xref: /netbsd-src/external/bsd/file/dist/src/encoding.c (revision 03dcb730d46d34d85c9f496c1f5a3a6a43f2b7b3)
1 /*	$NetBSD: encoding.c,v 1.5 2017/02/10 17:53:24 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 1.13 2015/06/04 19:16:28 christos Exp $")
42 #else
43 __RCSID("$NetBSD: encoding.c,v 1.5 2017/02/10 17:53:24 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 unsigned char *buf, size_t nbytes, unichar **ubuf, size_t *ulen, const char **code, const char **code_mime, const char **type)
76 {
77 	size_t mlen;
78 	int rv = 1, ucs_type;
79 	unsigned char *nbuf = NULL;
80 
81 	*type = "text";
82 	*ulen = 0;
83 	*code = "unknown";
84 	*code_mime = "binary";
85 
86 	mlen = (nbytes + 1) * sizeof((*ubuf)[0]);
87 	if ((*ubuf = CAST(unichar *, calloc((size_t)1, mlen))) == NULL) {
88 		file_oomem(ms, mlen);
89 		goto done;
90 	}
91 	mlen = (nbytes + 1) * sizeof(nbuf[0]);
92 	if ((nbuf = CAST(unsigned char *, calloc((size_t)1, mlen))) == NULL) {
93 		file_oomem(ms, mlen);
94 		goto done;
95 	}
96 
97 	if (looks_ascii(buf, nbytes, *ubuf, ulen)) {
98 		if (looks_utf7(buf, nbytes, *ubuf, ulen) > 0) {
99 			DPRINTF(("utf-7 %" SIZE_T_FORMAT "u\n", *ulen));
100 			*code = "UTF-7 Unicode";
101 			*code_mime = "utf-7";
102 		} else {
103 			DPRINTF(("ascii %" SIZE_T_FORMAT "u\n", *ulen));
104 			*code = "ASCII";
105 			*code_mime = "us-ascii";
106 		}
107 	} else if (looks_utf8_with_BOM(buf, nbytes, *ubuf, ulen) > 0) {
108 		DPRINTF(("utf8/bom %" SIZE_T_FORMAT "u\n", *ulen));
109 		*code = "UTF-8 Unicode (with BOM)";
110 		*code_mime = "utf-8";
111 	} else if (file_looks_utf8(buf, nbytes, *ubuf, ulen) > 1) {
112 		DPRINTF(("utf8 %" SIZE_T_FORMAT "u\n", *ulen));
113 		*code = "UTF-8 Unicode";
114 		*code_mime = "utf-8";
115 	} else if ((ucs_type = looks_ucs16(buf, nbytes, *ubuf, ulen)) != 0) {
116 		if (ucs_type == 1) {
117 			*code = "Little-endian UTF-16 Unicode";
118 			*code_mime = "utf-16le";
119 		} else {
120 			*code = "Big-endian UTF-16 Unicode";
121 			*code_mime = "utf-16be";
122 		}
123 		DPRINTF(("ucs16 %" SIZE_T_FORMAT "u\n", *ulen));
124 	} else if (looks_latin1(buf, nbytes, *ubuf, ulen)) {
125 		DPRINTF(("latin1 %" SIZE_T_FORMAT "u\n", *ulen));
126 		*code = "ISO-8859";
127 		*code_mime = "iso-8859-1";
128 	} else if (looks_extended(buf, nbytes, *ubuf, ulen)) {
129 		DPRINTF(("extended %" SIZE_T_FORMAT "u\n", *ulen));
130 		*code = "Non-ISO extended-ASCII";
131 		*code_mime = "unknown-8bit";
132 	} else {
133 		from_ebcdic(buf, nbytes, nbuf);
134 
135 		if (looks_ascii(nbuf, nbytes, *ubuf, ulen)) {
136 			DPRINTF(("ebcdic %" SIZE_T_FORMAT "u\n", *ulen));
137 			*code = "EBCDIC";
138 			*code_mime = "ebcdic";
139 		} else if (looks_latin1(nbuf, nbytes, *ubuf, ulen)) {
140 			DPRINTF(("ebcdic/international %" SIZE_T_FORMAT "u\n",
141 			    *ulen));
142 			*code = "International EBCDIC";
143 			*code_mime = "ebcdic";
144 		} else { /* Doesn't look like text at all */
145 			DPRINTF(("binary\n"));
146 			rv = 0;
147 			*type = "binary";
148 		}
149 	}
150 
151  done:
152 	free(nbuf);
153 
154 	return rv;
155 }
156 
157 /*
158  * This table reflects a particular philosophy about what constitutes
159  * "text," and there is room for disagreement about it.
