1 /*
2 * Copyright (c) Ian F. Darwin 1986-1995.
3 * Software written by Ian F. Darwin and others;
4 * maintained 1995-present by Christos Zoulas and others.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice immediately at the beginning of the file, without modification,
11 * this list of conditions, and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
20 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28 /*
29 * Encoding -- determine the character encoding of a text file.
30 *
31 * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit
32 * international characters.
33 */
34
35 #include "file.h"
36
37 #ifndef lint
38 FILE_RCSID("@(#)$File: encoding.c,v 1.10 2014/09/11 12:08:52 christos Exp $")
39 #endif /* lint */
40
41 #include "magic.h"
42 #include <string.h>
43 #include <memory.h>
44 #include <stdlib.h>
45
46
47 private int looks_ascii(const unsigned char *, size_t, unichar *, size_t *);
48 private int looks_utf8_with_BOM(const unsigned char *, size_t, unichar *,
49 size_t *);
50 private int looks_ucs16(const unsigned char *, size_t, unichar *, size_t *);
51 private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *);
52 private int looks_extended(const unsigned char *, size_t, unichar *, size_t *);
53 private void from_ebcdic(const unsigned char *, size_t, unsigned char *);
54
55 #ifdef DEBUG_ENCODING
56 #define DPRINTF(a) printf a
57 #else
58 #define DPRINTF(a)
59 #endif
60
61 /*
62 * Try to determine whether text is in some character code we can
63 * identify. Each of these tests, if it succeeds, will leave
64 * the text converted into one-unichar-per-character Unicode in
65 * ubuf, and the number of characters converted in ulen.
66 */
67 protected int
68 file_encoding(struct magic_set *ms, const unsigned char *buf, size_t nbytes, unichar **ubuf, size_t *ulen, const char **code, const char **code_mime, const char **type)
69 {
70 size_t mlen;
71 int rv = 1, ucs_type;
72 unsigned char *nbuf = NULL;
73
74 *type = "text";
75 *ulen = 0;
76 *code = "unknown";
77 *code_mime = "binary";
78
79 mlen = (nbytes + 1) * sizeof((*ubuf)[0]);
80 if ((*ubuf = CAST(unichar *, calloc((size_t)1, mlen))) == NULL) {
81 file_oomem(ms, mlen);
82 goto done;
83 }
84 mlen = (nbytes + 1) * sizeof(nbuf[0]);
85 if ((nbuf = CAST(unsigned char *, calloc((size_t)1, mlen))) == NULL) {
86 file_oomem(ms, mlen);
87 goto done;
88 }
89
90 if (looks_ascii(buf, nbytes, *ubuf, ulen)) {
91 DPRINTF(("ascii %" SIZE_T_FORMAT "u\n", *ulen));
92 *code = "ASCII";
93 *code_mime = "us-ascii";
94 } else if (looks_utf8_with_BOM(buf, nbytes, *ubuf, ulen) > 0) {
95 DPRINTF(("utf8/bom %" SIZE_T_FORMAT "u\n", *ulen));
96 *code = "UTF-8 Unicode (with BOM)";
97 *code_mime = "utf-8";
98 } else if (file_looks_utf8(buf, nbytes, *ubuf, ulen) > 1) {
99 DPRINTF(("utf8 %" SIZE_T_FORMAT "u\n", *ulen));
100 *code = "UTF-8 Unicode";
101 *code_mime = "utf-8";
102 } else if ((ucs_type = looks_ucs16(buf, nbytes, *ubuf, ulen)) != 0) {
103 if (ucs_type == 1) {
104 *code = "Little-endian UTF-16 Unicode";
105 *code_mime = "utf-16le";
106 } else {
107 *code = "Big-endian UTF-16 Unicode";
108 *code_mime = "utf-16be";
109 }
110 DPRINTF(("ucs16 %" SIZE_T_FORMAT "u\n", *ulen));
111 } else if (looks_latin1(buf, nbytes, *ubuf, ulen)) {
112 DPRINTF(("latin1 %" SIZE_T_FORMAT "u\n", *ulen));
113 *code = "ISO-8859";
114 *code_mime = "iso-8859-1";
115 } else if (looks_extended(buf, nbytes, *ubuf, ulen)) {
116 DPRINTF(("extended %" SIZE_T_FORMAT "u\n", *ulen));
117 *code = "Non-ISO extended-ASCII";
118 *code_mime = "unknown-8bit";
119 } else {
120 from_ebcdic(buf, nbytes, nbuf);
121
122 if (looks_ascii(nbuf, nbytes, *ubuf, ulen)) {
123 DPRINTF(("ebcdic %" SIZE_T_FORMAT "u\n", *ulen));
124 *code = "EBCDIC";
125 *code_mime = "ebcdic";
126 } else if (looks_latin1(nbuf, nbytes, *ubuf, ulen)) {
127 DPRINTF(("ebcdic/international %" SIZE_T_FORMAT "u\n",
128 *ulen));
129 *code = "International EBCDIC";
130 *code_mime = "ebcdic";
131 } else { /* Doesn't look like text at all */
132 DPRINTF(("binary\n"));
133 rv = 0;
134 *type = "binary";
135 }
136 }
137
138 done:
139 free(nbuf);
140
141 return rv;
142 }
143
144 /*
145 * This table reflects a particular philosophy about what constitutes
146 * "text," and there is room for disagreement about it.
