xref: /netbsd-src/external/bsd/file/dist/src/encoding.c (revision deb6f0161a9109e7de9b519dc8dfb9478668dcdd)
1 /*	$NetBSD: encoding.c,v 1.7 2018/10/19 00:11:48 christos Exp $	*/
2 
3 /*
4  * Copyright (c) Ian F. Darwin 1986-1995.
5  * Software written by Ian F. Darwin and others;
6  * maintained 1995-present by Christos Zoulas and others.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice immediately at the beginning of the file, without modification,
13  *    this list of conditions, and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  *
18  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21  * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
22  * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28  * SUCH DAMAGE.
29  */
30 /*
31  * Encoding -- determine the character encoding of a text file.
32  *
33  * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit
34  * international characters.
35  */
36 
37 #include "file.h"
38 
39 #ifndef	lint
40 #if 0
41 FILE_RCSID("@(#)$File: encoding.c,v 1.15 2018/10/15 16:29:16 christos Exp $")
42 #else
43 __RCSID("$NetBSD: encoding.c,v 1.7 2018/10/19 00:11:48 christos Exp $");
44 #endif
45 #endif	/* lint */
46 
47 #include "magic.h"
48 #include <string.h>
49 #include <memory.h>
50 #include <stdlib.h>
51 
52 
53 private int looks_ascii(const unsigned char *, size_t, unichar *, size_t *);
54 private int looks_utf8_with_BOM(const unsigned char *, size_t, unichar *,
55     size_t *);
56 private int looks_utf7(const unsigned char *, size_t, unichar *, size_t *);
57 private int looks_ucs16(const unsigned char *, size_t, unichar *, size_t *);
58 private int looks_latin1(const unsigned char *, size_t, unichar *, size_t *);
59 private int looks_extended(const unsigned char *, size_t, unichar *, size_t *);
60 private void from_ebcdic(const unsigned char *, size_t, unsigned char *);
61 
62 #ifdef DEBUG_ENCODING
63 #define DPRINTF(a) printf a
64 #else
65 #define DPRINTF(a)
66 #endif
67 
68 /*
69  * Try to determine whether text is in some character code we can
70  * identify.  Each of these tests, if it succeeds, will leave
71  * the text converted into one-unichar-per-character Unicode in
72  * ubuf, and the number of characters converted in ulen.
73  */
74 protected int
75 file_encoding(struct magic_set *ms, const struct buffer *b, unichar **ubuf,
76     size_t *ulen, const char **code, const char **code_mime, const char **type)
77 {
78 	const unsigned char *buf = CAST(const unsigned char *, b->fbuf);
79 	size_t nbytes = b->flen;
80 	size_t mlen;
81 	int rv = 1, ucs_type;
82 	unsigned char *nbuf = NULL;
83 	unichar *udefbuf;
84 	size_t udeflen;
85 
86 	if (ubuf == NULL)
87 		ubuf = &udefbuf;
88 	if (ulen == NULL)
89 		ulen = &udeflen;
90 
91 	*type = "text";
92 	*ulen = 0;
93 	*code = "unknown";
94 	*code_mime = "binary";
95 
96 	mlen = (nbytes + 1) * sizeof((*ubuf)[0]);
97 	if ((*ubuf = CAST(unichar *, calloc((size_t)1, mlen))) == NULL) {
98 		file_oomem(ms, mlen);
99 		goto done;
100 	}
101 	mlen = (nbytes + 1) * sizeof(nbuf[0]);
102 	if ((nbuf = CAST(unsigned char *, calloc((size_t)1, mlen))) == NULL) {
103 		file_oomem(ms, mlen);
104 		goto done;
105 	}
106 
107 	if (looks_ascii(buf, nbytes, *ubuf, ulen)) {
108 		if (looks_utf7(buf, nbytes, *ubuf, ulen) > 0) {
109 			DPRINTF(("utf-7 %" SIZE_T_FORMAT "u\n", *ulen));
110 			*code = "UTF-7 Unicode";
111 			*code_mime = "utf-7";
112 		} else {
113 			DPRINTF(("ascii %" SIZE_T_FORMAT "u\n", *ulen));
114 			*code = "ASCII";
115 			*code_mime = "us-ascii";
116 		}
117 	} else if (looks_utf8_with_BOM(buf, nbytes, *ubuf, ulen) > 0) {
118 		DPRINTF(("utf8/bom %" SIZE_T_FORMAT "u\n", *ulen));
119 		*code = "UTF-8 Unicode (with BOM)";
120 		*code_mime = "utf-8";
121 	} else if (file_looks_utf8(buf, nbytes, *ubuf, ulen) > 1) {
122 		DPRINTF(("utf8 %" SIZE_T_FORMAT "u\n", *ulen));
123 		*code = "UTF-8 Unicode";
124 		*code_mime = "utf-8";
125 	} else if ((ucs_type = looks_ucs16(buf, nbytes, *ubuf, ulen)) != 0) {
126 		if (ucs_type == 1) {
127 			*code = "Little-endian UTF-16 Unicode";
128 			*code_mime = "utf-16le";
129 		} else {
130 			*code = "Big-endian UTF-16 Unicode";