160  *
161  * Version 3.31 of the file command considered a file to be ASCII if
162  * each of its characters was approved by either the isascii() or
163  * isalpha() function.  On most systems, this would mean that any
164  * file consisting only of characters in the range 0x00 ... 0x7F
165  * would be called ASCII text, but many systems might reasonably
166  * consider some characters outside this range to be alphabetic,
167  * so the file command would call such characters ASCII.  It might
168  * have been more accurate to call this "considered textual on the
169  * local system" than "ASCII."
170  *
171  * It considered a file to be "International language text" if each
172  * of its characters was either an ASCII printing character (according
173  * to the real ASCII standard, not the above test), a character in
174  * the range 0x80 ... 0xFF, or one of the following control characters:
175  * backspace, tab, line feed, vertical tab, form feed, carriage return,
176  * escape.  No attempt was made to determine the language in which files
177  * of this type were written.
178  *
179  *
180  * The table below considers a file to be ASCII if all of its characters
181  * are either ASCII printing characters (again, according to the X3.4
182  * standard, not isascii()) or any of the following controls: bell,
183  * backspace, tab, line feed, form feed, carriage return, esc, nextline.
184  *
185  * I include bell because some programs (particularly shell scripts)
186  * use it literally, even though it is rare in normal text.  I exclude
187  * vertical tab because it never seems to be used in real text.  I also
188  * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85),
189  * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline
190  * character to.  It might be more appropriate to include it in the 8859
191  * set instead of the ASCII set, but it's got to be included in *something*
192  * we recognize or EBCDIC files aren't going to be considered textual.
193  * Some old Unix source files use SO/SI (^N/^O) to shift between Greek
194  * and Latin characters, so these should possibly be allowed.  But they
195  * make a real mess on VT100-style displays if they're not paired properly,
196  * so we are probably better off not calling them text.
197  *
198  * A file is considered to be ISO-8859 text if its characters are all
199  * either ASCII, according to the above definition, or printing characters
200  * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF.
201  *
202  * Finally, a file is considered to be international text from some other
203  * character code if its characters are all either ISO-8859 (according to
204  * the above definition) or characters in the range 0x80 ... 0x9F, which
205  * ISO-8859 considers to be control characters but the IBM PC and Macintosh
206  * consider to be printing characters.
207  */
208 
209 #define F 0   /* character never appears in text */
210 #define T 1   /* character appears in plain ASCII text */
211 #define I 2   /* character appears in ISO-8859 text */
212 #define X 3   /* character appears in non-ISO extended ASCII (Mac, IBM PC) */
213 
214 private char text_chars[256] = {
215 	/*                  BEL BS HT LF VT FF CR    */
216 	F, F, F, F, F, F, F, T, T, T, T, T, T, T, F, F,  /* 0x0X */
217 	/*                              ESC          */
218 	F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F,  /* 0x1X */
219 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x2X */
220 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x3X */
221 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x4X */
222 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x5X */
223 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,  /* 0x6X */
224 	T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F,  /* 0x7X */
225 	/*            NEL                            */
226 	X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X,  /* 0x8X */
227 	X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X,  /* 0x9X */
228 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xaX */
229 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xbX */
230 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xcX */
231 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xdX */
232 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I,  /* 0xeX */
233 	I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I   /* 0xfX */
234 };
235 
236 private int
237 looks_ascii(const unsigned char *buf, size_t nbytes, unichar *ubuf,
238     size_t *ulen)
239 {
240 	size_t i;
241 
242 	*ulen = 0;
243 
244 	for (i = 0; i < nbytes; i++) {
245 		int t = text_chars[buf[i]];
246 
247 		if (t != T)
248 			return 0;
249 
250 		ubuf[(*ulen)++] = buf[i];
251 	}
252 
253 	return 1;
254 }
255 
256 private int
257 looks_latin1(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
258 {
259 	size_t i;
260 
261 	*ulen = 0;
262 
263 	for (i = 0; i < nbytes; i++) {
264 		int t = text_chars[buf[i]];
265 
266 		if (t != T && t != I)
267 			return 0;
268 
269 		ubuf[(*ulen)++] = buf[i];
270 	}
271 
272 	return 1;
273 }
274 
275 private int
276 looks_extended(const unsigned char *buf, size_t nbytes, unichar *ubuf,
277     size_t *ulen)
278 {
279 	size_t i;
280 
281 	*ulen = 0;
282 
283 	for (i = 0; i < nbytes; i++) {
284 		int t = text_chars[buf[i]];
285 
286 		if (t != T && t != I && t != X)
287 			return 0;
288 
289 		ubuf[(*ulen)++] = buf[i];
290 	}
291 
292 	return 1;
293 }
294 
295 /*
296  * Decide whether some text looks like UTF-8. Returns:
297  *
298  *     -1: invalid UTF-8
299  *      0: uses odd control characters, so doesn't look like text
300  *      1: 7-bit text
301  *      2: definitely UTF-8 text (valid high-bit set bytes)
302  *
303  * If ubuf is non-NULL on entry, text is decoded into ubuf, *ulen;
304  * ubuf must be big enough!
