cp1251.h revision 9c019ec5
1
2/*
3 * CP1251
4 */
5
6static const unsigned short cp1251_2uni[128] = {
7  /* 0x80 */
8  0x0402, 0x0403, 0x201a, 0x0453, 0x201e, 0x2026, 0x2020, 0x2021,
9  0x20ac, 0x2030, 0x0409, 0x2039, 0x040a, 0x040c, 0x040b, 0x040f,
10  /* 0x90 */
11  0x0452, 0x2018, 0x2019, 0x201c, 0x201d, 0x2022, 0x2013, 0x2014,
12  0xfffd, 0x2122, 0x0459, 0x203a, 0x045a, 0x045c, 0x045b, 0x045f,
13  /* 0xa0 */
14  0x00a0, 0x040e, 0x045e, 0x0408, 0x00a4, 0x0490, 0x00a6, 0x00a7,
15  0x0401, 0x00a9, 0x0404, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x0407,
16  /* 0xb0 */
17  0x00b0, 0x00b1, 0x0406, 0x0456, 0x0491, 0x00b5, 0x00b6, 0x00b7,
18  0x0451, 0x2116, 0x0454, 0x00bb, 0x0458, 0x0405, 0x0455, 0x0457,
19  /* 0xc0 */
20  0x0410, 0x0411, 0x0412, 0x0413, 0x0414, 0x0415, 0x0416, 0x0417,
21  0x0418, 0x0419, 0x041a, 0x041b, 0x041c, 0x041d, 0x041e, 0x041f,
22  /* 0xd0 */
23  0x0420, 0x0421, 0x0422, 0x0423, 0x0424, 0x0425, 0x0426, 0x0427,
24  0x0428, 0x0429, 0x042a, 0x042b, 0x042c, 0x042d, 0x042e, 0x042f,
25  /* 0xe0 */
26  0x0430, 0x0431, 0x0432, 0x0433, 0x0434, 0x0435, 0x0436, 0x0437,
27  0x0438, 0x0439, 0x043a, 0x043b, 0x043c, 0x043d, 0x043e, 0x043f,
28  /* 0xf0 */
29  0x0440, 0x0441, 0x0442, 0x0443, 0x0444, 0x0445, 0x0446, 0x0447,
30  0x0448, 0x0449, 0x044a, 0x044b, 0x044c, 0x044d, 0x044e, 0x044f,
31};
32
33static int
34cp1251_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
35{
36  unsigned char c = *s;
37  if (c < 0x80) {
38    *pwc = (ucs4_t) c;
39    return 1;
40  }
41  else {
42    unsigned short wc = cp1251_2uni[c-0x80];
43    if (wc != 0xfffd) {
44      *pwc = (ucs4_t) wc;
45      return 1;
46    }
47  }
48  return RET_ILSEQ;
49}
50
51static const unsigned char cp1251_page00[32] = {
52  0xa0, 0x00, 0x00, 0x00, 0xa4, 0x00, 0xa6, 0xa7, /* 0xa0-0xa7 */
53  0x00, 0xa9, 0x00, 0xab, 0xac, 0xad, 0xae, 0x00, /* 0xa8-0xaf */
54  0xb0, 0xb1, 0x00, 0x00, 0x00, 0xb5, 0xb6, 0xb7, /* 0xb0-0xb7 */
55  0x00, 0x00, 0x00, 0xbb, 0x00, 0x00, 0x00, 0x00, /* 0xb8-0xbf */
56};
57static const unsigned char cp1251_page04[152] = {
58  0x00, 0xa8, 0x80, 0x81, 0xaa, 0xbd, 0xb2, 0xaf, /* 0x00-0x07 */
59  0xa3, 0x8a, 0x8c, 0x8e, 0x8d, 0x00, 0xa1, 0x8f, /* 0x08-0x0f */
60  0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7, /* 0x10-0x17 */
61  0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, /* 0x18-0x1f */
62  0xd0, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, /* 0x20-0x27 */
63  0xd8, 0xd9, 0xda, 0xdb, 0xdc, 0xdd, 0xde, 0xdf, /* 0x28-0x2f */
64  0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, /* 0x30-0x37 */
65  0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, /* 0x38-0x3f */
66  0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, /* 0x40-0x47 */
67  0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff, /* 0x48-0x4f */
68  0x00, 0xb8, 0x90, 0x83, 0xba, 0xbe, 0xb3, 0xbf, /* 0x50-0x57 */
69  0xbc, 0x9a, 0x9c, 0x9e, 0x9d, 0x00, 0xa2, 0x9f, /* 0x58-0x5f */
70  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x60-0x67 */
71  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x68-0x6f */
72  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x70-0x77 */
73  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x78-0x7f */
74  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x80-0x87 */
75  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x88-0x8f */
76  0xa5, 0xb4, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x90-0x97 */
77};
78static const unsigned char cp1251_page20[48] = {
79  0x00, 0x00, 0x00, 0x96, 0x97, 0x00, 0x00, 0x00, /* 0x10-0x17 */
80  0x91, 0x92, 0x82, 0x00, 0x93, 0x94, 0x84, 0x00, /* 0x18-0x1f */
81  0x86, 0x87, 0x95, 0x00, 0x00, 0x00, 0x85, 0x00, /* 0x20-0x27 */
82  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x28-0x2f */
83  0x89, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x30-0x37 */
84  0x00, 0x8b, 0x9b, 0x00, 0x00, 0x00, 0x00, 0x00, /* 0x38-0x3f */
85};
86
87static int
88cp1251_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
89{
90  unsigned char c = 0;
91  if (wc < 0x0080) {
92    *r = (unsigned char) wc;
93    return 1;
94  }
95  else if (wc >= 0x00a0 && wc < 0x00c0)
96    c = cp1251_page00[wc-0x00a0];
97  else if (wc >= 0x0400 && wc < 0x0498)
98    c = cp1251_page04[wc-0x0400];
99  else if (wc >= 0x2010 && wc < 0x2040)
100    c = cp1251_page20[wc-0x2010];
101  else if (wc == 0x20ac)
102    c = 0x88;
103  else if (wc == 0x2116)
104    c = 0xb9;
105  else if (wc == 0x2122)
106    c = 0x99;
107  if (c != 0) {
108    *r = c;
109    return 1;
110  }
111  return RET_ILSEQ;
112}
113