131 			*code_mime = "utf-16be";
132 		}
133 		DPRINTF(("ucs16 %" SIZE_T_FORMAT "u\n", *ulen));
134 	} else if (looks_latin1(buf, nbytes, *ubuf, ulen)) {
135 		DPRINTF(("latin1 %" SIZE_T_FORMAT "u\n", *ulen));
136 		*code = "ISO-8859";
137 		*code_mime = "iso-8859-1";
138 	} else if (looks_extended(buf, nbytes, *ubuf, ulen)) {
139 		DPRINTF(("extended %" SIZE_T_FORMAT "u\n", *ulen));
140 		*code = "Non-ISO extended-ASCII";
141 		*code_mime = "unknown-8bit";
142 	} else {
143 		from_ebcdic(buf, nbytes, nbuf);
144 
145 		if (looks_ascii(nbuf, nbytes, *ubuf, ulen)) {
146 			DPRINTF(("ebcdic %" SIZE_T_FORMAT "u\n", *ulen));
147 			*code = "EBCDIC";
148 			*code_mime = "ebcdic";
149 		} else if (looks_latin1(nbuf, nbytes, *ubuf, ulen)) {
150 			DPRINTF(("ebcdic/international %" SIZE_T_FORMAT "u\n",
151 			    *ulen));
152 			*code = "International EBCDIC";
153 			*code_mime = "ebcdic";
154 		} else { /* Doesn't look like text at all */
155 			DPRINTF(("binary\n"));
156 			rv = 0;
157 			*type = "binary";
158 		}
159 	}
160 
161  done:
162 	free(nbuf);
163 	if (ubuf == &udefbuf)
164 		free(udefbuf);
165 
166 	return rv;
167 }
168 
169 /*
170  * This table reflects a particular philosophy about what constitutes
171  * "text," and there is room for disagreement about it.
172  *
173  * Version 3.31 of the file command considered a file to be ASCII if
174  * each of its characters was approved by either the isascii() or
175  * isalpha() function.  On most systems, this would mean that any
176  * file consisting only of characters in the range 0x00 ... 0x7F
177  * would be called ASCII text, but many systems might reasonably
178  * consider some characters outside this range to be alphabetic,
179  * so the file command would call such characters ASCII.  It might
180  * have been more accurate to call this "considered textual on the
181  * local system" than "ASCII."
182  *
183  * It considered a file to be "International language text" if each
184  * of its characters was either an ASCII printing character (according
185  * to the real ASCII standard, not the above test), a character in
186  * the range 0x80 ... 0xFF, or one of the following control characters:
187  * backspace, tab, line feed, vertical tab, form feed, carriage return,
188  * escape.  No attempt was made to determine the language in which files
189  * of this type were written.
190  *
191  *
192  * The table below considers a file to be ASCII if all of its characters
193  * are either ASCII printing characters (again, according to the X3.4
194  * standard, not isascii()) or any of the following controls: bell,
195  * backspace, tab, line feed, form feed, carriage return, esc, nextline.
196  *
197  * I include bell because some programs (particularly shell scripts)
198  * use it literally, even though it is rare in normal text.  I exclude
199  * vertical tab because it never seems to be used in real text.  I also
200  * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85),
201  * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline
202  * character to.  It might be more appropriate to include it in the 8859
203  * set instead of the ASCII set, but it's got to be included in *something*
204  * we recognize or EBCDIC files aren't going to be considered textual.
205  * Some old Unix source files use SO/SI (^N/^O) to shift between Greek
206  * and Latin characters, so these should possibly be allowed.  But they
207  * make a real mess on VT100-style displays if they're not paired properly,
208  * so we are probably better off not calling them text.
209  *
210  * A file is considered to be ISO-8859 text if its characters are all
211  * either ASCII, according to the above definition, or printing characters
212  * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF.
213  *
214  * Finally, a file is considered to be international text from some other
215  * character code if its characters are all either ISO-8859 (according to
216  * the above definition) or characters in the range 0x80 ... 0x9F, which
217  * ISO-8859 considers to be control characters but the IBM PC and Macintosh
218  * consider to be printing characters.
