xref: /dragonfly/contrib/file/src/funcs.c (revision 2020c8fe)
1 /*
2  * Copyright (c) Christos Zoulas 2003.
3  * All Rights Reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice immediately at the beginning of the file, without modification,
10  *    this list of conditions, and the following disclaimer.
11  * 2. Redistributions in binary form must reproduce the above copyright
12  *    notice, this list of conditions and the following disclaimer in the
13  *    documentation and/or other materials provided with the distribution.
14  *
15  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18  * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
19  * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25  * SUCH DAMAGE.
26  */
27 #include "file.h"
28 
29 #ifndef	lint
30 FILE_RCSID("@(#)$File: funcs.c,v 1.60 2011/12/08 12:38:24 rrt Exp $")
31 #endif	/* lint */
32 
33 #include "magic.h"
34 #include <stdarg.h>
35 #include <stdlib.h>
36 #include <string.h>
37 #include <ctype.h>
38 #if defined(HAVE_WCHAR_H)
39 #include <wchar.h>
40 #endif
41 #if defined(HAVE_WCTYPE_H)
42 #include <wctype.h>
43 #endif
44 #if defined(HAVE_LIMITS_H)
45 #include <limits.h>
46 #endif
47 
48 #ifndef SIZE_MAX
49 #define SIZE_MAX	((size_t)~0)
50 #endif
51 
52 /*
53  * Like printf, only we append to a buffer.
54  */
55 protected int
56 file_vprintf(struct magic_set *ms, const char *fmt, va_list ap)
57 {
58 	int len;
59 	char *buf, *newstr;
60 
61 	len = vasprintf(&buf, fmt, ap);
62 	if (len < 0)
63 		goto out;
64 
65 	if (ms->o.buf != NULL) {
66 		len = asprintf(&newstr, "%s%s", ms->o.buf, buf);
67 		free(buf);
68 		if (len < 0)
69 			goto out;
70 		free(ms->o.buf);
71 		buf = newstr;
72 	}
73 	ms->o.buf = buf;
74 	return 0;
75 out:
76 	file_error(ms, errno, "vasprintf failed");
77 	return -1;
78 }
79 
80 protected int
81 file_printf(struct magic_set *ms, const char *fmt, ...)
82 {
83 	int rv;
84 	va_list ap;
85 
86 	va_start(ap, fmt);
87 	rv = file_vprintf(ms, fmt, ap);
88 	va_end(ap);
89 	return rv;
90 }
91 
92 /*
93  * error - print best error message possible
94  */
95 /*VARARGS*/
96 private void
97 file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
98     size_t lineno)
99 {
100 	/* Only the first error is ok */
101 	if (ms->event_flags & EVENT_HAD_ERR)
102 		return;
103 	if (lineno != 0) {
104 		free(ms->o.buf);
105 		ms->o.buf = NULL;
106 		file_printf(ms, "line %" SIZE_T_FORMAT "u: ", lineno);
107 	}
108 	file_vprintf(ms, f, va);
109 	if (error > 0)
110 		file_printf(ms, " (%s)", strerror(error));
111 	ms->event_flags |= EVENT_HAD_ERR;
112 	ms->error = error;
113 }
114 
115 /*VARARGS*/
116 protected void
117 file_error(struct magic_set *ms, int error, const char *f, ...)
118 {
119 	va_list va;
120 	va_start(va, f);
121 	file_error_core(ms, error, f, va, 0);
122 	va_end(va);
123 }
124 
125 /*
126  * Print an error with magic line number.
127  */
128 /*VARARGS*/
129 protected void
130 file_magerror(struct magic_set *ms, const char *f, ...)
