Return to cp936.h CVS log

Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / libiconv / lib

Annotation of embedaddon/libiconv/lib/cp936.h, revision 1.1.1.2

1.1       misho       1: /*
1.1.1.2 ! misho       2:  * Copyright (C) 2005, 2016 Free Software Foundation, Inc.
1.1       misho       3:  * This file is part of the GNU LIBICONV Library.
                      4:  *
                      5:  * The GNU LIBICONV Library is free software; you can redistribute it
                      6:  * and/or modify it under the terms of the GNU Library General Public
                      7:  * License as published by the Free Software Foundation; either version 2
                      8:  * of the License, or (at your option) any later version.
                      9:  *
                     10:  * The GNU LIBICONV Library is distributed in the hope that it will be
                     11:  * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
                     12:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
                     13:  * Library General Public License for more details.
                     14:  *
                     15:  * You should have received a copy of the GNU Library General Public
                     16:  * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
1.1.1.2 ! misho      17:  * If not, see <https://www.gnu.org/licenses/>.
1.1       misho      18:  */
                     19: 
                     20: /*
                     21:  * CP936
                     22:  */
                     23: 
                     24: /*
                     25:  * The IANA has CP936 as an alias of GBK. But GBK is an official Chinese
                     26:  * specification, whereas CP936 is de-facto maintained by Microsoft. And,
                     27:  * of course, Microsoft modified CP936 since 1999.
                     28:  *
                     29:  * The differences from GBK are:
                     30:  *
                     31:  * 1. A single character:
                     32:  *
                     33:  *    code   CP936.TXT
                     34:  *    0x80   0x20AC # EURO SIGN
                     35:  *
                     36:  * Some variants of CP936 (in JDK, Windows-2000, ICU) also add:
                     37:  *
                     38:  * 2. Private area mappings:
                     39:  *
                     40:  *              code                 Unicode
                     41:  *    0x{A1..A2}{40..7E,80..A0}  U+E4C6..U+E585
                     42:  *    0x{AA..AF,F8..FE}{A1..FE}  U+E000..U+E4C5
                     43:  *
                     44:  * We add them too because, although there are backward compatibility problems
                     45:  * when a character from a private area is moved to an official Unicode code
                     46:  * point, they are useful for some people in practice.
                     47:  */
                     48: 
                     49: static int
1.1.1.2 ! misho      50: cp936_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
1.1       misho      51: {
                     52:   /* Try GBK first. */
                     53:   {
                     54:     int ret = ces_gbk_mbtowc(conv,pwc,s,n);
                     55:     if (ret != RET_ILSEQ)
                     56:       return ret;
                     57:   }
                     58:   /* Then handle the additional mappings. */
                     59:   {
                     60:     unsigned char c = *s;
                     61:     if (c == 0x80) {
                     62:       *pwc = 0x20ac;
                     63:       return 1;
                     64:     }
                     65:     /* User-defined characters */
                     66:     if (c >= 0xa1 && c <= 0xa2) {
                     67:       if (n < 2)
                     68:         return RET_TOOFEW(0);
                     69:       {
                     70:         unsigned char c2 = s[1];
                     71:         if ((c2 >= 0x40 && c2 < 0x7f) || (c2 >= 0x80 && c2 < 0xa1)) {
                     72:           *pwc = 0xe4c6 + 96 * (c - 0xa1) + (c2 - (c2 >= 0x80 ? 0x41 : 0x40));
                     73:           return 2;
                     74:         }
                     75:       }
                     76:     } else if ((c >= 0xaa && c < 0xb0) || (c >= 0xf8 && c < 0xff)) {
                     77:       if (n < 2)
                     78:         return RET_TOOFEW(0);
                     79:       {
                     80:         unsigned char c2 = s[1];
                     81:         if (c2 >= 0xa1 && c2 < 0xff) {
                     82:           *pwc = 0xe000 + 94 * (c - (c >= 0xf8 ? 0xf2 : 0xaa)) + (c2 - 0xa1);
                     83:           return 2;
                     84:         }
                     85:       }
                     86:     }
                     87:   }
                     88:   return RET_ILSEQ;
                     89: }
                     90: 
                     91: static int
1.1.1.2 ! misho      92: cp936_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
1.1       misho      93: {
                     94:   /* Try GBK first. */
                     95:   {
                     96:     int ret = ces_gbk_wctomb(conv,r,wc,n);
                     97:     if (ret != RET_ILUNI)
                     98:       return ret;
                     99:   }
                    100:   /* Then handle the additional mappings. */
                    101:   if (wc >= 0xe000 && wc < 0xe586) {
                    102:     /* User-defined characters */
                    103:     if (n < 2)
                    104:       return RET_TOOFEW(0);
                    105:     if (wc < 0xe4c6) {
                    106:       unsigned int i = wc - 0xe000;
                    107:       unsigned int c1 = i / 94;
                    108:       unsigned int c2 = i % 94;
                    109:       r[0] = c1 + (c1 < 6 ? 0xaa : 0xf2);
                    110:       r[1] = c2 + 0xa1;
                    111:       return 2;
                    112:     } else {
                    113:       unsigned int i = wc - 0xe4c6;
                    114:       unsigned int c1 = i / 96;
                    115:       unsigned int c2 = i % 96;
                    116:       r[0] = c1 + 0xa1;
                    117:       r[1] = c2 + (c2 < 0x3f ? 0x40 : 0x41);
                    118:       return 2;
                    119:     }
                    120:   } else if (wc == 0x20ac) {
                    121:     r[0] = 0x80;
                    122:     return 1;
                    123:   }
                    124:   return RET_ILUNI;
                    125: }