1/* 2 * Copyright 2006, Ingo Weinhold <bonefish@cs.tu-berlin.de>. 3 * All rights reserved. Distributed under the terms of the MIT License. 4 */ 5 6#include "TextSnifferAddon.h" 7 8#include <MimeType.h> 9 10 11static int file_ascmagic(const unsigned char *buf, size_t nbytes, 12 BMimeType* mimeType); 13 14 15// constructor 16TextSnifferAddon::TextSnifferAddon() 17{ 18} 19 20// destructor 21TextSnifferAddon::~TextSnifferAddon() 22{ 23} 24 25// MinimalBufferSize 26size_t 27TextSnifferAddon::MinimalBufferSize() 28{ 29 return 512; 30} 31 32// GuessMimeType 33float 34TextSnifferAddon::GuessMimeType(const char* fileName, BMimeType* type) 35{ 36 // we check content only 37 return -1; 38} 39 40// GuessMimeType 41float 42TextSnifferAddon::GuessMimeType(BFile* file, const void* buffer, int32 length, 43 BMimeType* type) 44{ 45 if (file_ascmagic((const unsigned char*)buffer, length, type)) { 46 // If the buffer is very short, we return a lower priority. Maybe 47 // someone else knows better. 48 if (length < 20) 49 return .0f; 50 return 0.25f; 51 } 52 53 return -1; 54} 55 56 57// #pragma mark - ascmagic.c from the BSD file tool 58/* 59 * The following code has been taken from version 4.17 of the BSD file tool, 60 * file ascmagic.c, modified for our purpose. 61 */ 62 63/* 64 * Copyright (c) Ian F. Darwin 1986-1995. 65 * Software written by Ian F. Darwin and others; 66 * maintained 1995-present by Christos Zoulas and others. 67 * 68 * Redistribution and use in source and binary forms, with or without 69 * modification, are permitted provided that the following conditions 70 * are met: 71 * 1. Redistributions of source code must retain the above copyright 72 * notice immediately at the beginning of the file, without modification, 73 * this list of conditions, and the following disclaimer. 74 * 2. Redistributions in binary form must reproduce the above copyright 75 * notice, this list of conditions and the following disclaimer in the 76 * documentation and/or other materials provided with the distribution. 77 * 78 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 79 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 80 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 81 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR 82 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 83 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 84 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 85 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 86 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 87 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 88 * SUCH DAMAGE. 89 */ 90/* 91 * ASCII magic -- file types that we know based on keywords 92 * that can appear anywhere in the file. 93 * 94 * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000, 95 * to handle character codes other than ASCII on a unified basis. 96 * 97 * Joerg Wunsch <joerg@freebsd.org> wrote the original support for 8-bit 98 * international characters, now subsumed into this file. 99 */ 100 101#include <stdio.h> 102#include <string.h> 103#include <memory.h> 104#include <ctype.h> 105#include <stdlib.h> 106#include <unistd.h> 107#include "names.h" 108 109typedef unsigned long my_unichar; 110 111#define MAXLINELEN 300 /* longest sane line length */ 112#define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \ 113 || (x) == 0x85 || (x) == '\f') 114 115static int looks_ascii(const unsigned char *, size_t, my_unichar *, size_t *); 116static int looks_utf8(const unsigned char *, size_t, my_unichar *, size_t *); 117static int looks_unicode(const unsigned char *, size_t, my_unichar *, size_t *); 118static int looks_latin1(const unsigned char *, size_t, my_unichar *, size_t *); 119static int looks_extended(const unsigned char *, size_t, my_unichar *, size_t *); 120static void from_ebcdic(const unsigned char *, size_t, unsigned char *); 121static int ascmatch(const unsigned char *, const my_unichar *, size_t); 122 123 124static int 125file_ascmagic(const unsigned char *buf, size_t nbytes, BMimeType* mimeType) 126{ 127 size_t i; 128 unsigned char *nbuf = NULL; 129 my_unichar *ubuf = NULL; 130 size_t ulen; 131 struct names *p; 132 int rv = -1; 133 134 const char *code = NULL; 135 const char *code_mime = NULL; 136 const char *type = NULL; 137 const char *subtype = NULL; 138 const char *subtypeMimeGeneric = NULL; 139 const char *subtypeMimeSpecific = NULL; 140 141 int has_escapes = 0; 142 int has_backspace = 0; 143 int seen_cr = 0; 144 145 int n_crlf = 0; 146 int n_lf = 0; 147 int n_cr = 0; 148 int n_nel = 0; 149 150 int last_line_end = -1; 151 int has_long_lines = 0; 152 153 if ((nbuf = (unsigned char*)malloc((nbytes + 1) * sizeof(nbuf[0]))) == NULL) 154 goto done; 155 if ((ubuf = (my_unichar*)malloc((nbytes + 1) * sizeof(ubuf[0]))) == NULL) 156 goto done; 157 158 /* 159 * Then try to determine whether it's any character code we can 160 * identify. Each of these tests, if it succeeds, will leave 161 * the text converted into one-my_unichar-per-character Unicode in 162 * ubuf, and the number of characters converted in ulen. 163 */ 164 if (looks_ascii(buf, nbytes, ubuf, &ulen)) { 165 code = "ASCII"; 166 code_mime = "us-ascii"; 167 type = "text"; 168 } else if (looks_utf8(buf, nbytes, ubuf, &ulen)) { 169 code = "UTF-8 Unicode"; 170 code_mime = "utf-8"; 171 type = "text"; 172 } else if ((i = looks_unicode(buf, nbytes, ubuf, &ulen)) != 0) { 173 if (i == 1) 174 code = "Little-endian UTF-16 Unicode"; 175 else 176 code = "Big-endian UTF-16 Unicode"; 177 178 type = "character data"; 179 code_mime = "utf-16"; /* is this defined? */ 180 } else if (looks_latin1(buf, nbytes, ubuf, &ulen)) { 181 code = "ISO-8859"; 182 type = "text"; 183 code_mime = "iso-8859-1"; 184 } else if (looks_extended(buf, nbytes, ubuf, &ulen)) { 185 code = "Non-ISO extended-ASCII"; 186 type = "text"; 187 code_mime = "unknown"; 188 } else { 189 from_ebcdic(buf, nbytes, nbuf); 190 191 if (looks_ascii(nbuf, nbytes, ubuf, &ulen)) { 192 code = "EBCDIC"; 193 type = "character data"; 194 code_mime = "ebcdic"; 195 } else if (looks_latin1(nbuf, nbytes, ubuf, &ulen)) { 196 code = "International EBCDIC"; 197 type = "character data"; 198 code_mime = "ebcdic"; 199 } else { 200 rv = 0; 201 goto done; /* doesn't look like text at all */ 202 } 203 } 204 205 if (nbytes <= 1) { 206 rv = 0; 207 goto done; 208 } 209 210 /* 211 * for troff, look for . + letter + letter or .\"; 212 * this must be done to disambiguate tar archives' ./file 213 * and other trash from real troff input. 214 * 215 * I believe Plan 9 troff allows non-ASCII characters in the names 216 * of macros, so this test might possibly fail on such a file. 217 */ 218 if (*ubuf == '.') { 219 my_unichar *tp = ubuf + 1; 220 221 while (ISSPC(*tp)) 222 ++tp; /* skip leading whitespace */ 223 if ((tp[0] == '\\' && tp[1] == '\"') || 224 (isascii((unsigned char)tp[0]) && 225 isalnum((unsigned char)tp[0]) && 226 isascii((unsigned char)tp[1]) && 227 isalnum((unsigned char)tp[1]) && 228 ISSPC(tp[2]))) { 229 subtypeMimeGeneric = "text/x-source-code"; 230 subtypeMimeSpecific = "text/troff"; 231 subtype = "troff or preprocessor input"; 232 goto subtype_identified; 233 } 234 } 235 236 if ((*buf == 'c' || *buf == 'C') && ISSPC(buf[1])) { 237 subtypeMimeGeneric = "text/x-source-code"; 238 subtypeMimeSpecific = "text/fortran"; 239 subtype = "fortran program"; 240 goto subtype_identified; 241 } 242 243 /* look