Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 1 | /* |
| 2 | * encoding.c : implements the encoding conversion functions needed for XML |
| 3 | * |
| 4 | * Related specs: |
| 5 | * rfc2044 (UTF-8 and UTF-16) F. Yergeau Alis Technologies |
| 6 | * [ISO-10646] UTF-8 and UTF-16 in Annexes |
| 7 | * [ISO-8859-1] ISO Latin-1 characters codes. |
| 8 | * [UNICODE] The Unicode Consortium, "The Unicode Standard -- |
| 9 | * Worldwide Character Encoding -- Version 1.0", Addison- |
| 10 | * Wesley, Volume 1, 1991, Volume 2, 1992. UTF-8 is |
| 11 | * described in Unicode Technical Report #4. |
| 12 | * [US-ASCII] Coded Character Set--7-bit American Standard Code for |
| 13 | * Information Interchange, ANSI X3.4-1986. |
| 14 | * |
| 15 | * Original code from "Martin J. Duerst" <duerst@w3.org> |
| 16 | * |
| 17 | * See Copyright for the status of this software. |
| 18 | * |
| 19 | * $Id$ |
| 20 | * |
| 21 | * Daniel.Veillard@w3.org |
| 22 | */ |
| 23 | |
| 24 | #include "encoding.h" |
| 25 | |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 26 | /** |
| 27 | * isolat1ToUTF8: |
| 28 | * @out: a pointer ot an array of bytes to store the result |
| 29 | * @outlen: the lenght of @out |
| 30 | * @in: a pointer ot an array of ISO Latin 1 chars |
| 31 | * @inlen: the lenght of @in |
| 32 | * |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 33 | * Take a block of ISO Latin 1 chars in and try to convert it to an UTF-8 |
| 34 | * block of chars out. |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 35 | * return values: number of byte written, or -1 by lack of space. |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 36 | */ |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 37 | int |
| 38 | isolat1ToUTF8(unsigned char* out, int outlen, unsigned char* in, int inlen) |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 39 | { |
| 40 | unsigned char* outstart= out; |
| 41 | unsigned char* outend= out+outlen; |
| 42 | unsigned char* inend= in+inlen; |
| 43 | unsigned char c; |
| 44 | |
| 45 | while (in < inend) { |
| 46 | c= *in++; |
| 47 | if (c < 0x80) { |
| 48 | if (out >= outend) return -1; |
| 49 | *out++ = c; |
| 50 | } |
| 51 | else { |
| 52 | if (out >= outend) return -1; |
| 53 | *out++ = 0xC0 | (c >> 6); |
| 54 | if (out >= outend) return -1; |
| 55 | *out++ = 0x80 | (0x3F & c); |
| 56 | } |
| 57 | } |
| 58 | return out-outstart; |
| 59 | } |
| 60 | |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 61 | /** |
| 62 | * UTF8Toisolat1: |
| 63 | * @out: a pointer ot an array of bytes to store the result |
| 64 | * @outlen: the lenght of @out |
| 65 | * @in: a pointer ot an array of UTF-8 chars |
| 66 | * @inlen: the lenght of @in |
| 67 | * |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 68 | * Take a block of UTF-8 chars in and try to convert it to an ISO Latin 1 |
| 69 | * block of chars out. |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 70 | * TODO: need a fallback mechanism ... |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 71 | * return values: the number of byte written, or -1 by lack of space, or -2 |
| 72 | * if the transcoding failed. |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 73 | */ |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 74 | int |
| 75 | UTF8Toisolat1(unsigned char* out, int outlen, unsigned char* in, int inlen) |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 76 | { |
| 77 | unsigned char* outstart= out; |
| 78 | unsigned char* outend= out+outlen; |
| 79 | unsigned char* inend= in+inlen; |
| 80 | unsigned char c, d; |
| 81 | |
| 82 | while (in < inend) { |
| 83 | c= *in++; |
| 84 | if (c < 0x80) { |
| 85 | if (out >= outend) return -1; |
| 86 | *out++= c; |
| 87 | } |
| 88 | else if (((c & 0xFE) == 0xC2) && in<inend) { |
| 89 | if (out >= outend) return -1; |
| 90 | *out++= ((c & 0x03) << 6) | (*in++ & 0x3F); |
| 91 | } |
| 92 | else return -2; |
| 93 | } |
| 94 | return out-outstart; |
| 95 | } |
| 96 | |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 97 | /** |
| 98 | * UTF16ToUTF8: |
| 99 | * @out: a pointer ot an array of bytes to store the result |