219  */
220 
221 #define F 0   /* character never appears in text */
222 #define T 1   /* character appears in plain ASCII text */
223 #define I 2   /* character appears in ISO-8859 text */
224 #define X 3   /* character appears in non-ISO extended ASCII (Mac, IBM PC) */
225 
226 private char text_chars[256] = {
227 	/*                  BEL BS HT LF VT FF CR    */
228 	F, F, F, F, F, F, F, T, T, T, T, T, T, T, F, F,  /* 0x0X */
229 	/*                              ESC          */
230 	F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F,  /* 0x1X */
231 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x2X */
232 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x3X */
233 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x4X */
234 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x5X */
235 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x6X */
236 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F,  /* 0x7X */
237 	/*            NEL                            */
238 	X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X,  /* 0x8X */
239 	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,  /* 0x9X */
240 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xaX */
241 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xbX */
242 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xcX */
243 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xdX */
244 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xeX */
245 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I   /* 0xfX */
246 };
247 
248 private int
249 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf,
250     size_t *ulen)
251 {
252 	size_t i;
253 
254 	*ulen = 0;
255 
256 	for (i = 0; i < nbytes; i++) {
257 		int t = text_chars[buf[i]];
258 
259 		if (t != T)
260 			return 0;
261 
262 		ubuf[(*ulen)++] = buf[i];
263 	}
264 
265 	return 1;
266 }
267 
268 private int
269 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
270 {
271 	size_t i;
272 
273 	*ulen = 0;
274 
275 	for (i = 0; i < nbytes; i++) {
276 		int t = text_chars[buf[i]];
277 
278 		if (t != T && t != I)
279 			return 0;
280 
281 		ubuf[(*ulen)++] = buf[i];
282 	}
283 
284 	return 1;
285 }
286 
287 private int
288 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf,
289     size_t *ulen)
290 {
291 	size_t i;
292 
293 	*ulen = 0;
294 
295 	for (i = 0; i < nbytes; i++) {
296 		int t = text_chars[buf[i]];
297 
298 		if (t != T && t != I && t != X)
299 			return 0;
300 
301 		ubuf[(*ulen)++] = buf[i];
302 	}
303 
304 	return 1;
305 }
306 
307 /*
308  * Decide whether some text looks like UTF-8. Returns:
309  *
310  *     -1: invalid UTF-8
311  *      0: uses odd control characters, so doesn't look like text
312  *      1: 7-bit text
313  *      2: definitely UTF-8 text (valid high-bit set bytes)
314  *
315  * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen;
316  * ubuf must be big enough!
317  */
318 protected int
319 file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
320 {
321 	size_t i;
322 	int n;
323 	unichar c;
324 	int gotone = 0, ctrl = 0;
325 
326 	if (ubuf)
327 		*ulen = 0;
328 
329 	for (i = 0; i < nbytes; i++) {
330 		if ((buf[i] & 0x80) == 0) {	   /* 0xxxxxxx is plain ASCII */
331 			/*
332 			 * Even if the whole file is valid UTF-8 sequences,
333 			 * still reject it if it uses weird control characters.
334 			 */
335 
336 			if (text_chars[buf[i]] != T)
337 				ctrl = 1;
338 
339 			if (ubuf)
340 				ubuf[(*ulen)++] = buf[i];
341 		} else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */
342 			return -1;
343 		} else {			   /* 11xxxxxx begins UTF-8 */
344 			int following;
345 
346 			if ((buf[i] & 0x20) == 0) {		/* 110xxxxx */
347 				c = buf[i] & 0x1f;
348 				following = 1;
349 			} else if ((buf[i] & 0x10) == 0) {	/* 1110xxxx */
350 				c = buf[i] & 0x0f;
351 				following = 2;
352 			} else if ((buf[i] & 0x08) == 0) {	/* 11110xxx */
353 				c = buf[i] & 0x07;
354 				following = 3;
355 			} else if ((buf[i] & 0x04) == 0) {	/* 111110xx */
356 				c = buf[i] & 0x03;
357 				following = 4;
358 			} else if ((buf[i] & 0x02) == 0) {	/* 1111110x */
359 				c = buf[i] & 0x01;
360 				following = 5;
361 			} else
362 				return -1;
363 
364 			for (n = 0; n < following; n++) {
365 				i++;
366 				if (i >= nbytes)
367 					goto done;
368 
369 				if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40))
370 					return -1;
371 
372 				c = (c << 6) + (buf[i] & 0x3f);
373 			}
374 
375 			if (ubuf)
376 				ubuf[(*ulen)++] = c;
377 			gotone = 1;
378 		}
379 	}
380 done:
381 	return ctrl ? 0 : (gotone ? 2 : 1);
382 }
383 
384 /*
385  * Decide whether some text looks like UTF-8 with BOM. If there is no
386  * BOM, return -1; otherwise return the result of looks_utf8 on the
387  * rest of the text.