305  */
306 protected int
307 file_looks_utf8(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
308 {
309 	size_t i;
310 	int n;
311 	unichar c;
312 	int gotone = 0, ctrl = 0;
313 
314 	if (ubuf)
315 		*ulen = 0;
316 
317 	for (i = 0; i < nbytes; i++) {
318 		if ((buf[i] & 0x80) == 0) {	   /* 0xxxxxxx is plain ASCII */
319 			/*
320 			 * Even if the whole file is valid UTF-8 sequences,
321 			 * still reject it if it uses weird control characters.
322 			 */
323 
324 			if (text_chars[buf[i]] != T)
325 				ctrl = 1;
326 
327 			if (ubuf)
328 				ubuf[(*ulen)++] = buf[i];
329 		} else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */
330 			return -1;
331 		} else {			   /* 11xxxxxx begins UTF-8 */
332 			int following;
333 
334 			if ((buf[i] & 0x20) == 0) {		/* 110xxxxx */
335 				c = buf[i] & 0x1f;
336 				following = 1;
337 			} else if ((buf[i] & 0x10) == 0) {	/* 1110xxxx */
338 				c = buf[i] & 0x0f;
339 				following = 2;
340 			} else if ((buf[i] & 0x08) == 0) {	/* 11110xxx */
341 				c = buf[i] & 0x07;
342 				following = 3;
343 			} else if ((buf[i] & 0x04) == 0) {	/* 111110xx */
344 				c = buf[i] & 0x03;
345 				following = 4;
346 			} else if ((buf[i] & 0x02) == 0) {	/* 1111110x */
347 				c = buf[i] & 0x01;
348 				following = 5;
349 			} else
350 				return -1;
351 
352 			for (n = 0; n < following; n++) {
353 				i++;
354 				if (i >= nbytes)
355 					goto done;
356 
357 				if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40))
358 					return -1;
359 
360 				c = (c << 6) + (buf[i] & 0x3f);
361 			}
362 
363 			if (ubuf)
364 				ubuf[(*ulen)++] = c;
365 			gotone = 1;
366 		}
367 	}
368 done:
369 	return ctrl ? 0 : (gotone ? 2 : 1);
370 }
371 
372 /*
373  * Decide whether some text looks like UTF-8 with BOM. If there is no
374  * BOM, return -1; otherwise return the result of looks_utf8 on the
375  * rest of the text.
