg_uhc_mblen(s);
1231 pg_gb18030_verifier(const unsigned char *s, int len)
1236 l = mbl = pg_gb18030_mblen(s);
1251 pg_utf8_verifier(const unsigned char *s, int len)
1253 int l = pg_utf_mblen(s);
1258 if (!pg_utf8_islegal(s, l))
1265 * Check for validity of a single UTF-8 encoded character
1267 * This directly implements the rules in RFC3629. The bizarre-looking
1268 * restrictions on the second byte are meant to ensure that there isn't
1269 * more than one encoding of a given Unicode character point; that is,
1270 * you may not use a longer-than-necessary byte sequence with high order
1271 * zero bits to represent a character that would fit in fewer bytes.
1272 * To do otherwise is to create security hazards (eg, create an apparent
1273 * non-ASCII character that decodes to plain ASCII).
1275 * length is assumed to have been obtained by pg_utf_mblen(), and the
1276 * caller must have checked that that many bytes are present in the buffer.
1279 pg_utf8_islegal(const unsigned char *source, int length)
1286 /* reject lengths 5 and 6 for now */
1290 if (a < 0x80 || a > 0xBF)
1295 if (a < 0x80 || a > 0xBF)
1303 if (a < 0xA0 || a > 0xBF)
1307 if (a < 0x80 || a > 0x9F)
1311 if (a < 0x90 || a > 0xBF)
1315 if (a < 0x80 || a > 0x8F)
1319 if (a < 0x80 || a > 0xBF)
1326 if (a >= 0x80 && a < 0xC2)
1336 *-------------------------------------------------------------------
1337 * encoding info table
1338 * XXX must be sorted by the same order as enum pg_enc (in mb/pg_wchar.h)
1339 *-------------------------------------------------------------------
1341 pg_wchar_tbl pg_wchar_table[] = {
1342 {pg_ascii2wchar_with_len, pg_ascii_mblen, pg_ascii_dsplen, pg_ascii_verifier, 1}, /* PG_SQL_ASCII */
1343 {pg_eucjp2wchar_with_len, pg_eucjp_mblen, pg_eucjp_dsplen, pg_eucjp_verifier, 3}, /* PG_EUC_JP */
1344 {pg_euccn2wchar_with_len, pg_euccn_mblen, pg_euccn_dsplen, pg_euccn_verifier, 2}, /* PG_EUC_CN */
1345 {pg_euckr2wchar_with_len, pg_euckr_mblen, pg_euckr_dsplen, pg_euckr_verifier, 3}, /* PG_EUC_KR */
1346 {pg_euctw2wchar_with_len, pg_euctw_mblen, pg_euctw_dsplen, pg_euctw_verifier, 4}, /* PG_EUC_TW */
1347 {pg_eucjp2wchar_with_len, pg_eucjp_mblen, pg_eucjp_dsplen, pg_eucjp_verifier, 3}, /* PG_EUC_JIS_2004 */
1348 {pg_utf2wchar_with_len, pg_utf_mblen, pg_utf_dsplen, pg_utf8_verifier, 4}, /* PG_UTF8 */
1349 {pg_mule2wchar_with_len, pg_mule_mblen, pg_mule_dsplen, pg_mule_verifier, 4}, /* PG_MULE_INTERNAL */
1350 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN1 */
1351 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN2 */
1352 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN3 */
1353 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN4 */
1354 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN5 */
1355 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN6 */
1356 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN7 */
1357 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN8 */
1358 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN9 */
1359 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_LATIN10 */
1360 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1256 */
1361 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1258 */
1362 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN866 */
1363 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN874 */
1364 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_KOI8R */
1365 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1251 */
1366 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1252 */
1367 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* ISO-8859-5 */
1368 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* ISO-8859-6 */
1369 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* ISO-8859-7 */
1370 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* ISO-8859-8 */
1371 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1250 */
1372 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1253 */
1373 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1254 */
1374 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1255 */
1375 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_WIN1257 */
1376 {pg_latin12wchar_with_len, pg_latin1_mblen, pg_latin1_dsplen, pg_latin1_verifier, 1}, /* PG_KOI8U */
1377 {0, pg_sjis_mblen, pg_sjis_dsplen, pg_sjis_verifier, 2}, /* PG_SJIS */
1378 {0, pg_big5_mblen, pg_big5_dsplen, pg_big5_verifier, 2}, /* PG_BIG5 */
1379 {0, pg_gbk_mblen, pg_gbk_dsplen, pg_gbk_verifier, 2}, /* PG_GBK */
1380 {0, pg_uhc_mblen, pg_uhc_dsplen, pg_uhc_verifier, 2}, /* PG_UHC */
1381 {0, pg_gb18030_mblen, pg_gb18030_dsplen, pg_gb18030_verifier, 4}, /* PG_GB18030 */
1382 {0, pg_johab_mblen, pg_johab_dsplen, pg_johab_verifier, 3}, /* PG_JOHAB */
1383 {0, pg_sjis_mblen, pg_sjis_dsplen, pg_sjis_verifier, 2} /* PG_SHIFT_JIS_2004 */
1386 /* returns the byte length of a word for mule internal code */
1388 pg_mic_mblen(const unsigned char *mbstr)
1390 return pg_mule_mblen(mbstr);
1394 * Returns the byte length of a multibyte character.
