8 typedef size_t (*decode_func)(input_t *input, utf32 *buffer, size_t buffer_size);
23 /* state for utf-8 decoder */
24 utf32 utf8_part_decoded_min_code;
25 utf32 utf8_part_decoded_char;
26 size_t utf8_part_decoded_rest_len;
29 static input_error_callback_func input_error;
31 void set_input_error_callback(input_error_callback_func new_func)
33 input_error = new_func;
36 static size_t read_block(input_t *input, unsigned char *const read_buf,
39 if (input->kind == INPUT_FILE) {
40 FILE *file = input->in.file;
41 size_t const s = fread(read_buf, 1, n, file);
43 /* on OS/X ferror appears to return true on eof as well when running
44 * the application in gdb... */
45 if (!feof(file) && ferror(file))
46 input_error(0, 0, "read from input failed");
51 assert(input->kind == INPUT_STRING);
52 size_t len = strlen(input->in.string);
55 memcpy(read_buf, input->in.string, len);
56 input->in.string += len;
61 static size_t decode_iso_8859_1(input_t *input, utf32 *buffer,
64 unsigned char read_buf[buffer_size];
65 size_t const s = read_block(input, read_buf, sizeof(read_buf));
67 unsigned char const *src = read_buf;
68 unsigned char const *end = read_buf + s;
76 static size_t decode_iso_8859_15(input_t *input, utf32 *buffer,
79 unsigned char read_buf[buffer_size];
80 size_t const s = read_block(input, read_buf, sizeof(read_buf));
82 unsigned char const *src = read_buf;
83 unsigned char const *end = read_buf + s;
88 case 0xA4: tc = 0x20AC; break; // €
89 case 0xA6: tc = 0x0160; break; // Š
90 case 0xA8: tc = 0x0161; break; // š
91 case 0xB4: tc = 0x017D; break; // Ž
92 case 0xB8: tc = 0x017E; break; // ž
93 case 0xBC: tc = 0x0152; break; // Œ
94 case 0xBD: tc = 0x0153; break; // œ
95 case 0xBE: tc = 0x0178; break; // Ÿ
103 static size_t decode_utf8(input_t *input, utf32 *buffer, size_t buffer_size)
105 unsigned char read_buf[buffer_size];
108 size_t const s = read_block(input, read_buf, sizeof(read_buf));
110 if (input->utf8_part_decoded_rest_len > 0)
111 input_error(0, 0, "incomplete input char at end of input");
115 unsigned char const *src = read_buf;
116 unsigned char const *end = read_buf + s;
121 if (input->utf8_part_decoded_rest_len != 0) {
122 min_code = input->utf8_part_decoded_min_code;
123 decoded = input->utf8_part_decoded_char;
124 size_t const rest_len = input->utf8_part_decoded_rest_len;
125 input->utf8_part_decoded_rest_len = 0;
127 case 4: goto realign;
128 case 3: goto three_more;
129 case 2: goto two_more;
130 default: goto one_more;
135 if ((*src & 0x80) == 0) {
137 } else if ((*src & 0xE0) == 0xC0) {
139 decoded = *src++ & 0x1F;
142 input->utf8_part_decoded_min_code = min_code;
143 input->utf8_part_decoded_char = decoded;
144 input->utf8_part_decoded_rest_len = 1;
147 if ((*src & 0xC0) == 0x80) {
148 decoded = (decoded << 6) | (*src++ & 0x3F);
152 if (decoded < min_code ||
153 decoded > 0x10FFFF ||
154 (0xD800 <= decoded && decoded < 0xE000) || // high/low surrogates
155 (0xFDD0 <= decoded && decoded < 0xFDF0) || // noncharacters
156 (decoded & 0xFFFE) == 0xFFFE) { // noncharacters
157 input_error(0, 0, "invalid byte sequence in input");
159 } else if ((*src & 0xF0) == 0xE0) {
161 decoded = *src++ & 0x0F;
164 input->utf8_part_decoded_min_code = min_code;
165 input->utf8_part_decoded_char = decoded;
166 input->utf8_part_decoded_rest_len = 2;
169 if ((*src & 0xC0) == 0x80) {
170 decoded = (decoded << 6) | (*src++ & 0x3F);
175 } else if ((*src & 0xF8) == 0xF0) {
177 decoded = *src++ & 0x07;
180 input->utf8_part_decoded_min_code = min_code;
181 input->utf8_part_decoded_char = decoded;
182 input->utf8_part_decoded_rest_len = 3;
185 if ((*src & 0xC0) == 0x80) {
186 decoded = (decoded << 6) | (*src++ & 0x3F);
193 input_error(0, 0, "invalid byte sequence in input");
198 input->utf8_part_decoded_rest_len = 4;