for tokens from names.h - this is expensive! */ 244 245 i = 0; 246 while (i < ulen) { 247 size_t end; 248 249 /* 250 * skip past any leading space 251 */ 252 while (i < ulen && ISSPC(ubuf[i])) 253 i++; 254 if (i >= ulen) 255 break; 256 257 /* 258 * find the next whitespace 259 */ 260 for (end = i + 1; end < nbytes; end++) 261 if (ISSPC(ubuf[end])) 262 break; 263 264 /* 265 * compare the word thus isolated against the token list 266 */ 267 for (p = names; p < names + NNAMES; p++) { 268 if (ascmatch((const unsigned char *)p->name, ubuf + i, 269 end - i)) { 270 subtype = types[p->type].human; 271 subtypeMimeGeneric = types[p->type].generic_mime; 272 subtypeMimeSpecific = types[p->type].specific_mime; 273 goto subtype_identified; 274 } 275 } 276 277 i = end; 278 } 279 280subtype_identified: 281 282 /* 283 * Now try to discover other details about the file. 284 */ 285 for (i = 0; i < ulen; i++) { 286 if (ubuf[i] == '\n') { 287 if (seen_cr) 288 n_crlf++; 289 else 290 n_lf++; 291 last_line_end = i; 292 } else if (seen_cr) 293 n_cr++; 294 295 seen_cr = (ubuf[i] == '\r'); 296 if (seen_cr) 297 last_line_end = i; 298 299 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */ 300 n_nel++; 301 last_line_end = i; 302 } 303 304 /* If this line is _longer_ than MAXLINELEN, remember it. */ 305 if ((int)i > last_line_end + MAXLINELEN) 306 has_long_lines = 1; 307 308 if (ubuf[i] == '\033') 309 has_escapes = 1; 310 if (ubuf[i] == '\b') 311 has_backspace = 1; 312 } 313 314 rv = 1; 315done: 316 if (nbuf) 317 free(nbuf); 318 if (ubuf) 319 free(ubuf); 320 321 if (rv) { 322 // If we have identified the subtype, return it, otherwise just 323 // text/plain. 324 325 bool found = false; 326 if (subtypeMimeSpecific != NULL) { 327 mimeType->SetTo(subtypeMimeSpecific); 328 if (mimeType->IsInstalled()) 329 found = true; 330 } 331 if (!found && subtypeMimeGeneric != NULL) { 332 mimeType->SetTo(subtypeMimeGeneric); 333 if (mimeType->IsInstalled()) 334 found = true; 335 } 336 if (!found) 337 mimeType->SetTo("text/plain"); 338 } 339 340 return rv; 341} 342 343static int 344ascmatch(const unsigned char *s, const my_unichar *us, size_t ulen) 345{ 346 size_t i; 347 348 for (i = 0; i < ulen; i++) { 349 if (s[i] != us[i]) 350 return 0; 351 } 352 353 if (s[i]) 354 return 0; 355 else 356 return 1; 357} 358 359/* 360 * This table reflects a particular philosophy about what constitutes 361 * "text," and there is room for disagreement about it. 362 * 363 * Version 3.31 of the file command considered a file to be ASCII if 364 * each of its characters was approved by either the isascii() or 365 * isalpha() function. On most systems, this would mean that any 366 * file consisting only of characters in the range 0x00 ... 0x7F 367 * would be called ASCII text, but many systems might reasonably 368 * consider some characters outside this range to be alphabetic, 369 * so the file command would call such characters ASCII. It might 370 * have been more accurate to call this "considered textual on the 371 * local system" than "ASCII." 372 * 373 * It considered a file to be "International language text" if each 374 * of its characters was either an ASCII printing character (according 375 * to the real ASCII standard, not the above test), a character in 376 * the range 0x80 ... 0xFF, or one of the following control characters: 377 * backspace, tab, line feed, vertical tab, form feed, carriage return, 378 * escape. No attempt was made to determine the language in which files 379 * of this type were written. 380 * 381 * 382 * The table below considers a file to be ASCII if all of its characters 383 * are either ASCII printing characters (again, according to the X3.4 384 * standard, not isascii()) or any of the following controls: bell, 385 * backspace, tab, line feed, form feed, carriage return, esc, nextline. 