1397 pg_encoding_mblen(int encoding, const char *mbstr)
1399 Assert(PG_VALID_ENCODING(encoding));
1401 return ((encoding >= 0 &&
1402 encoding < sizeof(pg_wchar_table) / sizeof(pg_wchar_tbl)) ?
1403 ((*pg_wchar_table[encoding].mblen) ((const unsigned char *) mbstr)) :
1404 ((*pg_wchar_table[PG_SQL_ASCII].mblen) ((const unsigned char *) mbstr)));
1408 * Returns the display length of a multibyte character.
1411 pg_encoding_dsplen(int encoding, const char *mbstr)
1413 Assert(PG_VALID_ENCODING(encoding));
1415 return ((encoding >= 0 &&
1416 encoding < sizeof(pg_wchar_table) / sizeof(pg_wchar_tbl)) ?
1417 ((*pg_wchar_table[encoding].dsplen) ((const unsigned char *) mbstr)) :
1418 ((*pg_wchar_table[PG_SQL_ASCII].dsplen) ((const unsigned char *) mbstr)));
1422 * Verify the first multibyte character of the given string.
1423 * Return its byte length if good, -1 if bad. (See comments above for
1424 * full details of the mbverify API.)
1427 pg_encoding_verifymb(int encoding, const char *mbstr, int len)
1429 Assert(PG_VALID_ENCODING(encoding));
1431 return ((encoding >= 0 &&
1432 encoding < sizeof(pg_wchar_table) / sizeof(pg_wchar_tbl)) ?
1433 ((*pg_wchar_table[encoding].mbverify) ((const unsigned char *) mbstr, len)) :
1434 ((*pg_wchar_table[PG_SQL_ASCII].mbverify) ((const unsigned char *) mbstr, len)));
1438 * fetch maximum length of a given encoding
1441 pg_encoding_max_length(int encoding)
1443 Assert(PG_VALID_ENCODING(encoding));
1445 return pg_wchar_table[encoding].maxmblen;
1451 * fetch maximum length of the encoding for the current database
1454 pg_database_encoding_max_length(void)
1456 return pg_wchar_table[GetDatabaseEncoding()].maxmblen;
1460 * Verify mbstr to make sure that it is validly encoded in the current
1461 * database encoding. Otherwise same as pg_verify_mbstr().
1464 pg_verifymbstr(const char *mbstr, int len, bool noError)
1467 pg_verify_mbstr_len(GetDatabaseEncoding(), mbstr, len, noError) >= 0;
1471 * Verify mbstr to make sure that it is validly encoded in the specified
1475 pg_verify_mbstr(int encoding, const char *mbstr, int len, bool noError)
1477 return pg_verify_mbstr_len(encoding, mbstr, len, noError) >= 0;
1481 * Verify mbstr to make sure that it is validly encoded in the specified
1484 * mbstr is not necessarily zero terminated; length of mbstr is
1487 * If OK, return length of string in the encoding.
1488 * If a problem is found, return -1 when noError is
1489 * true; when noError is false, ereport() a descriptive message.