147 *
148 * Version 3.31 of the file command considered a file to be ASCII if
149 * each of its characters was approved by either the isascii() or
150 * isalpha() function. On most systems, this would mean that any
151 * file consisting only of characters in the range 0x00 ... 0x7F
152 * would be called ASCII text, but many systems might reasonably
153 * consider some characters outside this range to be alphabetic,
154 * so the file command would call such characters ASCII. It might
155 * have been more accurate to call this "considered textual on the
156 * local system" than "ASCII."
157 *
158 * It considered a file to be "International language text" if each
159 * of its characters was either an ASCII printing character (according
160 * to the real ASCII standard, not the above test), a character in
161 * the range 0x80 ... 0xFF, or one of the following control characters:
162 * backspace, tab, line feed, vertical tab, form feed, carriage return,
163 * escape. No attempt was made to determine the language in which files
164 * of this type were written.
165 *
166 *
167 * The table below considers a file to be ASCII if all of its characters
168 * are either ASCII printing characters (again, according to the X3.4
169 * standard, not isascii()) or any of the following controls: bell,
170 * backspace, tab, line feed, form feed, carriage return, esc, nextline.
171 *
172 * I include bell because some programs (particularly shell scripts)
173 * use it literally, even though it is rare in normal text. I exclude
174 * vertical tab because it never seems to be used in real text. I also
175 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85),
176 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline
177 * character to. It might be more appropriate to include it in the 8859
178 * set instead of the ASCII set, but it's got to be included in *something*
179 * we recognize or EBCDIC files aren't going to be considered textual.
180 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek
181 * and Latin characters, so these should possibly be allowed. But they
182 * make a real mess on VT100-style displays if they're not paired properly,
183 * so we are probably better off not calling them text.
184 *
185 * A file is considered to be ISO-8859 text if its characters are all
186 * either ASCII, according to the above definition, or printing characters
187 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF.
188 *
189 * Finally, a file is considered to be international text from some other
190 * character code if its characters are all either ISO-8859 (according to
191 * the above definition) or characters in the range 0x80 ... 0x9F, which
192 * ISO-8859 considers to be control characters but the IBM PC and Macintosh
193 * consider to be printing characters.
194 */
195
196 #define F 0 /* character never appears in text */
197 #define T 1 /* character appears in plain ASCII text */
198 #define I 2 /* character appears in ISO-8859 text */
199 #define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */
200
201 private char text_chars[256] = {
202 /* BEL BS HT LF VT FF CR */
203 F, F, F, F, F, F, F, T, T, T, T, T, T, T, F, F, /* 0x0X */
204 /* ESC */
205 F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */
206 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */
207 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */
208 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */
209 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */
210 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */
211 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */
212 /* NEL */
213 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */
214 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */
215 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */
216 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */
217 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */
218 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */
219 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */
220 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */
221 };
222
223 private int
224 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf,
225 size_t *ulen)
226 {
227 size_t i;
228
229 *ulen = 0;
230
231 for (i = 0; i < nbytes; i++) {
232 int t = text_chars[buf[i]];
233
234 if (t != T)
235 return 0;
236
237 ubuf[(*ulen)++] = buf[i];
238 }
239
240 return 1;
241 }
242
243 private int
244 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
245 {
246 size_t i;
247
248 *ulen = 0;
249
250 for (i = 0; i < nbytes; i++) {
251 int t = text_chars[buf[i]];
252
253 if (t != T && t != I)
254 return 0;
255
256 ubuf[(*ulen)++] = buf[i];
257 }
258
259 return 1;
260 }
261
262 private int
263 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf,
264 size_t *ulen)
265 {
266 size_t i;
267
268 *ulen = 0;
269
270 for (i = 0; i < nbytes; i++) {
271 int t = text_chars[buf[i]];
272
273 if (t != T && t != I && t != X)
274 return 0;
275
276 ubuf[(*ulen)++] = buf[i];
277 }
278
279 return 1;
280 }
281
282 /*
283 * Decide whether some text looks like UTF-8. Returns:
284 *
285 * -1: invalid UTF-8
286 * 0: uses odd control characters, so doesn't look like text
287 * 1: 7-bit text
288 * 2: definitely UTF-8 text (valid high-bit set bytes)
289 *
290 * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen;
291 * ubuf must be big enough!
292 */
293 protected int
294 file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
295 {
296 size_t i;
297 int n;
298 unichar c;
299 int gotone = 0, ctrl = 0;
300
301 if (ubuf)
302 *ulen = 0;
303
304 for (i = 0; i < nbytes; i++) {
305 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */
306 /*
307 * Even if the whole file is valid UTF-8 sequences,
308 * still reject it if it uses weird control characters.