386 * 387 * I include bell because some programs (particularly shell scripts) 388 * use it literally, even though it is rare in normal text. I exclude 389 * vertical tab because it never seems to be used in real text. I also 390 * include, with hesitation, the X3.64/ECMA-43 control nextline (0x85), 391 * because that's what the dd EBCDIC->ASCII table maps the EBCDIC newline 392 * character to. It might be more appropriate to include it in the 8859 393 * set instead of the ASCII set, but it's got to be included in *something* 394 * we recognize or EBCDIC files aren't going to be considered textual. 395 * Some old Unix source files use SO/SI (^N/^O) to shift between Greek 396 * and Latin characters, so these should possibly be allowed. But they 397 * make a real mess on VT100-style displays if they're not paired properly, 398 * so we are probably better off not calling them text. 399 * 400 * A file is considered to be ISO-8859 text if its characters are all 401 * either ASCII, according to the above definition, or printing characters 402 * from the ISO-8859 8-bit extension, characters 0xA0 ... 0xFF. 403 * 404 * Finally, a file is considered to be international text from some other 405 * character code if its characters are all either ISO-8859 (according to 406 * the above definition) or characters in the range 0x80 ... 0x9F, which 407 * ISO-8859 considers to be control characters but the IBM PC and Macintosh 408 * consider to be printing characters. 409 */ 410 411#define F 0 /* character never appears in text */ 412#define T 1 /* character appears in plain ASCII text */ 413#define I 2 /* character appears in ISO-8859 text */ 414#define X 3 /* character appears in non-ISO extended ASCII (Mac, IBM PC) */ 415 416static char text_chars[256] = { 417 /* BEL BS HT LF FF CR */ 418 F, F, F, F, F, F, F, T, T, T, T, F, T, T, F, F, /* 0x0X */ 419 /* ESC */ 420 F, F, F, F, F, F, F, F, F, F, F, T, F, F, F, F, /* 0x1X */ 421 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x2X */ 422 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x3X */ 423 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x4X */ 424 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x5X */ 425 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, /* 0x6X */ 426 T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, F, /* 0x7X */ 427 /* NEL */ 428 X, X, X, X, X, T, X, X, X, X, X, X, X, X, X, X, /* 0x8X */ 429 X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, X, /* 0x9X */ 430 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xaX */ 431 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xbX */ 432 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xcX */ 433 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xdX */ 434 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, /* 0xeX */ 435 I, I, I, I, I, I, I, I, I, I, I, I, I, I, I, I /* 0xfX */ 436}; 437 438static int 439looks_ascii(const unsigned char *buf, size_t nbytes, my_unichar *ubuf, 440 size_t *ulen) 441{ 442 int i; 443 444 *ulen = 0; 445 446 for (i = 0; i < (int)nbytes; i++) { 447 int t = text_chars[buf[i]]; 448 449 if (t != T) 450 return 0; 451 452 ubuf[(*ulen)++] = buf[i]; 453 } 454 455 return 1; 456} 457 458static int 459looks_latin1(const unsigned char *buf, size_t nbytes, my_unichar *ubuf, size_t *ulen) 460{ 461 int i; 462 463 *ulen = 0; 464 465 for (i = 0; i < (int)nbytes; i++) { 466 int t = text_chars[buf[i]]; 467 468 if (t != T && t != I) 469 return 0; 470 471 ubuf[(*ulen)++] = buf[i]; 472 } 473 474 return 1; 475} 476 477static int 478looks_extended(const