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 * ASCII magic -- file types that we know based on keywords
30 * that can appear anywhere in the file.
32 * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000,
33 * to handle character codes other than ASCII on a unified basis.
35 * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit
36 * international characters, now subsumed into this file.
52 FILE_RCSID("@(#)$File: ascmagic.c,v 1.64 2008/07/16 18:00:57 christos Exp $")
55 #define MAXLINELEN 300 /* longest sane line length */
56 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
57 || (x) == 0x85 || (x) == '\f')
59 private int looks_ascii(const unsigned char *, size_t, unichar *, size_t *);
60 private int looks_utf8_with_BOM(const unsigned char *, size_t, unichar *,
62 private int looks_ucs16(const unsigned char *, size_t, unichar *, size_t *);
63 private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *);
64 private int looks_extended(const unsigned char *, size_t, unichar *, size_t *);
65 private void from_ebcdic(const unsigned char *, size_t, unsigned char *);
66 private int ascmatch(const unsigned char *, const unichar *, size_t);
67 private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t);
71 file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
74 unsigned char *nbuf = NULL, *utf8_buf = NULL, *utf8_end;
77 const struct names *p;
79 int mime = ms->flags & MAGIC_MIME;
81 const char *code = NULL;
82 const char *code_mime = NULL;
83 const char *type = NULL;
84 const char *subtype = NULL;
85 const char *subtype_mime = NULL;
88 int has_backspace = 0;
96 size_t last_line_end = (size_t)-1;
97 int has_long_lines = 0;
100 * Undo the NUL-termination kindly provided by process()
101 * but leave at least one byte to look at
103 while (nbytes > 1 && buf[nbytes - 1] == '\0')
106 if ((nbuf = CAST(unsigned char *, calloc((size_t)1,
107 (nbytes + 1) * sizeof(nbuf[0])))) == NULL)
109 if ((ubuf = CAST(unichar *, calloc((size_t)1,
110 (nbytes + 1) * sizeof(ubuf[0])))) == NULL)
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_with_BOM(buf, nbytes, ubuf, &ulen) > 0) {
124 code = "UTF-8 Unicode (with BOM)";
127 } else if (file_looks_utf8(buf, nbytes, ubuf, &ulen) > 1) {
128 code = "UTF-8 Unicode";
131 } else if ((i = looks_ucs16(buf, nbytes, ubuf, &ulen)) != 0) {
133 code = "Little-endian UTF-16 Unicode";
135 code = "Big-endian UTF-16 Unicode";
137 type = "character data";
138 code_mime = "utf-16"; /* is this defined? */
139 } else if (looks_latin1(buf, nbytes, ubuf, &ulen)) {
142 code_mime = "iso-8859-1";
143 } else if (looks_extended(buf, nbytes, ubuf, &ulen)) {
144 code = "Non-ISO extended-ASCII";
146 code_mime = "unknown";
148 from_ebcdic(buf, nbytes, nbuf);
150 if (looks_ascii(nbuf, nbytes, ubuf, &ulen)) {
152 type = "character data";
153 code_mime = "ebcdic";
154 } else if (looks_latin1(nbuf, nbytes, ubuf, &ulen)) {
155 code = "International EBCDIC";
156 type = "character data";
157 code_mime = "ebcdic";
160 goto done; /* doesn't look like text at all */
169 /* Convert ubuf to UTF-8 and try text soft magic */
170 /* If original was ASCII or UTF-8, could use nbuf instead of
172 /* malloc size is a conservative overestimate; could be
173 re-converting improved, or at least realloced after
174 re-converting conversion. */
176 if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) {
177 file_oomem(ms, mlen);
180 if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen)) == NULL)
182 if (file_softmagic(ms, utf8_buf, utf8_end - utf8_buf, TEXTTEST) != 0) {
187 /* look for tokens from names.h - this is expensive! */
188 if ((ms->flags & MAGIC_NO_CHECK_TOKENS) != 0)
189 goto subtype_identified;
195 /* skip past any leading space */
196 while (i < ulen && ISSPC(ubuf[i]))
201 /* find the next whitespace */
202 for (end = i + 1; end < nbytes; end++)
203 if (ISSPC(ubuf[end]))
206 /* compare the word thus isolated against the token list */
207 for (p = names; p < names + NNAMES; p++) {
208 if (ascmatch((const unsigned char *)p->name, ubuf + i,
210 subtype = types[p->type].human;
211 subtype_mime = types[p->type].mime;
212 goto subtype_identified;
221 /* Now try to discover other details about the file. */
222 for (i = 0; i < ulen; i++) {
223 if (ubuf[i] == '\n') {
232 seen_cr = (ubuf[i] == '\r');
236 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
241 /* If this line is _longer_ than MAXLINELEN, remember it. */
242 if (i > last_line_end + MAXLINELEN)
245 if (ubuf[i] == '\033')
251 /* Beware, if the data has been truncated, the final CR could have
252 been followed by a LF. If we have HOWMANY bytes, it indicates
253 that the data might have been truncated, probably even before
254 this function was called. */
255 if (seen_cr && nbytes < HOWMANY)
259 if (mime & MAGIC_MIME_TYPE) {
261 if (file_printf(ms, subtype_mime) == -1)
264 if (file_printf(ms, "text/plain") == -1)
269 if ((mime == 0 || mime == MAGIC_MIME) && code_mime) {
270 if ((mime & MAGIC_MIME_TYPE) &&
271 file_printf(ms, " charset=") == -1)
273 if (file_printf(ms, code_mime) == -1)
277 if (mime == MAGIC_MIME_ENCODING)
278 file_printf(ms, "binary");
280 if (file_printf(ms, code) == -1)
284 if (file_printf(ms, " ") == -1)
286 if (file_printf(ms, subtype) == -1)
290 if (file_printf(ms, " ") == -1)
292 if (file_printf(ms, type) == -1)
296 if (file_printf(ms, ", with very long lines") == -1)
300 * Only report line terminators if we find one other than LF,
301 * or if we find none at all.