1492 pg_verify_mbstr_len(int encoding, const char *mbstr, int len, bool noError)
1494 mbverifier mbverify;
1497 Assert(PG_VALID_ENCODING(encoding));
1500 * In single-byte encodings, we need only reject nulls (\0).
1502 if (pg_encoding_max_length(encoding) <= 1)
1504 const char *nullpos = memchr(mbstr, 0, len);
1506 if (nullpos == NULL)
1510 report_invalid_encoding(encoding, nullpos, 1);
1513 /* fetch function pointer just once */
1514 mbverify = pg_wchar_table[encoding].mbverify;
1522 /* fast path for ASCII-subset characters */
1523 if (!IS_HIGHBIT_SET(*mbstr))
1534 report_invalid_encoding(encoding, mbstr, len);
1537 l = (*mbverify) ((const unsigned char *) mbstr, len);
1543 report_invalid_encoding(encoding, mbstr, len);
1554 * check_encoding_conversion_args: check arguments of a conversion function
1556 * "expected" arguments can be either an encoding ID or -1 to indicate that
1557 * the caller will check whether it accepts the ID.
1559 * Note: the errors here are not really user-facing, so elog instead of
1560 * ereport seems sufficient. Also, we trust that the "expected" encoding
1561 * arguments are valid encoding IDs, but we don't trust the actuals.
1564 check_encoding_conversion_args(int src_encoding,
1567 int expected_src_encoding,
1568 int expected_dest_encoding)
1570 if (!PG_VALID_ENCODING(src_encoding))
1571 elog(ERROR, "invalid source encoding ID: %d", src_encoding);
1572 if (src_encoding != expected_src_encoding && expected_src_encoding >= 0)
1573 elog(ERROR, "expected source encoding \"%s\", but got \"%s\"",
1574 pg_enc2name_tbl[expected_src_encoding].name,
1575 pg_enc2name_tbl[src_encoding].name);
1576 if (!PG_VALID_ENCODING(dest_encoding))
1577 elog(ERROR, "invalid destination encoding ID: %d", dest_encoding);
1578 if (dest_encoding != expected_dest_encoding && expected_dest_encoding >= 0)
1579 elog(ERROR, "expected destination encoding \"%s\", but got \"%s\"",
1580 pg_enc2name_tbl[expected_dest_encoding].name,
1581 pg_enc2name_tbl[dest_encoding].name);
1583 elog(ERROR, "encoding conversion length must not be negative");
1587 * report_invalid_encoding: complain about invalid multibyte character
1589 * note: len is remaining length of string, not length of character;
1590 * len must be greater than zero, as we always examine the first byte.
1593 report_invalid_encoding(int encoding, const char *mbstr, int len)
1595 int l = pg_encoding_mblen(encoding, mbstr);
1596 char buf[8 * 2 + 1];
1601 jlimit = Min(l, len);
1602 jlimit = Min(jlimit, 8); /* prevent buffer overrun */
1604 for (j = 0; j < jlimit; j++)
1605 p += sprintf(p, "%02x", (unsigned char) mbstr[j]);
1608 (errcode(ERRCODE_CHARACTER_NOT_IN_REPERTOIRE),
1609 errmsg("invalid byte sequence for encoding \"%s\": 0x%s",
1610 pg_enc2name_tbl[encoding].name,
1615 * report_untranslatable_char: complain about untranslatable character
1617 * note: len is remaining length of string, not length of character;
1618 * len must be greater than zero, as we always examine the first byte.
1621 report_untranslatable_char(int src_encoding, int dest_encoding,
1622 const char *mbstr, int len)
1624 int l = pg_encoding_mblen(src_encoding, mbstr);
1625 char buf[8 * 2 + 1];
1630 jlimit = Min(l, len);
1631 jlimit = Min(jlimit, 8); /* prevent buffer overrun */
1633 for (j = 0; j < jlimit; j++)
1634 p += sprintf(p, "%02x", (unsigned char) mbstr[j]);
1637 (errcode(ERRCODE_UNTRANSLATABLE_CHARACTER),
1638 errmsg("character 0x%s of encoding \"%s\" has no equivalent in \"%s\"",
1640 pg_enc2name_tbl[src_encoding].name,
1641 pg_enc2name_tbl[dest_encoding].name)));