131 {
132 	va_list va;
133 	va_start(va, f);
134 	file_error_core(ms, 0, f, va, ms->line);
135 	va_end(va);
136 }
137 
138 protected void
139 file_oomem(struct magic_set *ms, size_t len)
140 {
141 	file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
142 	    len);
143 }
144 
145 protected void
146 file_badseek(struct magic_set *ms)
147 {
148 	file_error(ms, errno, "error seeking");
149 }
150 
151 protected void
152 file_badread(struct magic_set *ms)
153 {
154 	file_error(ms, errno, "error reading");
155 }
156 
157 #ifndef COMPILE_ONLY
158 protected int
159 file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((unused)),
160     const void *buf, size_t nb)
161 {
162 	int m = 0, rv = 0, looks_text = 0;
163 	int mime = ms->flags & MAGIC_MIME;
164 	const unsigned char *ubuf = CAST(const unsigned char *, buf);
165 	unichar *u8buf = NULL;
166 	size_t ulen;
167 	const char *code = NULL;
168 	const char *code_mime = "binary";
169 	const char *type = NULL;
170 
171 
172 
173 	if (nb == 0) {
174 		if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
175 		    file_printf(ms, mime ? "application/x-empty" :
176 		    "empty") == -1)
177 			return -1;
178 		return 1;
179 	} else if (nb == 1) {
180 		if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
181 		    file_printf(ms, mime ? "application/octet-stream" :
182 		    "very short file (no magic)") == -1)
183 			return -1;
184 		return 1;
185 	}
186 
187 	if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
188 		looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen,
189 		    &code, &code_mime, &type);
190 	}
191 
192 #ifdef __EMX__
193 	if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
194 		switch (file_os2_apptype(ms, inname, buf, nb)) {
195 		case -1:
196 			return -1;
197 		case 0:
198 			break;
199 		default:
200 			return 1;
201 		}
202 	}
203 #endif
204 #if HAVE_FORK
205 	/* try compression stuff */
206 	if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0)
207 		if ((m = file_zmagic(ms, fd, inname, ubuf, nb)) != 0) {
208 			if ((ms->flags & MAGIC_DEBUG) != 0)
209 				(void)fprintf(stderr, "zmagic %d\n", m);
210 			goto done;
211 		}
212 #endif
213 	/* Check if we have a tar file */
214 	if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0)
215 		if ((m = file_is_tar(ms, ubuf, nb)) != 0) {
216 			if ((ms->flags & MAGIC_DEBUG) != 0)
217 				(void)fprintf(stderr, "tar %d\n", m);
218 			goto done;
219 		}
220 
221 	/* Check if we have a CDF file */
222 	if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0)
223 		if ((m = file_trycdf(ms, fd, ubuf, nb)) != 0) {
224 			if ((ms->flags & MAGIC_DEBUG) != 0)
225 				(void)fprintf(stderr, "cdf %d\n", m);
226 			goto done;
227 		}
228 
229 	/* try soft magic tests */
230 	if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0)
231 		if ((m = file_softmagic(ms, ubuf, nb, BINTEST,
232 		    looks_text)) != 0) {
233 			if ((ms->flags & MAGIC_DEBUG) != 0)
234 				(void)fprintf(stderr, "softmagic %d\n", m);
235 #ifdef BUILTIN_ELF
236 			if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
237 			    nb > 5 && fd != -1) {
238 				/*
239 				 * We matched something in the file, so this
240 				 * *might* be an ELF file, and the file is at
241 				 * least 5 bytes long, so if it's an ELF file
242 				 * it has at least one byte past the ELF magic
243 				 * number - try extracting information from the
244 				 * ELF headers that cannot easily * be
245 				 * extracted with rules in the magic file.
246 				 */
247 				if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0)
248 					if ((ms->flags & MAGIC_DEBUG) != 0)
249 						(void)fprintf(stderr,
250 						    "elf %d\n", m);
251 			}
252 #endif
253 			goto done;
254 		}
255 
256 	/* try text properties */
257 	if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
258 
259 		if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) {
260 			if ((ms->flags & MAGIC_DEBUG) != 0)
261 				(void)fprintf(stderr, "ascmagic %d\n", m);
262 			goto done;
263 		}
264 
265 		/* try to discover text encoding */
266 		if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
267 			if (looks_text == 0)
268 				if ((m = file_ascmagic_with_encoding( ms, ubuf,
269 				    nb, u8buf, ulen, code, type, looks_text))
270 				    != 0) {
271 					if ((ms->flags & MAGIC_DEBUG) != 0)
272 						(void)fprintf(stderr,
273 						    "ascmagic/enc %d\n", m);
274 					goto done;
275 				}
276 		}
277 	}
278 
279 	/* give up */
280 	m = 1;