309 */
310
311 if (text_chars[buf[i]] != T)
312 ctrl = 1;
313
314 if (ubuf)
315 ubuf[(*ulen)++] = buf[i];
316 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */
317 return -1;
318 } else { /* 11xxxxxx begins UTF-8 */
319 int following;
320
321 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */
322 c = buf[i] & 0x1f;
323 following = 1;
324 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */
325 c = buf[i] & 0x0f;
326 following = 2;
327 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */
328 c = buf[i] & 0x07;
329 following = 3;
330 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */
331 c = buf[i] & 0x03;
332 following = 4;
333 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */
334 c = buf[i] & 0x01;
335 following = 5;
336 } else
337 return -1;
338
339 for (n = 0; n < following; n++) {
340 i++;
341 if (i >= nbytes)
342 goto done;
343
344 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40))
345 return -1;
346
347 c = (c << 6) + (buf[i] & 0x3f);
348 }
349
350 if (ubuf)
351 ubuf[(*ulen)++] = c;
352 gotone = 1;
353 }
354 }
355 done:
356 return ctrl ? 0 : (gotone ? 2 : 1);
357 }
358
359 /*
360 * Decide whether some text looks like UTF-8 with BOM. If there is no
361 * BOM, return -1; otherwise return the result of looks_utf8 on the
362 * rest of the text.
363 */
364 private int
365 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf,
366 size_t *ulen)
367 {
368 if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf)
369 return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen);
370 else
371 return -1;
372 }
373
374 private int
375 looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf,
376 size_t *ulen)
377 {
378 int bigend;
379 size_t i;
380
381 if (nbytes < 2)
382 return 0;
383
384 if (buf[0] == 0xff && buf[1] == 0xfe)
385 bigend = 0;
386 else if (buf[0] == 0xfe && buf[1] == 0xff)
387 bigend = 1;
388 else
389 return 0;
390
391 *ulen = 0;
392
393 for (i = 2; i + 1 < nbytes; i += 2) {
394 /* XXX fix to properly handle chars > 65536 */
395
396 if (bigend)
397 ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i];
398 else
399 ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1];
400
401 if (ubuf[*ulen - 1] == 0xfffe)
402 return 0;
403 if (ubuf[*ulen - 1] < 128 &&
404 text_chars[(size_t)ubuf[*ulen - 1]] != T)
405 return 0;
406 }
407
408 return 1 + bigend;
409 }
410
411 #undef F
412 #undef T
413 #undef I
414 #undef X
415
416 /*
417 * This table maps each EBCDIC character to an (8-bit extended) ASCII
418 * character, as specified in the rationale for the dd(1) command in
419 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard.
420 *
421 * Unfortunately it does not seem to correspond exactly to any of the
422 * five variants of EBCDIC documented in IBM's _Enterprise Systems
423 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh
424 * Edition, July, 1999, pp. I-1 - I-4.
425 *
426 * Fortunately, though, all versions of EBCDIC, including this one, agree
427 * on most of the printing characters that also appear in (7-bit) ASCII.
428 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all.
429 *
430 * Fortunately too, there is general agreement that codes 0x00 through
431 * 0x3F represent control characters, 0x41 a nonbreaking space, and the
432 * remainder printing characters.
433 *
434 * This is sufficient to allow us to identify EBCDIC text and to distinguish
435 * between old-style and internationalized examples of text.
436 */
437
438 private unsigned char ebcdic_to_ascii[] = {
439 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15,
440 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31,
441 128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7,
442 144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26,
443 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|',
444 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~',
445 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?',
446 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"',
447 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201,
448 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208,
449 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215,
450 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231,
451 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237,
452 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243,
453 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249,
454 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255
455 };
456
457 #ifdef notdef
458 /*
459 * The following EBCDIC-to-ASCII table may relate more closely to reality,
460 * or at least to modern reality. It comes from
461 *
462 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html
463 *
464 * and maps the characters of EBCDIC code page 1047 (the code used for
465 * Unix-derived software on IBM's 390 systems) to the corresponding
466 * characters from ISO 8859-1.
467 *
468 * If this table is used instead of the above one, some of the special
469 * cases for the NEL character can be taken out of the code.
470 */
471
472 private unsigned char ebcdic_1047_to_8859[] = {
473 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F,
474 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F,
475 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07,
476 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A,
477 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C,
478 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E,
479 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F,
480 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22,
481 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1,
482 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4,
483 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE,
484 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7,
485 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5,
486 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF,
487 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5,
488 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F
489 };
490 #endif
491
492 /*
493 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII.
494 */
495 private void
496 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out)
497 {
498 size_t i;
499
500 for (i = 0; i < nbytes; i++) {
501 out[i] = ebcdic_to_ascii[buf[i]];
502 }
503 }