388  */
389 private int
390 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf,
391     size_t *ulen)
392 {
393 	if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf)
394 		return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen);
395 	else
396 		return -1;
397 }
398 
399 private int
400 looks_utf7(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
401 {
402 	if (nbytes > 4 && buf[0] == '+' && buf[1] == '/' && buf[2] == 'v')
403 		switch (buf[3]) {
404 		case '8':
405 		case '9':
406 		case '+':
407 		case '/':
408 			if (ubuf)
409 				*ulen = 0;
410 			return 1;
411 		default:
412 			return -1;
413 		}
414 	else
415 		return -1;
416 }
417 
418 private int
419 looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf,
420     size_t *ulen)
421 {
422 	int bigend;
423 	size_t i;
424 
425 	if (nbytes < 2)
426 		return 0;
427 
428 	if (buf[0] == 0xff && buf[1] == 0xfe)
429 		bigend = 0;
430 	else if (buf[0] == 0xfe && buf[1] == 0xff)
431 		bigend = 1;
432 	else
433 		return 0;
434 
435 	*ulen = 0;
436 
437 	for (i = 2; i + 1 < nbytes; i += 2) {
438 		/* XXX fix to properly handle chars > 65536 */
439 
440 		if (bigend)
441 			ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i];
442 		else
443 			ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1];
444 
445 		if (ubuf[*ulen - 1] == 0xfffe)
446 			return 0;
447 		if (ubuf[*ulen - 1] < 128 &&
448 		    text_chars[(size_t)ubuf[*ulen - 1]] != T)
449 			return 0;
450 	}
451 
452 	return 1 + bigend;
453 }
454 
455 #undef F
456 #undef T
457 #undef I
458 #undef X
459 
460 /*
461  * This table maps each EBCDIC character to an (8-bit extended) ASCII
462  * character, as specified in the rationale for the dd(1) command in
463  * draft 11.2 (September, 1991) of the POSIX P1003.2 standard.
464  *
465  * Unfortunately it does not seem to correspond exactly to any of the
466  * five variants of EBCDIC documented in IBM's _Enterprise Systems
467  * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh
468  * Edition, July, 1999, pp. I-1 - I-4.
469  *
470  * Fortunately, though, all versions of EBCDIC, including this one, agree
471  * on most of the printing characters that also appear in (7-bit) ASCII.
472  * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all.
473  *
474  * Fortunately too, there is general agreement that codes 0x00 through
475  * 0x3F represent control characters, 0x41 a nonbreaking space, and the
476  * remainder printing characters.
477  *
478  * This is sufficient to allow us to identify EBCDIC text and to distinguish
479  * between old-style and internationalized examples of text.
480  */
481 
482 private unsigned char ebcdic_to_ascii[] = {
483   0,   1,   2,   3, 156,   9, 134, 127, 151, 141, 142,  11,  12,  13,  14,  15,
484  16,  17,  18,  19, 157, 133,   8, 135,  24,  25, 146, 143,  28,  29,  30,  31,
485 128, 129, 130, 131, 132,  10,  23,  27, 136, 137, 138, 139, 140,   5,   6,   7,
486 144, 145,  22, 147, 148, 149, 150,   4, 152, 153, 154, 155,  20,  21, 158,  26,
487 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|',
488 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~',
489 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?',
490 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"',
491 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201,
492 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208,
493 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215,
494 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231,
495 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237,
496 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243,
497 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249,
498 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255
499 };
500 
501 #ifdef notdef
502 /*
503  * The following EBCDIC-to-ASCII table may relate more closely to reality,
504  * or at least to modern reality.  It comes from
505  *
506  *   http://ftp.s390.ibm.com/products/oe/bpxqp9.html
507  *
508  * and maps the characters of EBCDIC code page 1047 (the code used for
509  * Unix-derived software on IBM's 390 systems) to the corresponding
510  * characters from ISO 8859-1.
511  *
512  * If this table is used instead of the above one, some of the special
513  * cases for the NEL character can be taken out of the code.
514  */
515 
516 private unsigned char ebcdic_1047_to_8859[] = {
517 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F,
518 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F,
519 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07,
520 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A,
521 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C,
522 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E,
523 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F,
524 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22,
525 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1,
526 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4,
527 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE,
528 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7,
529 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5,
530 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF,
531 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5,
532 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F
533 };
534 #endif
535 
536 /*
537  * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII.
538  */
539 private void
540 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out)
541 {
542 	size_t i;
543 
544 	for (i = 0; i < nbytes; i++) {
545 		out[i] = ebcdic_to_ascii[buf[i]];
546 	}
547 }
548