unsigned char *buf, size_t nbytes, my_unichar *ubuf, 479 size_t *ulen) 480{ 481 int i; 482 483 *ulen = 0; 484 485 for (i = 0; i < (int)nbytes; i++) { 486 int t = text_chars[buf[i]]; 487 488 if (t != T && t != I && t != X) 489 return 0; 490 491 ubuf[(*ulen)++] = buf[i]; 492 } 493 494 return 1; 495} 496 497static int 498looks_utf8(const unsigned char *buf, size_t nbytes, my_unichar *ubuf, size_t *ulen) 499{ 500 int i, n; 501 my_unichar c; 502 int gotone = 0; 503 504 *ulen = 0; 505 506 for (i = 0; i < (int)nbytes; i++) { 507 if ((buf[i] & 0x80) == 0) { /* 0xxxxxxx is plain ASCII */ 508 /* 509 * Even if the whole file is valid UTF-8 sequences, 510 * still reject it if it uses weird control characters. 511 */ 512 513 if (text_chars[buf[i]] != T) 514 return 0; 515 516 ubuf[(*ulen)++] = buf[i]; 517 } else if ((buf[i] & 0x40) == 0) { /* 10xxxxxx never 1st byte */ 518 return 0; 519 } else { /* 11xxxxxx begins UTF-8 */ 520 int following; 521 522 if ((buf[i] & 0x20) == 0) { /* 110xxxxx */ 523 c = buf[i] & 0x1f; 524 following = 1; 525 } else if ((buf[i] & 0x10) == 0) { /* 1110xxxx */ 526 c = buf[i] & 0x0f; 527 following = 2; 528 } else if ((buf[i] & 0x08) == 0) { /* 11110xxx */ 529 c = buf[i] & 0x07; 530 following = 3; 531 } else if ((buf[i] & 0x04) == 0) { /* 111110xx */ 532 c = buf[i] & 0x03; 533 following = 4; 534 } else if ((buf[i] & 0x02) == 0) { /* 1111110x */ 535 c = buf[i] & 0x01; 536 following = 5; 537 } else 538 return 0; 539 540 for (n = 0; n < following; n++) { 541 i++; 542 if (i >= (int)nbytes) 543 goto done; 544 545 if ((buf[i] & 0x80) == 0 || (buf[i] & 0x40)) 546 return 0; 547 548 c = (c << 6) + (buf[i] & 0x3f); 549 } 550 551 ubuf[(*ulen)++] = c; 552 gotone = 1; 553 } 554 } 555done: 556 return gotone; /* don't claim it's UTF-8 if it's all 7-bit */ 557} 558 559static int 560looks_unicode(const unsigned char *buf, size_t nbytes, my_unichar *ubuf, 561 size_t *ulen) 562{ 563 int bigend; 564 int i; 565 566 if (nbytes < 2) 567 return 0; 568 569 if (buf[0] == 0xff && buf[1] == 0xfe) 570 bigend = 0; 571 else if (buf[0] == 0xfe && buf[1] == 0xff) 572 bigend = 1; 573 else 574 return 0; 575 576 *ulen = 0; 577 578 for (i = 2; i + 1 < (int)nbytes; i += 2) { 579 /* XXX fix to properly handle chars > 65536 */ 580 581 if (bigend) 582 ubuf[(*ulen)++] = buf[i + 1] + 256 * buf[i]; 583 else 584 ubuf[(*ulen)++] = buf[i] + 256 * buf[i + 1]; 585 586 if (ubuf[*ulen - 1] == 0xfffe) 587 return 0; 588 if (ubuf[*ulen - 1] < 128 && 589 text_chars[(size_t)ubuf[*ulen - 1]] != T) 590 return 0; 591 } 592 593 return 1 + bigend; 594} 595 596#undef F 597#undef T 598#undef I 599#undef X 600 601/* 602 * This table maps each EBCDIC character to an (8-bit extended) ASCII 603 * character, as specified in the rationale for the dd(1) command in 604 * draft 11.2 (September, 1991) of the POSIX P1003.2 standard. 605 * 606 * Unfortunately it does not seem to correspond exactly to any of the 607 * five variants of EBCDIC documented in IBM's _Enterprise Systems 608 * Architecture/390: Principles of Operation_, SA22-7201-06, Seventh 609 * Edition, July, 1999, pp. I-1 - I-4. 610 * 611 * Fortunately, though, all versions of EBCDIC, including this one, agree 612 * on most of the printing characters that also appear in (7-bit) ASCII. 613 * Of these, only '|', '!', '~', '^', '[', and ']' are in question at all. 614 * 615 * Fortunately too, there is general agreement that codes 0x00 through 616 * 0x3F represent control characters, 0x41 a nonbreaking space, and the 617 * remainder printing characters. 