1 /*
2 * Copyright (c) Ian F. Darwin 1986-1995.
3 * Software written by Ian F. Darwin and others;
4 * maintained 1995-present by Christos Zoulas and others.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice immediately at the beginning of the file, without modification,
11 * this list of conditions, and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
20 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28 /*
29 * ASCII magic -- try to detect text encoding.
30 *
31 * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000,
32 * to handle character codes other than ASCII on a unified basis.
33 */
34
35 #include "file.h"
36
37 #ifndef lint
38 FILE_RCSID("@(#)$File: ascmagic.c,v 1.104 2019/05/07 02:27:11 christos Exp $")
39 #endif /* lint */
40
41 #include "magic.h"
42 #include <string.h>
43 #include <memory.h>
44 #include <ctype.h>
45 #include <stdlib.h>
46 #ifdef HAVE_UNISTD_H
47 #include <unistd.h>
48 #endif
49
50 #define MAXLINELEN 300 /* longest sane line length */
51 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
52 || (x) == 0x85 || (x) == '\f')
53
54 private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t);
55 private size_t trim_nuls(const unsigned char *, size_t);
56
57 /*
58 * Undo the NUL-termination kindly provided by process()
59 * but leave at least one byte to look at
60 */
61 private size_t
trim_nuls(const unsigned char * buf,size_t nbytes)62 trim_nuls(const unsigned char *buf, size_t nbytes)
63 {
64 while (nbytes > 1 && buf[nbytes - 1] == '\0')
65 nbytes--;
66
67 return nbytes;
68 }
69
70 protected int
file_ascmagic(struct magic_set * ms,const struct buffer * b,int text)71 file_ascmagic(struct magic_set *ms, const struct buffer *b, int text)
72 {
73 unichar *ubuf = NULL;
74 size_t ulen = 0;
75 int rv = 1;
76 struct buffer bb;
77
78 const char *code = NULL;
79 const char *code_mime = NULL;
80 const char *type = NULL;
81
82 bb = *b;
83 bb.flen = trim_nuls(CAST(const unsigned char *, b->fbuf), b->flen);
84 /*
85 * Avoid trimming at an odd byte if the original buffer was evenly
86 * sized; this avoids losing the last character on UTF-16 LE text
87 */
88 if ((bb.flen & 1) && !(b->flen & 1))
89 bb.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 efree(ubuf);
100
101 return rv;
102 }
103
104 protected int
file_ascmagic_with_encoding(struct magic_set * ms,const struct buffer * b,unichar * ubuf,size_t ulen,const char * code,const char * type,int text)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 = CAST(const unsigned char *, b->fbuf);
111 size_t nbytes = b->flen;
112 unsigned char *utf8_buf = NULL, *utf8_end;
113 size_t mlen, i, len;
114 int rv = -1;
115 int mime = ms->flags & MAGIC_MIME;
116 int need_separator = 0;
117
118 const char *subtype = NULL;
119 const char *subtype_mime = NULL;
120
121 int has_escapes = 0;
122 int has_backspace = 0;
123 int seen_cr = 0;
124
125 int n_crlf = 0;
126 int n_lf = 0;
127 int n_cr = 0;
128 int n_nel = 0;
129 int executable = 0;
130
131 size_t last_line_end = CAST(size_t, -1);
132 int has_long_lines = 0;
133
134 nbytes = trim_nuls(buf, nbytes);
135
136 /* If we have fewer than 2 bytes, give up. */
137 if (nbytes <= 1) {
138 rv = 0;
139 goto done;
140 }
141
142 if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
143 /* Convert ubuf to UTF-8 and try text soft magic */
144 /* malloc size is a conservative overestimate; could be
145 improved, or at least realloced after conversion. */
146 mlen = ulen * 6;
147 if ((utf8_buf = CAST(unsigned char *, emalloc(mlen))) == NULL) {
148 file_oomem(ms, mlen);
149 goto done;
150 }
151 if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen))
152 == NULL)
153 goto done;
154 buffer_init(&bb, b->fd, &b->st, utf8_buf,
155 CAST(size_t, utf8_end - utf8_buf));
156
157 if ((rv = file_softmagic(ms, &bb, NULL, NULL,
158 TEXTTEST, text)) == 0)
159 rv = -1;
160 else
161 need_separator = 1;
162 buffer_fini(&bb);
163 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
164 rv = rv == -1 ? 0 : 1;
165 goto done;
166 }
167 }
168 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION)))
169 return 0;
170
171 /* Now try to discover other details about the file. */
172 for (i = 0; i < ulen; i++) {
173 if (ubuf[i] == '\n') {
174 if (seen_cr)
175 n_crlf++;
176 else
177 n_lf++;
178 last_line_end = i;
179 } else if (seen_cr)
180 n_cr++;
181
182 seen_cr = (ubuf[i] == '\r');
183 if (seen_cr)
184 last_line_end = i;
185
186 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
187 n_nel++;
188 last_line_end = i;
189 }
190
191 /* If this line is _longer_ than MAXLINELEN, remember it. */
192 if (i > last_line_end + MAXLINELEN)
193 has_long_lines = 1;
194
195 if (ubuf[i] == '\033')
196 has_escapes = 1;
197 if (ubuf[i] == '\b')
198 has_backspace = 1;
199 }
200
201 /* Beware, if the data has been truncated, the final CR could have
202 been followed by a LF. If we have ms->bytes_max bytes, it indicates
203 that the data might have been truncated, probably even before
