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.
15 * 3. All advertising materials mentioning features or use of this software
16 * must display the following acknowledgement:
17 * This product includes software developed by Ian F. Darwin and others.
18 * 4. The name of the author may not be used to endorse or promote products
19 * derived from this software without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
25 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * ASCII magic -- file types that we know based on keywords
35 * that can appear anywhere in the file.
37 * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000,
38 * to handle character codes other than ASCII on a unified basis.
40 * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit
41 * international characters, now subsumed into this file.
57 FILE_RCSID("@(#)$Id: ascmagic.c,v 1.40 2003/11/20 00:25:39 christos Exp $")
60 typedef unsigned long unichar;
62 #define MAXLINELEN 300 /* longest sane line length */
63 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
64 || (x) == 0x85 || (x) == '\f')
66 private int looks_ascii(const unsigned char *, size_t, unichar *, size_t *);
67 private int looks_utf8(const unsigned char *, size_t, unichar *, size_t *);
68 private int looks_unicode(const unsigned char *, size_t, unichar *, size_t *);
69 private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *);
70 private int looks_extended(const unsigned char *, size_t, unichar *, size_t *);
71 private void from_ebcdic(const unsigned char *, size_t, unsigned char *);
72 private int ascmatch(const unsigned char *, const unichar *, size_t);
76 file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
79 unsigned char nbuf[HOWMANY+1]; /* one extra for terminating '\0' */
80 unichar ubuf[HOWMANY+1]; /* one extra for terminating '\0' */
84 const char *code = NULL;
85 const char *code_mime = NULL;
86 const char *type = NULL;
87 const char *subtype = NULL;
88 const char *subtype_mime = NULL;
91 int has_backspace = 0;
98 int last_line_end = -1;
99 int has_long_lines = 0;
102 * Undo the NUL-termination kindly provided by process()
103 * but leave at least one byte to look at
106 while (nbytes > 1 && buf[nbytes - 1] == '\0')
109 /* nbuf and ubuf relies on this */
110 if (nbytes > HOWMANY)
114 * Then try to determine whether it's any character code we can
115 * identify. Each of these tests, if it succeeds, will leave
116 * the text converted into one-unichar-per-character Unicode in
117 * ubuf, and the number of characters converted in ulen.
119 if (looks_ascii(buf, nbytes, ubuf, &ulen)) {
121 code_mime = "us-ascii";
123 } else if (looks_utf8(buf, nbytes, ubuf, &ulen)) {
124 code = "UTF-8 Unicode";
127 } else if ((i = looks_unicode(buf, nbytes, ubuf, &ulen)) != 0) {
129 code = "Little-endian UTF-16 Unicode";
131 code = "Big-endian UTF-16 Unicode";
133 type = "character data";
134 code_mime = "utf-16"; /* is this defined? */
135 } else if (looks_latin1(buf, nbytes, ubuf, &ulen)) {
138 code_mime = "iso-8859-1";
139 } else if (looks_extended(buf, nbytes, ubuf, &ulen)) {
140 code = "Non-ISO extended-ASCII";
142 code_mime = "unknown";
144 from_ebcdic(buf, nbytes, nbuf);
146 if (looks_ascii(nbuf, nbytes, ubuf, &ulen)) {
148 type = "character data";
149 code_mime = "ebcdic";
150 } else if (looks_latin1(nbuf, nbytes, ubuf, &ulen)) {
151 code = "International EBCDIC";
152 type = "character data";
153 code_mime = "ebcdic";
155 return 0; /* doesn't look like text at all */
160 * for troff, look for . + letter + letter or .\";
161 * this must be done to disambiguate tar archives' ./file
162 * and other trash from real troff input.
164 * I believe Plan 9 troff allows non-ASCII characters in the names
165 * of macros, so this test might possibly fail on such a file.
168 unichar *tp = ubuf + 1;
171 ++tp; /* skip leading whitespace */
172 if ((tp[0] == '\\' && tp[1] == '\"') ||
173 (isascii((unsigned char)tp[0]) &&
174 isalnum((unsigned char)tp[0]) &&
175 isascii((unsigned char)tp[1]) &&
176 isalnum((unsigned char)tp[1]) &&
178 subtype_mime = "text/troff";
179 subtype = "troff or preprocessor input";
180 goto subtype_identified;
184 if ((*buf == 'c' || *buf == 'C') && ISSPC(buf[1])) {
185 subtype_mime = "text/fortran";
186 subtype = "fortran program";
187 goto subtype_identified;
190 /* look for tokens from names.h - this is expensive! */
197 * skip past any leading space
199 while (i < ulen && ISSPC(ubuf[i]))
205 * find the next whitespace
207 for (end = i + 1; end < nbytes; end++)
208 if (ISSPC(ubuf[end]))
212 * compare the word thus isolated against the token list
214 for (p = names; p < names + NNAMES; p++) {
215 if (ascmatch((const unsigned char *)p->name, ubuf + i,
217 subtype = types[p->type].human;
218 subtype_mime = types[p->type].mime;
219 goto subtype_identified;
229 * Now try to discover other details about the file.