281 	if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
282 	    file_printf(ms, mime ? "application/octet-stream" : "data") == -1) {
283 	    rv = -1;
284 	}
285  done:
286 	if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
287 		if (ms->flags & MAGIC_MIME_TYPE)
288 			if (file_printf(ms, "; charset=") == -1)
289 				rv = -1;
290 		if (file_printf(ms, "%s", code_mime) == -1)
291 			rv = -1;
292 	}
293 	free(u8buf);
294 	if (rv)
295 		return rv;
296 
297 	return m;
298 }
299 #endif
300 
301 protected int
302 file_reset(struct magic_set *ms)
303 {
304 	if (ms->mlist == NULL) {
305 		file_error(ms, 0, "no magic files loaded");
306 		return -1;
307 	}
308 	if (ms->o.buf) {
309 		free(ms->o.buf);
310 		ms->o.buf = NULL;
311 	}
312 	if (ms->o.pbuf) {
313 		free(ms->o.pbuf);
314 		ms->o.pbuf = NULL;
315 	}
316 	ms->event_flags &= ~EVENT_HAD_ERR;
317 	ms->error = -1;
318 	return 0;
319 }
320 
321 #define OCTALIFY(n, o)	\
322 	/*LINTED*/ \
323 	(void)(*(n)++ = '\\', \
324 	*(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
325 	*(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
326 	*(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
327 	(o)++)
328 
329 protected const char *
330 file_getbuffer(struct magic_set *ms)
331 {
332 	char *pbuf, *op, *np;
333 	size_t psize, len;
334 
335 	if (ms->event_flags & EVENT_HAD_ERR)
336 		return NULL;
337 
338 	if (ms->flags & MAGIC_RAW)
339 		return ms->o.buf;
340 
341 	if (ms->o.buf == NULL)
342 		return NULL;
343 
344 	/* * 4 is for octal representation, + 1 is for NUL */
345 	len = strlen(ms->o.buf);
346 	if (len > (SIZE_MAX - 1) / 4) {
347 		file_oomem(ms, len);
348 		return NULL;
349 	}
350 	psize = len * 4 + 1;
351 	if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
352 		file_oomem(ms, psize);
353 		return NULL;
354 	}
355 	ms->o.pbuf = pbuf;
356 
357 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
358 	{
359 		mbstate_t state;
360 		wchar_t nextchar;
361 		int mb_conv = 1;
362 		size_t bytesconsumed;
363 		char *eop;
364 		(void)memset(&state, 0, sizeof(mbstate_t));
365 
366 		np = ms->o.pbuf;
367 		op = ms->o.buf;
368 		eop = op + len;
369 
370 		while (op < eop) {
371 			bytesconsumed = mbrtowc(&nextchar, op,
372 			    (size_t)(eop - op), &state);
373 			if (bytesconsumed == (size_t)(-1) ||
374 			    bytesconsumed == (size_t)(-2)) {
375 				mb_conv = 0;
376 				break;
377 			}
378 
379 			if (iswprint(nextchar)) {
380 				(void)memcpy(np, op, bytesconsumed);
381 				op += bytesconsumed;
382 				np += bytesconsumed;
383 			} else {
384 				while (bytesconsumed-- > 0)
385 					OCTALIFY(np, op);
386 			}
387 		}
388 		*np = '\0';
389 
390 		/* Parsing succeeded as a multi-byte sequence */
391 		if (mb_conv != 0)
392 			return ms->o.pbuf;
393 	}
394 #endif
395 
396 	for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
397 		if (isprint((unsigned char)*op)) {
398 			*np++ = *op++;
399 		} else {
400 			OCTALIFY(np, op);
401 		}
402 	}
403 	*np = '\0';
404 	return ms->o.pbuf;
405 }
406 
407 protected int
408 file_check_mem(struct magic_set *ms, unsigned int level)
409 {
410 	size_t len;
411 
412 	if (level >= ms->c.len) {
413 		len = (ms->c.len += 20) * sizeof(*ms->c.li);
414 		ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
415 		    malloc(len) :
416 		    realloc(ms->c.li, len));
417 		if (ms->c.li == NULL) {
418 			file_oomem(ms, len);
419 			return -1;
420 		}
421 	}
422 	ms->c.li[level].got_match = 0;
423 #ifdef ENABLE_CONDITIONALS
424 	ms->c.li[level].last_match = 0;
425 	ms->c.li[level].last_cond = COND_NONE;
426 #endif /* ENABLE_CONDITIONALS */
427 	return 0;
428 }
429 
430 protected size_t
431 file_printedlen(const struct magic_set *ms)
432 {
433 	return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
434 }
435 
436 protected int
437 file_replace(struct magic_set *ms, const char *pat, const char *rep)
438 {
439 	regex_t rx;
440 	int rc;
441 
442 	rc = regcomp(&rx, pat, REG_EXTENDED);
443 	if (rc) {
444 		char errmsg[512];
445 		(void)regerror(rc, &rx, errmsg, sizeof(errmsg));
446 		file_magerror(ms, "regex error %d, (%s)", rc, errmsg);
447 		return -1;
448 	} else {
449 		regmatch_t rm;
450 		int nm = 0;
451 		while (regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
452 			ms->o.buf[rm.rm_so] = '\0';
453 			if (file_printf(ms, "%s%s", rep,
454 			    rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
455 				return -1;
456 			nm++;
457 		}
458 		regfree(&rx);
459 		return nm;
460 	}
461 }
462