1 /* $NetBSD: ascmagic.c,v 1.1.1.12 2019/12/17 02:23:53 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 * ASCII magic -- try to detect text encoding. 32 * 33 * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000, 34 * to handle character codes other than ASCII on a unified basis. 35 */ 36 37 #include "file.h" 38 39 #ifndef lint 40 #if 0 41 FILE_RCSID("@(#)$File: ascmagic.c,v 1.105 2019/06/08 20:49:14 christos Exp $") 42 #else 43 __RCSID("$NetBSD: ascmagic.c,v 1.1.1.12 2019/12/17 02:23:53 christos Exp $"); 44 #endif 45 #endif /* lint */ 46 47 #include "magic.h" 48 #include <string.h> 49 #include <ctype.h> 50 #include <stdlib.h> 51 #ifdef HAVE_UNISTD_H 52 #include <unistd.h> 53 #endif 54 55 #define MAXLINELEN 300 /* longest sane line length */ 56 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \ 57 || (x) == 0x85 || (x) == '\f') 58 59 private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t); 60 private size_t trim_nuls(const unsigned char *, size_t); 61 62 /* 63 * Undo the NUL-termination kindly provided by process() 64 * but leave at least one byte to look at 65 */ 66 private size_t 67 trim_nuls(const unsigned char *buf, size_t nbytes) 68 { 69 while (nbytes > 1 && buf[nbytes - 1] == '\0') 70 nbytes--; 71 72 return nbytes; 73 } 74 75 protected int 76 file_ascmagic(struct magic_set *ms, const struct buffer *b, int text) 77 { 78 unichar *ubuf = NULL; 79 size_t ulen = 0; 80 int rv = 1; 81 struct buffer bb; 82 83 const char *code = NULL; 84 const char *code_mime = NULL; 85 const char *type = NULL; 86 87 bb = *b; 88 bb.flen = trim_nuls(CAST(const unsigned char *, b->fbuf), b->flen); 89 /* 90 * Avoid trimming at an odd byte if the original buffer was evenly 91 * sized; this avoids losing the last character on UTF-16 LE text 92 */ 93 if ((bb.flen & 1) && !(b->flen & 1)) 94 bb.flen++; 95 96 /* If file doesn't look like any sort of text, give up. */ 97 if (file_encoding(ms, &bb, &ubuf, &ulen, &code, &code_mime, 98 &type) == 0) 99 rv = 0; 100 else 101 rv = file_ascmagic_with_encoding(ms, &bb, 102 ubuf, ulen, code, type, text); 103 104 free(ubuf); 105 106 return rv; 107 } 108 109 protected int 110 file_ascmagic_with_encoding(struct magic_set *ms, 111 const struct buffer *b, unichar *ubuf, size_t ulen, const char *code, 112 const char *type, int text) 113 { 114 struct buffer bb; 115 const unsigned char *buf = CAST(const unsigned char *, b->fbuf); 116 size_t nbytes = b->flen; 117 unsigned char *utf8_buf = NULL, *utf8_end; 118 size_t mlen, i, len; 119 int rv = -1; 120 int mime = ms->flags & MAGIC_MIME; 121 int need_separator = 0; 122 123 const char *subtype = NULL; 124 const char *subtype_mime = NULL; 125 126 int has_escapes = 0; 127 int has_backspace = 0; 128 int seen_cr = 0; 129 130 int n_crlf = 0; 131 int n_lf = 0; 132 int n_cr = 0; 133 int n_nel = 0; 134 int executable = 0; 135 136 size_t last_line_end = CAST(size_t, -1); 137 int has_long_lines = 0; 138 139 nbytes = trim_nuls(buf, nbytes); 140 141 /* If we have fewer than 2 bytes, give up. */ 142 if (nbytes <= 1) { 143 rv = 0; 144 goto done; 145 } 146 147 if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) { 148 /* Convert ubuf to UTF-8 and try text soft magic */ 149 /* malloc size is a conservative overestimate; could be 150 improved, or at least realloced after conversion. */ 151 mlen = ulen * 6; 152 if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) { 153 file_oomem(ms, mlen); 154 goto done; 155 } 156 if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen)) 157 == NULL) 158 goto done; 159 buffer_init(&bb, b->fd, &b->st, utf8_buf, 160 CAST(size_t, utf8_end - utf8_buf)); 161 162 if ((rv = file_softmagic(ms, &bb, NULL, NULL, 163 TEXTTEST, text)) == 0) 164 rv = -1; 165 else 166 need_separator = 1; 167 buffer_fini(&bb); 168 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) { 169 rv = rv == -1 ? 0 : 1; 170 goto done; 171 } 172 } 173 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) 174 return 0; 175 176 /* Now try to discover other details about the file. */ 177 for (i = 0; i < ulen; i++) { 178 if (ubuf[i] == '\n') { 179 if (seen_cr) 180 n_crlf++; 181 else 182 n_lf++; 183 last_line_end = i; 184 } else if (seen_cr) 185 n_cr++; 186 187 seen_cr = (ubuf[i] == '\r'); 188 if (seen_cr) 189 last_line_end = i; 190 191 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */ 192 n_nel++; 193 last_line_end = i; 194 } 195 196 /* If this line is _longer_ than MAXLINELEN, remember it. */ 197 if (i > last_line_end + MAXLINELEN) 198 has_long_lines = 1; 199 200 if (ubuf[i] == '\033') 201 has_escapes = 1; 202 if (ubuf[i] == '\b') 203 has_backspace = 1; 204 } 205 206 /* Beware, if the data has been truncated, the final CR could have 207 been followed by a LF. If we have ms->bytes_max bytes, it indicates 208 that the data might have been truncated, probably even before 209 this function was called. */ 210 if (seen_cr && nbytes < ms->bytes_max) 211 n_cr++; 212 213 if (strcmp(type, "binary") == 0) { 214 rv = 0; 215 goto done; 216 } 217 len = file_printedlen(ms); 218 if (mime) { 219 if ((mime & MAGIC_MIME_TYPE) != 0) { 220 if (len) { 221 /* 222 * Softmagic printed something, we 223 * are either done, or we need a separator 224 */ 225 if ((ms->flags & MAGIC_CONTINUE) == 0) { 226 rv = 1; 227 goto done; 228 } 229 if (need_separator && file_separator(ms) == -1) 230 goto done; 231 } 232 if (subtype_mime) { 233 if (file_printf(ms, "%s", subtype_mime) == -1) 234 goto done; 235 } else { 236 if (file_printf(ms, "text/plain") == -1) 237 goto done; 238 } 239 } 240 } else { 241 if (len) { 242 switch (file_replace(ms, " text$", ", ")) { 243 case 0: 244 switch (file_replace(ms, " text executable$", 245 ", ")) { 246 case 0: 247 if (file_printf(ms, ", ") == -1) 248 goto done; 249 break; 250 case -1: 251 goto done; 252 default: 253 executable = 1; 254 break; 255 } 256 break; 257 case -1: 258 goto done; 259 default: 260 break; 261 } 262 } 263 264 if (file_printf(ms, "%s", code) == -1) 265 goto done; 266 267 if (subtype) { 268 if (file_printf(ms, " %s", subtype) == -1) 269 goto done; 270 } 271 272 if (file_printf(ms, " %s", type) == -1) 273 goto done; 274 275 if (executable) 276 if (file_printf(ms, " executable") == -1) 277 goto done; 278 279 if (has_long_lines) 280 if (file_printf(ms, ", with very long lines") == -1) 281 goto done; 282 283 /* 284 * Only report line terminators if we find one other than LF, 285 * or if we find none at all. 286 */ 287 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) || 288 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) { 289 if (file_printf(ms, ", with") == -1) 290 goto done; 291 292 if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) { 293 if (file_printf(ms, " no") == -1) 294 goto done; 295 } else { 296 if (n_crlf) { 297 if (file_printf(ms, " CRLF") == -1) 298 goto done; 299 if (n_cr || n_lf || n_nel) 300 if (file_printf(ms, ",") == -1) 301 goto done; 302 } 303 if (n_cr) { 304 if (file_printf(ms, " CR") == -1) 305 goto done; 306 if (n_lf || n_nel) 307 if (file_printf(ms, ",") == -1) 308 goto done; 309 } 310 if (n_lf) { 311 if (file_printf(ms, " LF") == -1) 312 goto done; 313 if (n_nel) 314 if (file_printf(ms, ",") == -1) 315 goto done; 316 } 317 if (n_nel) 318 if (file_printf(ms, " NEL") == -1) 319 goto done; 320 } 321 322 if (file_printf(ms, " line terminators") == -1) 323 goto done; 324 } 325 326 if (has_escapes) 327 if (file_printf(ms, ", with escape sequences") == -1) 328 goto done; 329 if (has_backspace) 330 if (file_printf(ms, ", with overstriking") == -1) 331 goto done; 332 } 333 rv = 1; 334 done: 335 free(utf8_buf); 336 337 return rv; 338 } 339 340 /* 341 * Encode Unicode string as UTF-8, returning pointer to character 342 * after end of string, or NULL if an invalid character is found. 343 */ 344 private unsigned char * 345 encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen) 346 { 347 size_t i; 348 unsigned char *end = buf + len; 349 350 for (i = 0; i < ulen; i++) { 351 if (ubuf[i] <= 0x7f) { 352 if (end - buf < 1) 353 return NULL; 354 *buf++ = CAST(unsigned char, ubuf[i]); 355 } else if (ubuf[i] <= 0x7ff) { 356 if (end - buf < 2) 357 return NULL; 358 *buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0); 359 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80); 360 } else if (ubuf[i] <= 0xffff) { 361 if (end - buf < 3) 362 return NULL; 363 *buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0); 364 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80); 365 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80); 366 } else if (ubuf[i] <= 0x1fffff) { 367 if (end - buf < 4) 368 return NULL; 369 *buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0); 370 *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80); 371 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80); 372 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80); 373 } else if (ubuf[i] <= 0x3ffffff) { 374 if (end - buf < 5) 375 return NULL; 376 *buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8); 377 *buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80); 378 *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80); 379 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80); 380 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80); 381 } else if (ubuf[i] <= 0x7fffffff) { 382 if (end - buf < 6) 383 return NULL; 384 *buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc); 385 *buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80); 386 *buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80); 387 *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80); 388 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80); 389 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80); 390 } else /* Invalid character */ 391 return NULL; 392 } 393 394 return buf; 395 } 396