201 } while ((*src & 0xC0) == 0x80 || (*src & 0xF8) == 0xF8);
207 /* we're done when we could read more than 1 char */
213 static size_t decode_windows_1252(input_t *input, utf32 *buffer,
216 unsigned char read_buf[buffer_size];
217 size_t const s = read_block(input, read_buf, sizeof(read_buf));
219 unsigned char const *src = read_buf;
220 unsigned char const *end = read_buf + s;
225 case 0x80: tc = 0x20AC; break; // €
226 case 0x82: tc = 0x201A; break; // ‚
227 case 0x83: tc = 0x0192; break; // ƒ
228 case 0x84: tc = 0x201E; break; // „
229 case 0x85: tc = 0x2026; break; // …
230 case 0x86: tc = 0x2020; break; // †
231 case 0x87: tc = 0x2021; break; // ‡
232 case 0x88: tc = 0x02C6; break; // ˆ
233 case 0x89: tc = 0x2030; break; // ‰
234 case 0x8A: tc = 0x0160; break; // Š
235 case 0x8B: tc = 0x2039; break; // ‹
236 case 0x8C: tc = 0x0152; break; // Œ
237 case 0x8E: tc = 0x017D; break; // Ž
238 case 0x91: tc = 0x2018; break; // ‘
239 case 0x92: tc = 0x2019; break; // ’
240 case 0x93: tc = 0x201C; break; // “
241 case 0x94: tc = 0x201D; break; // ”
242 case 0x95: tc = 0x2022; break; // •
243 case 0x96: tc = 0x2013; break; // –
244 case 0x97: tc = 0x2014; break; // —
245 case 0x98: tc = 0x02DC; break; // ˜
246 case 0x99: tc = 0x2122; break; // ™
247 case 0x9A: tc = 0x0161; break; // š
248 case 0x9B: tc = 0x203A; break; // ›
249 case 0x9C: tc = 0x0153; break; // œ
250 case 0x9E: tc = 0x017E; break; // ž
251 case 0x9F: tc = 0x0178; break; // Ÿ
259 typedef struct named_decoder_t {
264 static named_decoder_t const decoders[] = {
265 { "CP819", decode_iso_8859_1 }, // official alias
266 { "IBM819", decode_iso_8859_1 }, // official alias
267 { "ISO-8859-1", decode_iso_8859_1 }, // official alias
268 { "ISO-8859-15", decode_iso_8859_15 }, // official name
269 { "ISO8859-1", decode_iso_8859_1 },
270 { "ISO8859-15", decode_iso_8859_15 },
271 { "ISO_8859-1", decode_iso_8859_1 }, // official alias
272 { "ISO_8859-15", decode_iso_8859_15 }, // official alias
273 { "ISO_8859-1:1987", decode_iso_8859_1 }, // official name
274 { "Latin-9", decode_iso_8859_15 }, // official alias
275 { "UTF-8", decode_utf8 }, // official name
276 { "csISOLatin1", decode_iso_8859_1 }, // official alias
277 { "cp1252", decode_windows_1252 },
278 { "iso-ir-100", decode_iso_8859_1 }, // official alias
279 { "l1", decode_iso_8859_1 }, // official alias
280 { "latin1", decode_iso_8859_1 }, // official alias
281 { "windows-1252", decode_windows_1252 }, // official name
286 /** strcasecmp is not part of C99 so we need our own implementation here */
287 static int my_strcasecmp(const char *s1, const char *s2)
289 for ( ; *s1 != 0; ++s1, ++s2) {
290 if (tolower(*s1) != tolower(*s2))
293 return (unsigned char)*s1 - (unsigned char)*s2;
296 static void choose_decoder(input_t *result, const char *encoding)
298 if (encoding == NULL) {
299 result->decode = decode_utf8;
301 for (named_decoder_t const *i = decoders; i->name != NULL; ++i) {
302 if (my_strcasecmp(encoding, i->name) != 0)
304 result->decode = i->decoder;
307 if (result->decode == NULL) {
308 fprintf(stderr, "error: input encoding \"%s\" not supported\n",
310 result->decode = decode_utf8;
315 input_t *input_from_stream(FILE *file, const char *encoding)
317 input_t *result = XMALLOCZ(input_t);
318 result->kind = INPUT_FILE;
319 result->in.file = file;
321 choose_decoder(result, encoding);
326 input_t *input_from_string(const char *string, const char *encoding)
328 input_t *result = XMALLOCZ(input_t);
329 result->kind = INPUT_STRING;
330 result->in.string = string;
332 choose_decoder(result, encoding);
337 size_t decode(input_t *input, utf32 *buffer, size_t buffer_size)
339 return input->decode(input, buffer, buffer_size);
342 void input_free(input_t *input)