| 100 | * @outlen: the lenght of @out |
| 101 | * @in: a pointer ot an array of UTF-16 chars (array of unsigned shorts) |
| 102 | * @inlen: the lenght of @in |
| 103 | * |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 104 | * Take a block of UTF-16 ushorts in and try to convert it to an UTF-8 |
| 105 | * block of chars out. |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 106 | * return values: number of byte written, or -1 by lack of space. |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 107 | */ |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 108 | int |
| 109 | UTF16ToUTF8(unsigned char* out, int outlen, unsigned short* in, int inlen) |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 110 | { |
| 111 | unsigned char* outstart= out; |
| 112 | unsigned char* outend= out+outlen; |
| 113 | unsigned short* inend= in+inlen; |
| 114 | unsigned int c, d; |
| 115 | int bits; |
| 116 | |
| 117 | while (in < inend) { |
| 118 | c= *in++; |
| 119 | if ((c & 0xFC00) == 0xD800) { /* surrogates */ |
| 120 | if ((in<inend) && (((d=*in++) & 0xFC00) == 0xDC00)) { |
| 121 | c &= 0x03FF; |
| 122 | c <<= 10; |
| 123 | c |= d & 0x03FF; |
| 124 | c += 0x10000; |
| 125 | } |
| 126 | else return -1; |
| 127 | } |
| 128 | |
| 129 | /* assertion: c is a single UTF-4 value */ |
| 130 | |
| 131 | if (out >= outend) return -1; |
| 132 | if (c < 0x80) { *out++= c; bits= -6; } |
| 133 | else if (c < 0x800) { *out++= (c >> 6) | 0xC0; bits= 0; } |
| 134 | else if (c < 0x10000) { *out++= (c >> 12) | 0xE0; bits= 6; } |
| 135 | else { *out++= (c >> 18) | 0xF0; bits= 12; } |
| 136 | |
| 137 | for ( ; bits < 0; bits-= 6) { |
| 138 | if (out >= outend) return -1; |
| 139 | *out++= (c >> bits) & 0x3F; |
| 140 | } |
| 141 | } |
| 142 | return out-outstart; |
| 143 | } |
| 144 | |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 145 | /** |
| 146 | * UTF8ToUTF16: |
| 147 | * @out: a pointer ot an array of shorts to store the result |
| 148 | * @outlen: the lenght of @out (number of shorts) |
| 149 | * @in: a pointer ot an array of UTF-8 chars |
| 150 | * @inlen: the lenght of @in |
| 151 | * |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 152 | * Take a block of UTF-8 chars in and try to convert it to an UTF-16 |
| 153 | * block of chars out. |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 154 | * TODO: need a fallback mechanism ... |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 155 | * return values: the number of byte written, or -1 by lack of space, or -2 |
| 156 | * if the transcoding failed. |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 157 | */ |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 158 | int |
| 159 | UTF8ToUTF16(unsigned short* out, int outlen, unsigned char* in, int inlen) |
Daniel Veillard | 891e404 | 1998-10-19 00:43:02 +0000 | [diff] [blame] | 160 | { |
| 161 | unsigned short* outstart= out; |
| 162 | unsigned short* outend= out+outlen; |
| 163 | unsigned char* inend= in+inlen; |
| 164 | unsigned int c, d, trailing; |
| 165 | |
| 166 | while (in < inend) { |
| 167 | d= *in++; |
| 168 | if (d < 0x80) { c= d; trailing= 0; } |
| 169 | else if (d < 0xC0) return -2; /* trailing byte in leading position */ |
| 170 | else if (d < 0xE0) { c= d & 0x1F; trailing= 1; } |
| 171 | else if (d < 0xF0) { c= d & 0x0F; trailing= 2; } |
| 172 | else if (d < 0xF8) { c= d & 0x07; trailing= 3; } |
| 173 | else return -2; /* no chance for this in UTF-16 */ |
| 174 | |
| 175 | for ( ; trailing; trailing--) { |
| 176 | if ((in >= inend) || (((d= *in++) & 0xC0) != 0x80)) return -1; |
| 177 | c <<= 6; |
| 178 | c |= d & 0x3F; |
| 179 | } |
| 180 | |
| 181 | /* assertion: c is a single UTF-4 value */ |
| 182 | if (c < 0x10000) { |
| 183 | if (out >= outend) return -1; |
| 184 | *out++ = c; |
| 185 | } |
| 186 | else if (c < 0x110000) { |
| 187 | if (out+1 >= outend) return -1; |
| 188 | c -= 0x10000; |
| 189 | *out++ = 0xD800 | (c >> 10); |
| 190 | *out++ = 0xDC00 | (c & 0x03FF); |
| 191 | } |
| 192 | else return -1; |
| 193 | } |
| 194 | return out-outstart; |
| 195 | } |
| 196 | |
Daniel Veillard | 97b5877 | 1998-10-20 06:14:16 +0000 | [diff] [blame^] | 197 | |