231 for (i = 0; i < ulen; i++) {
232 if (i > last_line_end + MAXLINELEN)
235 if (ubuf[i] == '\033')
240 if (ubuf[i] == '\r' && (i + 1 < ulen && ubuf[i + 1] == '\n')) {
244 if (ubuf[i] == '\r' && (i + 1 >= ulen || ubuf[i + 1] != '\n')) {
248 if (ubuf[i] == '\n' && ((int)i - 1 < 0 || ubuf[i - 1] != '\r')){
252 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
258 if ((ms->flags & MAGIC_MIME)) {
260 if (file_printf(ms, subtype_mime) == -1)
263 if (file_printf(ms, "text/plain") == -1)
268 if (file_printf(ms, "; charset=") == -1)
270 if (file_printf(ms, code_mime) == -1)
274 if (file_printf(ms, code) == -1)
278 if (file_printf(ms, " ") == -1)
280 if (file_printf(ms, subtype) == -1)
284 if (file_printf(ms, " ") == -1)
286 if (file_printf(ms, type) == -1)
290 if (file_printf(ms, ", with very long lines") == -1)
294 * Only report line terminators if we find one other than LF,
295 * or if we find none at all.
297 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
298 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
299 if (file_printf(ms, ", with") == -1)
302 if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
303 if (file_printf(ms, " no") == -1)
307 if (file_printf(ms, " CRLF") == -1)
309 if (n_cr || n_lf || n_nel)
310 if (file_printf(ms, ",") == -1)
314 if (file_printf(ms, " CR") == -1)
317 if (file_printf(ms, ",") == -1)
321 if (file_printf(ms, " LF") == -1)
324 if (file_printf(ms, ",") == -1)
328 if (file_printf(ms, " NEL") == -1)
332 if (file_printf(ms, " line terminators") == -1)
337 if (file_printf(ms, ", with escape sequences") == -1)
340 if (file_printf(ms, ", with overstriking") == -1)
348 ascmatch(const unsigned char *s, const unichar *us, size_t ulen)
352 for (i = 0; i < ulen; i++) {
364 * This table reflects a particular philosophy about what constitutes
365 * "text," and there is room for disagreement about it.
367 * Version 3.31 of the file command considered a file to be ASCII if
368 * each of its characters was approved by either the isascii() or
369 * isalpha() function. On most systems, this would mean that any
370 * file consisting only of characters in the range 0x00 ... 0x7F
371 * would be called ASCII text, but many systems might reasonably
372 * consider some characters outside this range to be alphabetic,
373 * so the file command would call such characters ASCII. It might
374 * have been more accurate to call this "considered textual on the
375 * local system" than "ASCII."
377 * It considered a file to be "International language text" if each
378 * of its characters was either an ASCII printing character (according
379 * to the real ASCII standard, not the above test), a character in
380 * the range 0x80 ... 0xFF, or one of the following control characters:
381 * backspace, tab, line feed, vertical tab, form feed, carriage return,
382 * escape. No attempt was made to determine the language in which files
383 * of this type were written.
386 * The table below considers a file to be ASCII if all of its characters
387 * are either ASCII printing characters (again, according to the X3.4
388 * standard, not isascii()) or any of the following controls: bell,
389 * backspace, tab, line feed, form feed, carriage return, esc, nextline.
391 * I include bell because some programs (particularly shell scripts)
392 * use it literally, even though it is rare in normal text. I exclude
393 * vertical tab because it never seems to be used in real text. I also
394 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85),
395 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline
396 * character to. It might be more appropriate to include it in the 8859
397 * set instead of the ASCII set, but it's got to be included in *something*
398 * we recognize or EBCDIC files aren't going to be considered textual.
399 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek
400 * and Latin characters, so these should possibly be allowed. But they
401 * make a real mess on VT100-style displays if they're not paired properly,
402 * so we are probably better off not calling them text.
404 * A file is considered to be ISO-8859 text if its characters are all
405 * either ASCII, according to the above definition, or printing characters
406 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF.
