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.
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
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.
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
29 * Encoding -- determine the character encoding of a text file.
31 * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit
32 * international characters.
38 FILE_RCSID("@(#)$File: encoding.c,v 1.13 2015/06/04 19:16:28 christos Exp $")
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 *,
50 private int looks_utf7(const unsigned char *, size_t, unichar *, size_t *);
51 private int looks_ucs16(const unsigned char *, size_t, unichar *, size_t *);
52 private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *);
53 private int looks_extended(const unsigned char *, size_t, unichar *, size_t *);
54 private void from_ebcdic(const unsigned char *, size_t, unsigned char *);
57 #define DPRINTF(a) printf a
63 * Try to determine whether text is in some character code we can
64 * identify. Each of these tests, if it succeeds, will leave
65 * the text converted into one-unichar-per-character Unicode in
66 * ubuf, and the number of characters converted in ulen.
69 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)
73 unsigned char *nbuf = NULL;
78 *code_mime = "binary";
80 mlen = (nbytes + 1) * sizeof((*ubuf)[0]);
81 if ((*ubuf = CAST(unichar *, calloc((size_t)1, mlen))) == NULL) {
85 mlen = (nbytes + 1) * sizeof(nbuf[0]);
86 if ((nbuf = CAST(unsigned char *, calloc((size_t)1, mlen))) == NULL) {
91 if (looks_ascii(buf, nbytes, *ubuf, ulen)) {
92 if (looks_utf7(buf, nbytes, *ubuf, ulen) > 0) {
93 DPRINTF(("utf-7 %" SIZE_T_FORMAT "u\n", *ulen));
94 *code = "UTF-7 Unicode";
97 DPRINTF(("ascii %" SIZE_T_FORMAT "u\n", *ulen));
99 *code_mime = "us-ascii";
101 } else if (looks_utf8_with_BOM(buf, nbytes, *ubuf, ulen) > 0) {
102 DPRINTF(("utf8/bom %" SIZE_T_FORMAT "u\n", *ulen));
103 *code = "UTF-8 Unicode (with BOM)";
104 *code_mime = "utf-8";
105 } else if (file_looks_utf8(buf, nbytes, *ubuf, ulen) > 1) {
106 DPRINTF(("utf8 %" SIZE_T_FORMAT "u\n", *ulen));
107 *code = "UTF-8 Unicode";
108 *code_mime = "utf-8";
109 } else if ((ucs_type = looks_ucs16(buf, nbytes, *ubuf, ulen)) != 0) {
111 *code = "Little-endian UTF-16 Unicode";
112 *code_mime = "utf-16le";
114 *code = "Big-endian UTF-16 Unicode";
115 *code_mime = "utf-16be";
117 DPRINTF(("ucs16 %" SIZE_T_FORMAT "u\n", *ulen));
118 } else if (looks_latin1(buf, nbytes, *ubuf, ulen)) {
119 DPRINTF(("latin1 %" SIZE_T_FORMAT "u\n", *ulen));
121 *code_mime = "iso-8859-1";
122 } else if (looks_extended(buf, nbytes, *ubuf, ulen)) {
123 DPRINTF(("extended %" SIZE_T_FORMAT "u\n", *ulen));
124 *code = "Non-ISO extended-ASCII";
125 *code_mime = "unknown-8bit";
127 from_ebcdic(buf, nbytes, nbuf);
129 if (looks_ascii(nbuf, nbytes, *ubuf, ulen)) {
130 DPRINTF(("ebcdic %" SIZE_T_FORMAT "u\n", *ulen));
132 *code_mime = "ebcdic";
133 } else if (looks_latin1(nbuf, nbytes, *ubuf, ulen)) {
134 DPRINTF(("ebcdic/international %" SIZE_T_FORMAT "u\n",
136 *code = "International EBCDIC";
137 *code_mime = "ebcdic";
138 } else { /* Doesn't look like text at all */
139 DPRINTF(("binary\n"));
152 * This table reflects a particular philosophy about what constitutes
153 * "text," and there is room for disagreement about it.
155 * Version 3.31 of the file command considered a file to be ASCII if
156 * each of its characters was approved by either the isascii() or
157 * isalpha() function. On most systems, this would mean that any
158 * file consisting only of characters in the range 0x00 ... 0x7F
159 * would be called ASCII text, but many systems might reasonably
160 * consider some characters outside this range to be alphabetic,
161 * so the file command would call such characters ASCII. It might
162 * have been more accurate to call this "considered textual on the
163 * local system" than "ASCII."
165 * It considered a file to be "International language text" if each
166 * of its characters was either an ASCII printing character (according
167 * to the real ASCII standard, not the above test), a character in
168 * the range 0x80 ... 0xFF, or one of the following control characters:
169 * backspace, tab, line feed, vertical tab, form feed, carriage return,
170 * escape. No attempt was made to determine the language in which files
171 * of this type were written.
174 * The table below considers a file to be ASCII if all of its characters
175 * are either ASCII printing characters (again, according to the X3.4
176 * standard, not isascii()) or any of the following controls: bell,
177 * backspace, tab, line feed, form feed, carriage return, esc, nextline.
179 * I include bell because some programs (particularly shell scripts)
180 * use it literally, even though it is rare in normal text. I exclude
181 * vertical tab because it never seems to be used in real text. I also
182 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85),
183 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline
184 * character to. It might be more appropriate to include it in the 8859
185 * set instead of the ASCII set, but it's got to be included in *something*
186 * we recognize or EBCDIC files aren't going to be considered textual.