303 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
304 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
305 if (file_printf(ms, ", with") == -1)
308 if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
309 if (file_printf(ms, " no") == -1)
313 if (file_printf(ms, " CRLF") == -1)
315 if (n_cr || n_lf || n_nel)
316 if (file_printf(ms, ",") == -1)
320 if (file_printf(ms, " CR") == -1)
323 if (file_printf(ms, ",") == -1)
327 if (file_printf(ms, " LF") == -1)
330 if (file_printf(ms, ",") == -1)
334 if (file_printf(ms, " NEL") == -1)
338 if (file_printf(ms, " line terminators") == -1)
343 if (file_printf(ms, ", with escape sequences") == -1)
346 if (file_printf(ms, ", with overstriking") == -1)
362 ascmatch(const unsigned char *s, const unichar *us, size_t ulen)
366 for (i = 0; i < ulen; i++) {
378 * This table reflects a particular philosophy about what constitutes
379 * "text," and there is room for disagreement about it.
381 * Version 3.31 of the file command considered a file to be ASCII if
382 * each of its characters was approved by either the isascii() or
383 * isalpha() function. On most systems, this would mean that any
384 * file consisting only of characters in the range 0x00 ... 0x7F
385 * would be called ASCII text, but many systems might reasonably
386 * consider some characters outside this range to be alphabetic,
387 * so the file command would call such characters ASCII. It might
388 * have been more accurate to call this "considered textual on the
389 * local system" than "ASCII."
391 * It considered a file to be "International language text" if each
392 * of its characters was either an ASCII printing character (according
393 * to the real ASCII standard, not the above test), a character in
394 * the range 0x80 ... 0xFF, or one of the following control characters:
395 * backspace, tab, line feed, vertical tab, form feed, carriage return,
396 * escape. No attempt was made to determine the language in which files
397 * of this type were written.
400 * The table below considers a file to be ASCII if all of its characters
401 * are either ASCII printing characters (again, according to the X3.4
402 * standard, not isascii()) or any of the following controls: bell,
403 * backspace, tab, line feed, form feed, carriage return, esc, nextline.
405 * I include bell because some programs (particularly shell scripts)
406 * use it literally, even though it is rare in normal text. I exclude
407 * vertical tab because it never seems to be used in real text. I also
408 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85),
409 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline
410 * character to. It might be more appropriate to include it in the 8859
411 * set instead of the ASCII set, but it's got to be included in *something*
412 * we recognize or EBCDIC files aren't going to be considered textual.
413 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek
414 * and Latin characters, so these should possibly be allowed. But they
415 * make a real mess on VT100-style displays if they're not paired properly,
416 * so we are probably better off not calling them text.
418 * A file is considered to be ISO-8859 text if its characters are all
419 * either ASCII, according to the above definition, or printing characters
420 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF.
422 * Finally, a file is considered to be international text from some other
423 * character code if its characters are all either ISO-8859 (according to
424 * the above definition) or characters in the range 0x80 ... 0x9F, which
425 * ISO-8859 considers to be control characters but the IBM PC and Macintosh
426 * consider to be printing characters.
429 #define F 0 /* character never appears in text */
430 #define T 1 /* character appears in plain ASCII text */
431 #define I 2 /* character appears in ISO-8859 text */
432 #define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */
434 private char text_chars[256] = {
435 /* BEL BS HT LF FF CR */
436 F, F, F, F, F, F, F, T, T, T, T, F, T, T, F, F, /* 0x0X */
438 F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */
439 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */
440 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */
441 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */
442 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */
443 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */
444 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */
446 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */
447 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */
448 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */
449 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */
450 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */
451 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */
452 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */
453 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */
457 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf,
464 for (i = 0; i < nbytes; i++) {
465 int t = text_chars[buf[i]];
470 ubuf[(*ulen)++] = buf[i];
477 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
483 for (i = 0; i < nbytes; i++) {
484 int t = text_chars[buf[i]];
486 if (t != T && t != I)
489 ubuf[(*ulen)++] = buf[i];
496 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf,
503 for (i = 0; i < nbytes; i++) {
504 int t = text_chars[buf[i]];
506 if (t != T && t != I && t != X)
509 ubuf[(*ulen)++] = buf[i];
516 * Encode Unicode string as UTF-8, returning pointer to character
517 * after end of string, or NULL if an invalid character is found.