376  */
377 private int
378 looks_utf8_with_BOM(const unsigned char *buf, size_t nbytes, unichar *ubuf,
379     size_t *ulen)
380 {
381 	if (nbytes > 3 && buf[0] == 0xef && buf[1] == 0xbb && buf[2] == 0xbf)
382 		return file_looks_utf8(buf + 3, nbytes - 3, ubuf, ulen);
383 	else
384 		return -1;
385 }
386 
387 private int
388 looks_utf7(const unsigned char *buf, size_t nbytes, unichar *ubuf, size_t *ulen)
389 {
390 	if (nbytes > 4 && buf[0] == '+' && buf[1] == '/' && buf[2] == 'v')
391 		switch (buf[3]) {
392 		case '8':
393 		case '9':
394 		case '+':
395 		case '/':
396 			if (ubuf)
397 				*ulen = 0;
398 			return 1;
399 		default:
400 			return -1;
401 		}
402 	else
403 		return -1;
404 }
405 
406 private int
407 looks_ucs16(const unsigned char *buf, size_t nbytes, unichar *ubuf,
408     size_t *ulen)
409 {
410 	int bigend;
411 	size_t i;
412 
413 	if (nbytes < 2)
414 		return 0;
415 
416 	if (buf[0] == 0xff && buf[1] == 0xfe)
417 		bigend = 0;
418 	else if (buf[0] == 0xfe && buf[1] == 0xff)
419 		bigend = 1;
420 	else
421 		return 0;
422 
423 	*ulen = 0;
424 
425 	for (i = 2; i + 1 < nbytes; i += 2) {
426 		/* XXX fix to properly handle chars > 65536 */
427 
428 		if (bigend)
429 			ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i];
430 		else
431 			ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1];
432 
433 		if (ubuf[*ulen - 1] == 0xfffe)
434 			return 0;
435 		if (ubuf[*ulen - 1] < 128 &&
436 		    text_chars[(size_t)ubuf[*ulen - 1]] != T)
437 			return 0;
438 	}
439 
440 	return 1 + bigend;
441 }
442 
443 #undef F
444 #undef T
445 #undef I
446 #undef X
447 
448 /*
449  * This table maps each EBCDIC character to an (8-bit extended) ASCII
450  * character, as specified in the rationale for the dd(1) command in
451  * draft 11.2 (September, 1991) of the POSIX P1003.2 standard.
452  *
453  * Unfortunately it does not seem to correspond exactly to any of the
454  * five variants of EBCDIC documented in IBM's _Enterprise Systems
455  * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh
456  * Edition, July, 1999, pp. I-1 - I-4.
457  *
458  * Fortunately, though, all versions of EBCDIC, including this one, agree
459  * on most of the printing characters that also appear in (7-bit) ASCII.
460  * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all.
461  *
462  * Fortunately too, there is general agreement that codes 0x00 through
463  * 0x3F represent control characters, 0x41 a nonbreaking space, and the
464  * remainder printing characters.
465  *
466  * This is sufficient to allow us to identify EBCDIC text and to distinguish
467  * between old-style and internationalized examples of text.
468  */
469 
470 private unsigned char ebcdic_to_ascii[] = {
471   0,   1,   2,   3, 156,   9, 134, 127, 151, 141, 142,  11,  12,  13,  14,  15,
472  16,  17,  18,  19, 157, 133,   8, 135,  24,  25, 146, 143,  28,  29,  30,  31,
473 128, 129, 130, 131, 132,  10,  23,  27, 136, 137, 138, 139, 140,   5,   6,   7,
474 144, 145,  22, 147, 148, 149, 150,   4, 152, 153, 154, 155,  20,  21, 158,  26,
475 ' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|',
476 '&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~',
477 '-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?',
478 186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"',
479 195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201,
480 202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208,
481 209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215,
482 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231,
483 '{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237,
484 '}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243,
485 '\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249,
486 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255
487 };
488 
489 #ifdef notdef
490 /*
491  * The following EBCDIC-to-ASCII table may relate more closely to reality,
492  * or at least to modern reality.  It comes from
493  *
494  *   http://ftp.s390.ibm.com/products/oe/bpxqp9.html
495  *
496  * and maps the characters of EBCDIC code page 1047 (the code used for
497  * Unix-derived software on IBM's 390 systems) to the corresponding
498  * characters from ISO 8859-1.
499  *
500  * If this table is used instead of the above one, some of the special
501  * cases for the NEL character can be taken out of the code.
502  */
503 
504 private unsigned char ebcdic_1047_to_8859[] = {
505 0x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F,
506 0x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F,
507 0x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07,
508 0x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A,
509 0x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C,
510 0x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E,
511 0x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F,
512 0xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22,
513 0xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1,
514 0xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4,
515 0xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE,
516 0xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7,
517 0x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5,
518 0x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF,
519 0x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5,
520 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F
521 };
522 #endif
523 
524 /*
525  * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII.
526  */
527 private void
528 from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out)
529 {
530 	size_t i;
531 
532 	for (i = 0; i < nbytes; i++) {
533 		out[i] = ebcdic_to_ascii[buf[i]];
534 	}
535 }
536