187 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek
188 * and Latin characters, so these should possibly be allowed. But they
189 * make a real mess on VT100-style displays if they're not paired properly,
190 * so we are probably better off not calling them text.
192 * A file is considered to be ISO-8859 text if its characters are all
193 * either ASCII, according to the above definition, or printing characters
194 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF.
196 * Finally, a file is considered to be international text from some other
197 * character code if its characters are all either ISO-8859 (according to
198 * the above definition) or characters in the range 0x80 ... 0x9F, which
199 * ISO-8859 considers to be control characters but the IBM PC and Macintosh
200 * consider to be printing characters.
203 #define F 0 /* character never appears in text */
204 #define T 1 /* character appears in plain ASCII text */
205 #define I 2 /* character appears in ISO-8859 text */
206 #define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */
208 private char text_chars[256] = {
209 /* BEL BS HT LF VT FF CR */
210 F, F, F, F, F, F, F, T, T, T, T, T, T, T, F, F, /* 0x0X */
212 F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */
213 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */
214 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */
215 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */
216 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */
217 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */
218 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */
220 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */
221 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */
222 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */
223 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */
224 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */
225 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */
226 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */
227 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */
231 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf,
238 for (i = 0; i < nbytes; i++) {
239 int t = text_chars[buf[i]];
244 ubuf[(*ulen)++] = buf[i];
251 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
257 for (i = 0; i < nbytes; i++) {
258 int t = text_chars[buf[i]];
260 if (t != T && t != I)
263 ubuf[(*ulen)++] = buf[i];
270 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf,
277 for (i = 0; i < nbytes; i++) {
278 int t = text_chars[buf[i]];
280 if (t != T && t != I && t != X)
283 ubuf[(*ulen)++] = buf[i];
290 * Decide whether some text looks like UTF-8. Returns:
293 * 0: uses odd control characters, so doesn't look like text
295 * 2: definitely UTF-8 text (valid high-bit set bytes)
297 * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen;
298 * ubuf must be big enough!
301 file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
306 int gotone = 0, ctrl = 0;
311 for (i = 0; i < nbytes; i++) {
312 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */
314 * Even if the whole file is valid UTF-8 sequences,
315 * still reject it if it uses weird control characters.
318 if (text_chars[buf[i]] != T)
322 ubuf[(*ulen)++] = buf[i];
323 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */
325 } else { /* 11xxxxxx begins UTF-8 */
328 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */
331 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */
334 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */
337 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */
340 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */
346 for (n = 0; n < following; n++) {
351 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40))
354 c = (c << 6) + (buf[i] & 0x3f);
363 return ctrl ? 0 : (gotone ? 2 : 1);
367 * Decide whether some text looks like UTF-8 with BOM. If there is no
368 * BOM, return -1; otherwise return the result of looks_utf8 on the
372 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf,
375 if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf)
376 return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen);
382 looks_utf7(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
384 if (nbytes > 4 && buf[0] == '+' && buf[1] == '/' && buf[2] == 'v')
401 looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf,
410 if (buf[0] == 0xff && buf[1] == 0xfe)
412 else if (buf[0] == 0xfe && buf[1] == 0xff)
419 for (i = 2; i + 1 < nbytes; i += 2) {
420 /* XXX fix to properly handle chars > 65536 */
423 ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i];
425 ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1];
427 if (ubuf[*ulen - 1] == 0xfffe)
429 if (ubuf[*ulen - 1] < 128 &&
430 text_chars[(size_t)ubuf[*ulen - 1]] != T)
443 * This table maps each EBCDIC character to an (8-bit extended) ASCII
444 * character, as specified in the rationale for the dd(1) command in
445 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard.
447 * Unfortunately it does not seem to correspond exactly to any of the
448 * five variants of EBCDIC documented in IBM's _Enterprise Systems
449 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh
450 * Edition, July, 1999, pp. I-1 - I-4.
452 * Fortunately, though, all versions of EBCDIC, including this one, agree
453 * on most of the printing characters that also appear in (7-bit) ASCII.
454 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all.
456 * Fortunately too, there is general agreement that codes 0x00 through
457 * 0x3F represent control characters, 0x41 a nonbreaking space, and the
458 * remainder printing characters.
460 * This is sufficient to allow us to identify EBCDIC text and to distinguish
461 * between old-style and internationalized examples of text.
464 private unsigned char ebcdic_to_ascii[] = {
465 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15,
466 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31,
467 128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7,
468 144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26,
469 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|',
470 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~',
471 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?',
472 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"',
473 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201,
474 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208,
475 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215,
476 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231,
477 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237,
478 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243,
479 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249,
480 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255
485 * The following EBCDIC-to-ASCII table may relate more closely to reality,
486 * or at least to modern reality. It comes from
488 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html
490 * and maps the characters of EBCDIC code page 1047 (the code used for
491 * Unix-derived software on IBM's 390 systems) to the corresponding
492 * characters from ISO 8859-1.
494 * If this table is used instead of the above one, some of the special
495 * cases for the NEL character can be taken out of the code.
498 private unsigned char ebcdic_1047_to_8859[] = {
499 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F,
500 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F,
501 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07,
502 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A,
503 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C,
504 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E,
505 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F,
506 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22,
507 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1,
508 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4,
509 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE,
510 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7,
511 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5,
512 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF,
513 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5,
514 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F
519 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII.
522 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out)
526 for (i = 0; i < nbytes; i++) {
527 out[i] = ebcdic_to_ascii[buf[i]];