CP1258 handles combining characters.
[libiconv.git] / lib / isoir165.h
blob031385497f407e2307a85770b8c248703f681764
1 /*
2 * Copyright (C) 1999-2000 Free Software Foundation, Inc.
3 * This file is part of the GNU LIBICONV Library.
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.
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.
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.
17 * If not, write to the Free Software Foundation, Inc., 59 Temple Place -
18 * Suite 330, Boston, MA 02111-1307, USA.
22 * ISO-IR-165
26 * ISO-IR-165 is an extension of GB 2312, consisting of:
27 * 1. GB 6345.1-86 corrections:
28 * Two corrections to GB 2312, at 0x2367 and 0x6F71.
29 * 2. GB 6345.1-86 additions:
30 * - 6 new full-width pinyin characters in row 0x28.
31 * - ISO646-CN in row 0x2A.
32 * - 32 half-width pinyin characters in row 0x2B.
33 * 3. GB 8565.2-88 additions:
34 * - 50 characters in row 0x2D.
35 * - 92 characters in row 0x2E.
36 * - 93 characters in row 0x2F.
37 * - 470 characters in rows 0x7A-0x7E.
38 * 4. ISO-IR-165 additions:
39 * - 22 characters in row 0x26.
40 * - 94 characters in row 0x2C.
41 * - 44 new characters in row 0x2D.
42 * - 1 new character in row 0x2F.
44 * The conversion table was created from the following sources:
45 * Ad 1. The 0x2367 correction is already integrated in the unicode.org
46 * GB2312.TXT table. The 0x6F71 mapping is the same in the unicode.org
47 * GB2312.TXT and UNIHAN.TXT table and in Koichi Yasuoka's Uni2GB table,
48 * so we assume it's correct.
49 * The unicode.org UNIHAN.TXT table about GB 8565 is not usable: it has
50 * extraneous code points at rows 0x28, 0x2C, 0x2D. Note also that it does
51 * not list the 69 non-hanzi in row 0x2F. Moreover, it has the characters
52 * 0x2F7A-0x2F7D shifted down by one to 0x2F79-0x2F7C.
53 * Therefore we take the GB8565 and ISO-IR-165 data from Koichi Yasuoka's
54 * Uni2GB table.
55 * Ad 1. Yasuoka maps 0x2367 to U+0261 (small script g) and 0x2840 to U+FF47
56 * (full-width small normal g). While coherent with ISO-IR's 165.pdf,
57 * this disagrees with Ken Lunde's book: He says that ISO-IR-165
58 * includes the GB6345 correction, i.e. maps 0x2367 to U+FF47 or U+0067
59 * and _not_ to U+0261 (small script g).
60 * To overcome the confusion, we just map both 0x2367 and 0x2840 to
61 * U+FF47.
62 * Ad 2. Row 0x28: Add a mapping from 0x283F to U+01F9.
63 * Row 0x2A: Mapping is well-known, also present in Koichi Yasuoka's
64 * table.
65 * Row 0x2B: Typed in by hand from appendix E in Ken Lunde's book.
66 * When converting from Unicode to ISO-IR-165, prefer the half-width
67 * range 0x2B{21..40} to the full-width range 0x28{21..40}.
68 * Ad 3. Rows 0x2D, 0x2E: Both Koichi Yasuoka's Uni2GB table and the UNIHAN.TXT
69 * data for GB 8565 agree here.
70 * Row 0x2F: Taken from Koichi Yasuoka's Uni2GB table.
71 * Rows 0x7A-0x7E: Koichi Yasuoka's Uni2GB table and the UNIHAN.TXT
72 * data for GB 8565 agree here mostly. Differences:
73 * 0x7C38 -> U+6F26 or U+527A ? We choose U+6F26.
74 * 0x7C5A -> U+7A40 or U+6996 ? We choose U+6996.
75 * Ad 4. Row 0x26: Mapping unknown.
76 * Rows 0x2C, 0x2D: Both Koichi Yasuoka's Uni2GB table and the UNIHAN.TXT
77 * data for GB 8565 (!) agree here.
78 * Row 0x2F: Taken from Koichi Yasuoka's Uni2GB table.
81 #include "isoir165ext.h"
83 static int
84 isoir165_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
86 int ret;
88 /* Map full-width pinyin (row 0x28) like half-width pinyin (row 0x2B). */
89 if (s[0] == 0x28) {
90 if (n >= 2) {
91 unsigned char c2 = s[1];
92 if (c2 >= 0x21 && c2 <= 0x40) {
93 unsigned char buf[2];
94 buf[0] = 0x2b;
95 buf[1] = c2;
96 ret = isoir165ext_mbtowc(conv,pwc,buf,2);
97 if (ret != RET_ILSEQ)
98 return ret;
102 /* Try the GB2312 -> Unicode table. */
103 ret = gb2312_mbtowc(conv,pwc,s,n);
104 if (ret != RET_ILSEQ)
105 return ret;
106 /* Row 0x2A is GB_1988-80. */
107 if (s[0] == 0x2a) {
108 if (n >= 2) {
109 unsigned char c2 = s[1];
110 if (c2 >= 0x21 && c2 < 0x7f) {
111 int ret = iso646_cn_mbtowc(conv,pwc,s+1,1);
112 if (ret != 1) abort();
113 return 2;
115 return RET_ILSEQ;
117 return RET_TOOFEW(0);
119 /* Try the ISO-IR-165 extensions -> Unicode table. */
120 ret = isoir165ext_mbtowc(conv,pwc,s,n);
121 return ret;
124 static int
125 isoir165_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
127 unsigned char buf[1];
128 int ret;
130 /* Try the Unicode -> GB2312 table table. */
131 ret = gb2312_wctomb(conv,r,wc,n);
132 if (ret != RET_ILSEQ) {
133 if (ret != 2) abort();
134 if (!(r[0] == 0x28 && r[1] >= 0x21 && r[1] <= 0x40))
135 return ret;
137 /* Row 0x2A is GB_1988-80. */
138 ret = iso646_cn_wctomb(conv,buf,wc,1);
139 if (ret != RET_ILSEQ) {
140 if (ret != 1) abort();
141 if (buf[0] >= 0x21 && buf[0] < 0x7f) {
142 if (n >= 2) {
143 r[0] = 0x2a;
144 r[1] = buf[0];
145 return 2;
147 return RET_TOOSMALL;
150 /* Try the Unicode -> ISO-IR-165 extensions table. */
151 ret = isoir165ext_wctomb(conv,r,wc,n);
152 return ret;