519 private unsigned char *
520 encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen)
523 unsigned char *end = buf + len;
525 for (i = 0; i < ulen; i++) {
526 if (ubuf[i] <= 0x7f) {
529 *buf++ = (unsigned char)ubuf[i];
530 } else if (ubuf[i] <= 0x7ff) {
533 *buf++ = (unsigned char)((ubuf[i] >> 6) + 0xc0);
534 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
535 } else if (ubuf[i] <= 0xffff) {
538 *buf++ = (unsigned char)((ubuf[i] >> 12) + 0xe0);
539 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80);
540 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
541 } else if (ubuf[i] <= 0x1fffff) {
544 *buf++ = (unsigned char)((ubuf[i] >> 18) + 0xf0);
545 *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
546 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80);
547 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
548 } else if (ubuf[i] <= 0x3ffffff) {
551 *buf++ = (unsigned char)((ubuf[i] >> 24) + 0xf8);
552 *buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80);
553 *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
554 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80);
555 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
556 } else if (ubuf[i] <= 0x7fffffff) {
559 *buf++ = (unsigned char)((ubuf[i] >> 30) + 0xfc);
560 *buf++ = (unsigned char)(((ubuf[i] >> 24) & 0x3f) + 0x80);
561 *buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80);
562 *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
563 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80);
564 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
565 } else /* Invalid character */
573 * Decide whether some text looks like UTF-8. Returns:
576 * 0: uses odd control characters, so doesn't look like text
578 * 2: definitely UTF-8 text (valid high-bit set bytes)
580 * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen;
581 * ubuf must be big enough!
584 file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
589 int gotone = 0, ctrl = 0;
594 for (i = 0; i < nbytes; i++) {
595 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */
597 * Even if the whole file is valid UTF-8 sequences,
598 * still reject it if it uses weird control characters.
601 if (text_chars[buf[i]] != T)
605 ubuf[(*ulen)++] = buf[i];
606 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */
608 } else { /* 11xxxxxx begins UTF-8 */
611 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */
614 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */
617 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */
620 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */
623 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */
629 for (n = 0; n < following; n++) {
634 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40))
637 c = (c << 6) + (buf[i] & 0x3f);
646 return ctrl ? 0 : (gotone ? 2 : 1);
650 * Decide whether some text looks like UTF-8 with BOM. If there is no
651 * BOM, return -1; otherwise return the result of looks_utf8 on the
655 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf,
658 if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf)
659 return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen);
665 looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf,
674 if (buf[0] == 0xff && buf[1] == 0xfe)
676 else if (buf[0] == 0xfe && buf[1] == 0xff)
683 for (i = 2; i + 1 < nbytes; i += 2) {
684 /* XXX fix to properly handle chars > 65536 */
687 ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i];
689 ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1];
691 if (ubuf[*ulen - 1] == 0xfffe)
693 if (ubuf[*ulen - 1] < 128 &&
694 text_chars[(size_t)ubuf[*ulen - 1]] != T)
707 * This table maps each EBCDIC character to an (8-bit extended) ASCII
708 * character, as specified in the rationale for the dd(1) command in
709 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard.
711 * Unfortunately it does not seem to correspond exactly to any of the
712 * five variants of EBCDIC documented in IBM's _Enterprise Systems
713 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh
714 * Edition, July, 1999, pp. I-1 - I-4.
716 * Fortunately, though, all versions of EBCDIC, including this one, agree
717 * on most of the printing characters that also appear in (7-bit) ASCII.
718 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all.
720 * Fortunately too, there is general agreement that codes 0x00 through
721 * 0x3F represent control characters, 0x41 a nonbreaking space, and the
722 * remainder printing characters.
724 * This is sufficient to allow us to identify EBCDIC text and to distinguish
725 * between old-style and internationalized examples of text.
728 private unsigned char ebcdic_to_ascii[] = {
729 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15,
730 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31,
731 128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7,
732 144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26,
733 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|',
734 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~',
735 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?',
736 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"',
737 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201,
738 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208,
739 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215,
740 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231,
741 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237,
742 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243,
743 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249,
744 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255
749 * The following EBCDIC-to-ASCII table may relate more closely to reality,
750 * or at least to modern reality. It comes from
752 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html
754 * and maps the characters of EBCDIC code page 1047 (the code used for
755 * Unix-derived software on IBM's 390 systems) to the corresponding
756 * characters from ISO 8859-1.
758 * If this table is used instead of the above one, some of the special
759 * cases for the NEL character can be taken out of the code.
762 private unsigned char ebcdic_1047_to_8859[] = {
763 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F,
764 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F,
765 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07,
766 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A,
767 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C,
768 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E,
769 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F,
770 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22,
771 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1,
772 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4,
773 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE,
774 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7,
775 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5,
776 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF,
777 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5,
778 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F
783 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII.
786 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out)
790 for (i = 0; i < nbytes; i++) {
791 out[i] = ebcdic_to_ascii[buf[i]];