204 this function was called. */
205 if (seen_cr && nbytes < ms->bytes_max)
206 n_cr++;
207
208 if (strcmp(type, "binary") == 0) {
209 rv = 0;
210 goto done;
211 }
212 len = file_printedlen(ms);
213 if (mime) {
214 if ((mime & MAGIC_MIME_TYPE) != 0) {
215 if (len) {
216 /*
217 * Softmagic printed something, we
218 * are either done, or we need a separator
219 */
220 if ((ms->flags & MAGIC_CONTINUE) == 0) {
221 rv = 1;
222 goto done;
223 }
224 if (need_separator && file_separator(ms) == -1)
225 goto done;
226 }
227 if (subtype_mime) {
228 if (file_printf(ms, "%s", subtype_mime) == -1)
229 goto done;
230 } else {
231 if (file_printf(ms, "text/plain") == -1)
232 goto done;
233 }
234 }
235 } else {
236 if (len) {
237 switch (file_replace(ms, " text$", ", ")) {
238 case 0:
239 switch (file_replace(ms, " text executable$",
240 ", ")) {
241 case 0:
242 if (file_printf(ms, ", ") == -1)
243 goto done;
244 break;
245 case -1:
246 goto done;
247 default:
248 executable = 1;
249 break;
250 }
251 break;
252 case -1:
253 goto done;
254 default:
255 break;
256 }
257 }
258
259 if (file_printf(ms, "%s", code) == -1)
260 goto done;
261
262 if (subtype) {
263 if (file_printf(ms, " %s", subtype) == -1)
264 goto done;
265 }
266
267 if (file_printf(ms, " %s", type) == -1)
268 goto done;
269
270 if (executable)
271 if (file_printf(ms, " executable") == -1)
272 goto done;
273
274 if (has_long_lines)
275 if (file_printf(ms, ", with very long lines") == -1)
276 goto done;
277
278 /*
279 * Only report line terminators if we find one other than LF,
280 * or if we find none at all.
281 */
282 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
283 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
284 if (file_printf(ms, ", with") == -1)
285 goto done;
286
287 if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
288 if (file_printf(ms, " no") == -1)
289 goto done;
290 } else {
291 if (n_crlf) {
292 if (file_printf(ms, " CRLF") == -1)
293 goto done;
294 if (n_cr || n_lf || n_nel)
295 if (file_printf(ms, ",") == -1)
296 goto done;
297 }
298 if (n_cr) {
299 if (file_printf(ms, " CR") == -1)
300 goto done;
301 if (n_lf || n_nel)
302 if (file_printf(ms, ",") == -1)
303 goto done;
304 }
305 if (n_lf) {
306 if (file_printf(ms, " LF") == -1)
307 goto done;
308 if (n_nel)
309 if (file_printf(ms, ",") == -1)
310 goto done;
311 }
312 if (n_nel)
313 if (file_printf(ms, " NEL") == -1)
314 goto done;
315 }
316
317 if (file_printf(ms, " line terminators") == -1)
318 goto done;
319 }
320
321 if (has_escapes)
322 if (file_printf(ms, ", with escape sequences") == -1)
323 goto done;
324 if (has_backspace)
325 if (file_printf(ms, ", with overstriking") == -1)
326 goto done;
327 }
328 rv = 1;
329 done:
330 if (utf8_buf)
331 efree(utf8_buf);
332
333 return rv;
334 }
335
336 /*
337 * Encode Unicode string as UTF-8, returning pointer to character
338 * after end of string, or NULL if an invalid character is found.
339 */
340 private unsigned char *
encode_utf8(unsigned char * buf,size_t len,unichar * ubuf,size_t ulen)341 encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen)
342 {
343 size_t i;
344 unsigned char *end = buf + len;
345
346 for (i = 0; i < ulen; i++) {
347 if (ubuf[i] <= 0x7f) {
348 if (end - buf < 1)
349 return NULL;
350 *buf++ = CAST(unsigned char, ubuf[i]);
351 } else if (ubuf[i] <= 0x7ff) {
352 if (end - buf < 2)
353 return NULL;
354 *buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0);
355 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
356 } else if (ubuf[i] <= 0xffff) {
357 if (end - buf < 3)
358 return NULL;
359 *buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0);
360 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
361 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
362 } else if (ubuf[i] <= 0x1fffff) {
363 if (end - buf < 4)
364 return NULL;
365 *buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0);
366 *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
367 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
368 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
369 } else if (ubuf[i] <= 0x3ffffff) {
370 if (end - buf < 5)
371 return NULL;
372 *buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8);
373 *buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
374 *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
375 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
376 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
377 } else if (ubuf[i] <= 0x7fffffff) {
378 if (end - buf < 6)
379 return NULL;
380 *buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc);
381 *buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80);
382 *buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
383 *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
384 *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
385 *buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
386 } else /* Invalid character */
387 return NULL;
388 }
389
390 return buf;
391 }
392