408 * Finally, a file is considered to be international text from some other
409 * character code if its characters are all either ISO-8859 (according to
410 * the above definition) or characters in the range 0x80 ... 0x9F, which
411 * ISO-8859 considers to be control characters but the IBM PC and Macintosh
412 * consider to be printing characters.
415 #define F 0 /* character never appears in text */
416 #define T 1 /* character appears in plain ASCII text */
417 #define I 2 /* character appears in ISO-8859 text */
418 #define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */
420 private char text_chars[256] = {
421 /* BEL BS HT LF FF CR */
422 F, F, F, F, F, F, F, T, T, T, T, F, T, T, F, F, /* 0x0X */
424 F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */
425 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */
426 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */
427 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */
428 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */
429 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */
430 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */
432 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */
433 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */
434 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */
435 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */
436 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */
437 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */
438 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */
439 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */
443 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf,
450 for (i = 0; i < nbytes; i++) {
451 int t = text_chars[buf[i]];
456 ubuf[(*ulen)++] = buf[i];
463 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
469 for (i = 0; i < nbytes; i++) {
470 int t = text_chars[buf[i]];
472 if (t != T && t != I)
475 ubuf[(*ulen)++] = buf[i];
482 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf,
489 for (i = 0; i < nbytes; i++) {
490 int t = text_chars[buf[i]];
492 if (t != T && t != I && t != X)
495 ubuf[(*ulen)++] = buf[i];
502 looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
510 for (i = 0; i < nbytes; i++) {
511 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */
513 * Even if the whole file is valid UTF-8 sequences,
514 * still reject it if it uses weird control characters.
517 if (text_chars[buf[i]] != T)
520 ubuf[(*ulen)++] = buf[i];
521 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */
523 } else { /* 11xxxxxx begins UTF-8 */
526 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */
529 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */
532 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */
535 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */
538 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */
544 for (n = 0; n < following; n++) {
549 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40))
552 c = (c << 6) + (buf[i] & 0x3f);
560 return gotone; /* don't claim it's UTF-8 if it's all 7-bit */
564 looks_unicode(const unsigned char *buf, size_t nbytes, unichar *ubuf,
573 if (buf[0] == 0xff && buf[1] == 0xfe)
575 else if (buf[0] == 0xfe && buf[1] == 0xff)
582 for (i = 2; i + 1 < nbytes; i += 2) {
583 /* XXX fix to properly handle chars > 65536 */
586 ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i];
588 ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1];
590 if (ubuf[*ulen - 1] == 0xfffe)
592 if (ubuf[*ulen - 1] < 128 &&
593 text_chars[(size_t)ubuf[*ulen - 1]] != T)
606 * This table maps each EBCDIC character to an (8-bit extended) ASCII
607 * character, as specified in the rationale for the dd(1) command in
608 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard.
610 * Unfortunately it does not seem to correspond exactly to any of the
611 * five variants of EBCDIC documented in IBM's _Enterprise Systems
612 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh
613 * Edition, July, 1999, pp. I-1 - I-4.
615 * Fortunately, though, all versions of EBCDIC, including this one, agree
616 * on most of the printing characters that also appear in (7-bit) ASCII.
617 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all.
619 * Fortunately too, there is general agreement that codes 0x00 through
620 * 0x3F represent control characters, 0x41 a nonbreaking space, and the
621 * remainder printing characters.
623 * This is sufficient to allow us to identify EBCDIC text and to distinguish
624 * between old-style and internationalized examples of text.
627 private unsigned char ebcdic_to_ascii[] = {
628 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15,
629 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31,
630 128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7,
631 144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26,
632 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|',
633 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~',
634 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?',
635 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"',
636 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201,
637 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208,
638 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215,
639 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231,
640 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237,
641 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243,
642 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249,
643 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255
648 * The following EBCDIC-to-ASCII table may relate more closely to reality,
649 * or at least to modern reality. It comes from
651 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html
653 * and maps the characters of EBCDIC code page 1047 (the code used for
654 * Unix-derived software on IBM's 390 systems) to the corresponding
655 * characters from ISO 8859-1.
657 * If this table is used instead of the above one, some of the special
658 * cases for the NEL character can be taken out of the code.
661 private unsigned char ebcdic_1047_to_8859[] = {
662 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F,
663 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F,
664 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07,
665 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A,
666 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C,
667 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E,
668 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F,
669 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22,
670 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1,
671 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4,
672 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE,
673 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7,
674 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5,
675 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF,
676 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5,
677 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F
682 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII.
685 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out)
689 for (i = 0; i < nbytes; i++) {
690 out[i] = ebcdic_to_ascii[buf[i]];