1 /* $NetBSD: ascmagic.c,v 1.1.1.14 2021/04/09 18:58:01 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 1.109 2021/02/05 23:01:40 christos Exp $") 42 #else 43 __RCSID("$NetBSD: ascmagic.c,v 1.1.1.14 2021/04/09 18:58:01 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, file_unichar_t *, 60 size_t); 61 private size_t trim_nuls(const unsigned char *, size_t); 62 63 /* 64 * Undo the NUL-termination kindly provided by process() 65 * but leave at least one byte to look at 66 */ 67 private size_t 68 trim_nuls(const unsigned char *buf, size_t nbytes) 69 { 70 while (nbytes > 1 && buf[nbytes - 1] == '\0') 71 nbytes--; 72 73 return nbytes; 74 } 75 76 protected int 77 file_ascmagic(struct magic_set *ms, const struct buffer *b, int text) 78 { 79 file_unichar_t *ubuf = NULL; 80 size_t ulen = 0; 81 int rv = 1; 82 struct buffer bb; 83 84 const char *code = NULL; 85 const char *code_mime = NULL; 86 const char *type = NULL; 87 88 bb = *b; 89 bb.flen = trim_nuls(CAST(const unsigned char *, b->fbuf), b->flen); 90 /* 91 * Avoid trimming at an odd byte if the original buffer was evenly 92 * sized; this avoids losing the last character on UTF-16 LE text 93 */ 94 if ((bb.flen & 1) && !(b->flen & 1)) 95 bb.flen++; 96 97 /* If file doesn't look like any sort of text, give up. */ 98 if (file_encoding(ms, &bb, &ubuf, &ulen, &code, &code_mime, 99 &type) == 0) 100 rv = 0; 101 else 102 rv = file_ascmagic_with_encoding(ms, &bb, 103 ubuf, ulen, code, type, text); 104 105 free(ubuf); 106 107 return rv; 108 } 109 110 protected int 111 file_ascmagic_with_encoding(struct magic_set *ms, const struct buffer *b, 112 file_unichar_t *ubuf, size_t ulen, const char *code, const char *type, 113 int text) 114 { 115 struct buffer bb; 116 const unsigned char *buf = CAST(const unsigned char *, b->fbuf); 117 size_t nbytes = b->flen; 118 unsigned char *utf8_buf = NULL, *utf8_end; 119 size_t mlen, i, len; 120 int rv = -1; 121 int mime = ms->flags & MAGIC_MIME; 122 int need_separator = 0; 123 124 const char *subtype = 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 size_t 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 } 173 174 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) { 175 rv = 0; 176 goto done; 177 } 178 179 /* Now try to discover other details about the file. */ 180 for (i = 0; i < ulen; i++) { 181 if (ubuf[i] == '\n') { 182 if (seen_cr) 183 n_crlf++; 184 else 185 n_lf++; 186 last_line_end = i; 187 } else if (seen_cr) 188 n_cr++; 189 190 seen_cr = (ubuf[i] == '\r'); 191 if (seen_cr) 192 last_line_end = i; 193 194 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */ 195 n_nel++; 196 last_line_end = i; 197 } 198 199 /* If this line is _longer_ than MAXLINELEN, remember it. */ 200 if (i > last_line_end + MAXLINELEN) { 201 size_t ll = i - last_line_end; 202 if (ll > has_long_lines) 203 has_long_lines = ll; 204 } 205 206 if (ubuf[i] == '\033') 207 has_escapes = 1; 208 if (ubuf[i] == '\b') 209 has_backspace = 1; 210 } 211 212 /* Beware, if the data has been truncated, the final CR could have 213 been followed by a LF. If we have ms->bytes_max bytes, it indicates 214 that the data might have been truncated, probably even before 215 this function was called. */ 216 if (seen_cr && nbytes < ms->bytes_max) 217 n_cr++; 218 219 if (strcmp(type, "binary") == 0) { 220 rv = 0; 221 goto done; 222 } 223 len = file_printedlen(ms); 224 if (mime) { 225 if ((mime & MAGIC_MIME_TYPE) != 0) { 226 if (len) { 227 /* 228 * Softmagic printed something, we 229 * are either done, or we need a separator 230 */ 231 if ((ms->flags & MAGIC_CONTINUE) == 0) { 232 rv = 1; 233 goto done; 234 } 235 if (need_separator && file_separator(ms) == -1) 236 goto done; 237 } else { 238 if (file_printf(ms, "text/plain") == -1) 239 goto done; 240 } 241 } 242 } else { 243 if (len) { 244 switch (file_replace(ms, " text$", ", ")) { 245 case 0: 246 switch (file_replace(ms, " text executable$", 247 ", ")) { 248 case 0: 249 if (file_printf(ms, ", ") == -1) 250 goto done; 251 break; 252 case -1: 253 goto done; 254 default: 255 executable = 1; 256 break; 257 } 258 break; 259 case -1: 260 goto done; 261 default: 262 break; 263 } 264 } 265 266 if (file_printf(ms, "%s", code) == -1) 267 goto done; 268 269 if (subtype) { 270 if (file_printf(ms, " %s", subtype) == -1) 271 goto done; 272 } 273 274 if (file_printf(ms, " %s", type) == -1) 275 goto done; 276 277 if (executable) 278 if (file_printf(ms, " executable") == -1) 279 goto done; 280 281 if (has_long_lines) 282 if (file_printf(ms, ", with very long lines (%zu)", 283 has_long_lines) == -1) 284 goto done; 285 286 /* 287 * Only report line terminators if we find one other than LF, 288 * or if we find none at all. 289 */ 290 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) || 291 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) { 292 if (file_printf(ms, ", with") == -1) 293 goto done; 294 295 if (n_crlf == 0 && n_cr == 0 && 296 n_nel == 0 && n_lf == 0) { 297 if (file_printf(ms, " no") == -1) 298 goto done; 299 } else { 300 if (n_crlf) { 301 if (file_printf(ms, " CRLF") == -1) 302 goto done; 303 if (n_cr || n_lf || n_nel) 304 if (file_printf(ms, ",") == -1) 305 goto done; 306 } 307 if (n_cr) { 308 if (file_printf(ms, " CR") == -1) 309 goto done; 310 if (n_lf || n_nel) 311 if (file_printf(ms, ",") == -1) 312 goto done; 313 } 314 if (n_lf) { 315 if (file_printf(ms, " LF") == -1) 316 goto done; 317 if (n_nel) 318 if (file_printf(ms, ",") == -1) 319 goto done; 320 } 321 if (n_nel) 322 if (file_printf(ms, " NEL") == -1) 323 goto done; 324 } 325 326 if (file_printf(ms, " line terminators") == -1) 327 goto done; 328 } 329 330 if (has_escapes) 331 if (file_printf(ms, ", with escape sequences") == -1) 332 goto done; 333 if (has_backspace) 334 if (file_printf(ms, ", with overstriking") == -1) 335 goto done; 336 } 337 rv = 1; 338 done: 339 free(utf8_buf); 340 341 return rv; 342 } 343 344 /* 345 * Encode Unicode string as UTF-8, returning pointer to character 346 * after end of string, or NULL if an invalid character is found. 347 */ 348 private unsigned char * 349 encode_utf8(unsigned char *buf, size_t len, file_unichar_t *ubuf, size_t ulen) 350 { 351 size_t i; 352 unsigned char *end = buf + len; 353 354 for (i = 0; i < ulen; i++) { 355 if (ubuf[i] <= 0x7f) { 356 if (end - buf < 1) 357 return NULL; 358 *buf++ = CAST(unsigned char, ubuf[i]); 359 continue; 360 } 361 if (ubuf[i] <= 0x7ff) { 362 if (end - buf < 2) 363 return NULL; 364 *buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0); 365 goto out1; 366 } 367 if (ubuf[i] <= 0xffff) { 368 if (end - buf < 3) 369 return NULL; 370 *buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0); 371 goto out2; 372 } 373 if (ubuf[i] <= 0x1fffff) { 374 if (end - buf < 4) 375 return NULL; 376 *buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0); 377 goto out3; 378 } 379 if (ubuf[i] <= 0x3ffffff) { 380 if (end - buf < 5) 381 return NULL; 382 *buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8); 383 goto out4; 384 } 385 if (ubuf[i] <= 0x7fffffff) { 386 if (end - buf < 6) 387 return NULL; 388 *buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc); 389 goto out5; 390 } 391 /* Invalid character */ 392 return NULL; 393 out5: *buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80); 394 out4: *buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80); 395 out3: *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80); 396 out2: *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80); 397 out1: *buf++ = CAST(unsigned char, ((ubuf[i] >> 0) & 0x3f) + 0x80); 398 } 399 400 return buf; 401 } 402