618 * 619 * This is sufficient to allow us to identify EBCDIC text and to distinguish 620 * between old-style and internationalized examples of text. 621 */ 622 623static unsigned char ebcdic_to_ascii[] = { 624 0, 1, 2, 3, 156, 9, 134, 127, 151, 141, 142, 11, 12, 13, 14, 15, 625 16, 17, 18, 19, 157, 133, 8, 135, 24, 25, 146, 143, 28, 29, 30, 31, 626128, 129, 130, 131, 132, 10, 23, 27, 136, 137, 138, 139, 140, 5, 6, 7, 627144, 145, 22, 147, 148, 149, 150, 4, 152, 153, 154, 155, 20, 21, 158, 26, 628' ', 160, 161, 162, 163, 164, 165, 166, 167, 168, 213, '.', '<', '(', '+', '|', 629'&', 169, 170, 171, 172, 173, 174, 175, 176, 177, '!', '$', '*', ')', ';', '~', 630'-', '/', 178, 179, 180, 181, 182, 183, 184, 185, 203, ',', '%', '_', '>', '?', 631186, 187, 188, 189, 190, 191, 192, 193, 194, '`', ':', '#', '@', '\'','=', '"', 632195, 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 196, 197, 198, 199, 200, 201, 633202, 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', '^', 204, 205, 206, 207, 208, 634209, 229, 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', 210, 211, 212, '[', 214, 215, 635216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, ']', 230, 231, 636'{', 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 232, 233, 234, 235, 236, 237, 637'}', 'J', 'K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 238, 239, 240, 241, 242, 243, 638'\\',159, 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 244, 245, 246, 247, 248, 249, 639'0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 250, 251, 252, 253, 254, 255 640}; 641 642#ifdef notdef 643/* 644 * The following EBCDIC-to-ASCII table may relate more closely to reality, 645 * or at least to modern reality. It comes from 646 * 647 * http://ftp.s390.ibm.com/products/oe/bpxqp9.html 648 * 649 * and maps the characters of EBCDIC code page 1047 (the code used for 650 * Unix-derived software on IBM's 390 systems) to the corresponding 651 * characters from ISO 8859-1. 652 * 653 * If this table is used instead of the above one, some of the special 654 * cases for the NEL character can be taken out of the code. 655 */ 656 657static unsigned char ebcdic_1047_to_8859[] = { 6580x00,0x01,0x02,0x03,0x9C,0x09,0x86,0x7F,0x97,0x8D,0x8E,0x0B,0x0C,0x0D,0x0E,0x0F, 6590x10,0x11,0x12,0x13,0x9D,0x0A,0x08,0x87,0x18,0x19,0x92,0x8F,0x1C,0x1D,0x1E,0x1F, 6600x80,0x81,0x82,0x83,0x84,0x85,0x17,0x1B,0x88,0x89,0x8A,0x8B,0x8C,0x05,0x06,0x07, 6610x90,0x91,0x16,0x93,0x94,0x95,0x96,0x04,0x98,0x99,0x9A,0x9B,0x14,0x15,0x9E,0x1A, 6620x20,0xA0,0xE2,0xE4,0xE0,0xE1,0xE3,0xE5,0xE7,0xF1,0xA2,0x2E,0x3C,0x28,0x2B,0x7C, 6630x26,0xE9,0xEA,0xEB,0xE8,0xED,0xEE,0xEF,0xEC,0xDF,0x21,0x24,0x2A,0x29,0x3B,0x5E, 6640x2D,0x2F,0xC2,0xC4,0xC0,0xC1,0xC3,0xC5,0xC7,0xD1,0xA6,0x2C,0x25,0x5F,0x3E,0x3F, 6650xF8,0xC9,0xCA,0xCB,0xC8,0xCD,0xCE,0xCF,0xCC,0x60,0x3A,0x23,0x40,0x27,0x3D,0x22, 6660xD8,0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0xAB,0xBB,0xF0,0xFD,0xFE,0xB1, 6670xB0,0x6A,0x6B,0x6C,0x6D,0x6E,0x6F,0x70,0x71,0x72,0xAA,0xBA,0xE6,0xB8,0xC6,0xA4, 6680xB5,0x7E,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7A,0xA1,0xBF,0xD0,0x5B,0xDE,0xAE, 6690xAC,0xA3,0xA5,0xB7,0xA9,0xA7,0xB6,0xBC,0xBD,0xBE,0xDD,0xA8,0xAF,0x5D,0xB4,0xD7, 6700x7B,0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0xAD,0xF4,0xF6,0xF2,0xF3,0xF5, 6710x7D,0x4A,0x4B,0x4C,0x4D,0x4E,0x4F,0x50,0x51,0x52,0xB9,0xFB,0xFC,0xF9,0xFA,0xFF, 6720x5C,0xF7,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5A,0xB2,0xD4,0xD6,0xD2,0xD3,0xD5, 6730x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0xB3,0xDB,0xDC,0xD9,0xDA,0x9F 674}; 675#endif 676 677/* 678 * Copy buf[0 ... nbytes-1] into out[], translating EBCDIC to ASCII. 679 */ 680static void 681from_ebcdic(const unsigned char *buf, size_t nbytes, unsigned char *out) 682{ 683 int i; 684 685 for (i = 0; i < (int)nbytes; i++) { 686 out[i] = ebcdic_to_ascii[buf[i]]; 687 } 688} 689