1 /* $NetBSD: ascmagic.c,v 1.1.1.9 2018/04/15 19:32:48 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 * 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 1.98 2017/11/02 20:25:39 christos Exp $") 42 #else 43 __RCSID("$NetBSD: ascmagic.c,v 1.1.1.9 2018/04/15 19:32:48 christos Exp $"); 44 #endif 45 #endif /* lint */ 46 47 #include "magic.h" 48 #include <string.h> 49 #include <memory.h> 50 #include <ctype.h> 51 #include <stdlib.h> 52 #ifdef HAVE_UNISTD_H 53 #include <unistd.h> 54 #endif 55 56 #define MAXLINELEN 300 /* longest sane line length */ 57 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \ 58 || (x) == 0x85 || (x) == '\f') 59 60 private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, 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 unichar *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(b->fbuf, b->flen); 90 91 /* If file doesn't look like any sort of text, give up. */ 92 if (file_encoding(ms, &bb, &ubuf, &ulen, &code, &code_mime, 93 &type) == 0) 94 rv = 0; 95 else 96 rv = file_ascmagic_with_encoding(ms, &bb, 97 ubuf, ulen, code, type, text); 98 99 free(ubuf); 100 101 return rv; 102 } 103 104 protected int 105 file_ascmagic_with_encoding(struct magic_set *ms, 106 const struct buffer *b, unichar *ubuf, size_t ulen, const char *code, 107 const char *type, int text) 108 { 109 struct buffer bb; 110 const unsigned char *buf = b->fbuf; 111 size_t nbytes = b->flen; 112 unsigned char *utf8_buf = NULL, *utf8_end; 113 size_t mlen, i; 114 int rv = -1; 115 int mime = ms->flags & MAGIC_MIME; 116 117 const char *subtype = NULL; 118 const char *subtype_mime = NULL; 119 120 int has_escapes = 0; 121 int has_backspace = 0; 122 int seen_cr = 0; 123 124 int n_crlf = 0; 125 int n_lf = 0; 126 int n_cr = 0; 127 int n_nel = 0; 128 int executable = 0; 129 130 size_t last_line_end = (size_t)-1; 131 int has_long_lines = 0; 132 133 nbytes = trim_nuls(buf, nbytes); 134 135 /* If we have fewer than 2 bytes, give up. */ 136 if (nbytes <= 1) { 137 rv = 0; 138 goto done; 139 } 140 141 if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) { 142 /* Convert ubuf to UTF-8 and try text soft magic */ 143 /* malloc size is a conservative overestimate; could be 144 improved, or at least realloced after conversion. */ 145 mlen = ulen * 6; 146 if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) { 147 file_oomem(ms, mlen); 148 goto done; 149 } 150 if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen)) 151 == NULL) 152 goto done; 153 buffer_init(&bb, b->fd, utf8_buf, 154 (size_t)(utf8_end - utf8_buf)); 155 156 if ((rv = file_softmagic(ms, &bb, NULL, NULL, 157 TEXTTEST, text)) == 0) 158 rv = -1; 159 buffer_fini(&bb); 160 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) { 161 rv = rv == -1 ? 0 : 1; 162 goto done; 163 } 164 } 165 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) 166 return 0; 167 168 /* Now try to discover other details about the file. */ 169 for (i = 0; i < ulen; i++) { 170 if (ubuf[i] == '\n') { 171 if (seen_cr) 172 n_crlf++; 173 else 174 n_lf++; 175 last_line_end = i; 176 } else if (seen_cr) 177 n_cr++; 178 179 seen_cr = (ubuf[i] == '\r'); 180 if (seen_cr) 181 last_line_end = i; 182 183 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */ 184 n_nel++; 185 last_line_end = i; 186 } 187 188 /* If this line is _longer_ than MAXLINELEN, remember it. */ 189 if (i > last_line_end + MAXLINELEN) 190 has_long_lines = 1; 191 192 if (ubuf[i] == '\033') 193 has_escapes = 1; 194 if (ubuf[i] == '\b') 195 has_backspace = 1; 196 } 197 198 /* Beware, if the data has been truncated, the final CR could have 199 been followed by a LF. If we have ms->bytes_max bytes, it indicates 200 that the data might have been truncated, probably even before 201 this function was called. */ 202 if (seen_cr && nbytes < ms->bytes_max) 203 n_cr++; 204 205 if (strcmp(type, "binary") == 0) { 206 rv = 0; 207 goto done; 208 } 209 if (mime) { 210 if (!file_printedlen(ms) && (mime & MAGIC_MIME_TYPE) != 0) { 211 if (subtype_mime) { 212 if (file_printf(ms, "%s", subtype_mime) == -1) 213 goto done; 214 } else { 215 if (file_printf(ms, "text/plain") == -1) 216 goto done; 217 } 218 } 219 } else { 220 if (file_printedlen(ms)) { 221 switch (file_replace(ms, " text$", ", ")) { 222 case 0: 223 switch (file_replace(ms, " text executable$", 224 ", ")) { 225 case 0: 226 if (file_printf(ms, ", ") == -1) 227 goto done; 228 break; 229 case -1: 230 goto done; 231 default: 232 executable = 1; 233 break; 234 } 235 break; 236 case -1: 237 goto done; 238 default: 239 break; 240 } 241 } 242 243 if (file_printf(ms, "%s", code) == -1) 244 goto done; 245 246 if (subtype) { 247 if (file_printf(ms, " %s", subtype) == -1) 248 goto done; 249 } 250 251 if (file_printf(ms, " %s", type) == -1) 252 goto done; 253 254 if (executable) 255 if (file_printf(ms, " executable") == -1) 256 goto done; 257 258 if (has_long_lines) 259 if (file_printf(ms, ", with very long lines") == -1) 260 goto done; 261 262 /* 263 * Only report line terminators if we find one other than LF, 264 * or if we find none at all. 265 */ 266 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) || 267 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) { 268 if (file_printf(ms, ", with") == -1) 269 goto done; 270 271 if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) { 272 if (file_printf(ms, " no") == -1) 273 goto done; 274 } else { 275 if (n_crlf) { 276 if (file_printf(ms, " CRLF") == -1) 277 goto done; 278 if (n_cr || n_lf || n_nel) 279 if (file_printf(ms, ",") == -1) 280 goto done; 281 } 282 if (n_cr) { 283 if (file_printf(ms, " CR") == -1) 284 goto done; 285 if (n_lf || n_nel) 286 if (file_printf(ms, ",") == -1) 287 goto done; 288 } 289 if (n_lf) { 290 if (file_printf(ms, " LF") == -1) 291 goto done; 292 if (n_nel) 293 if (file_printf(ms, ",") == -1) 294 goto done; 295 } 296 if (n_nel) 297 if (file_printf(ms, " NEL") == -1) 298 goto done; 299 } 300 301 if (file_printf(ms, " line terminators") == -1) 302 goto done; 303 } 304 305 if (has_escapes) 306 if (file_printf(ms, ", with escape sequences") == -1) 307 goto done; 308 if (has_backspace) 309 if (file_printf(ms, ", with overstriking") == -1) 310 goto done; 311 } 312 rv = 1; 313 done: 314 free(utf8_buf); 315 316 return rv; 317 } 318 319 /* 320 * Encode Unicode string as UTF-8, returning pointer to character 321 * after end of string, or NULL if an invalid character is found. 322 */ 323 private unsigned char * 324 encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen) 325 { 326 size_t i; 327 unsigned char *end = buf + len; 328 329 for (i = 0; i < ulen; i++) { 330 if (ubuf[i] <= 0x7f) { 331 if (end - buf < 1) 332 return NULL; 333 *buf++ = (unsigned char)ubuf[i]; 334 } else if (ubuf[i] <= 0x7ff) { 335 if (end - buf < 2) 336 return NULL; 337 *buf++ = (unsigned char)((ubuf[i] >> 6) + 0xc0); 338 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); 339 } else if (ubuf[i] <= 0xffff) { 340 if (end - buf < 3) 341 return NULL; 342 *buf++ = (unsigned char)((ubuf[i] >> 12) + 0xe0); 343 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); 344 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); 345 } else if (ubuf[i] <= 0x1fffff) { 346 if (end - buf < 4) 347 return NULL; 348 *buf++ = (unsigned char)((ubuf[i] >> 18) + 0xf0); 349 *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80); 350 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); 351 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); 352 } else if (ubuf[i] <= 0x3ffffff) { 353 if (end - buf < 5) 354 return NULL; 355 *buf++ = (unsigned char)((ubuf[i] >> 24) + 0xf8); 356 *buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80); 357 *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80); 358 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); 359 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); 360 } else if (ubuf[i] <= 0x7fffffff) { 361 if (end - buf < 6) 362 return NULL; 363 *buf++ = (unsigned char)((ubuf[i] >> 30) + 0xfc); 364 *buf++ = (unsigned char)(((ubuf[i] >> 24) & 0x3f) + 0x80); 365 *buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80); 366 *buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80); 367 *buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80); 368 *buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80); 369 } else /* Invalid character */ 370 return NULL; 371 } 372 373 return buf; 374 } 375