xref: /dragonfly/contrib/file/src/funcs.c (revision a563ca70)
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 1.57 2011/05/11 01:02:41 christos 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)) != 0) {
232 			if ((ms->flags & MAGIC_DEBUG) != 0)
233 				(void)fprintf(stderr, "softmagic %d\n", m);
234 #ifdef BUILTIN_ELF
235 			if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
236 			    nb > 5 && fd != -1) {
237 				/*
238 				 * We matched something in the file, so this
239 				 * *might* be an ELF file, and the file is at
240 				 * least 5 bytes long, so if it's an ELF file
241 				 * it has at least one byte past the ELF magic
242 				 * number - try extracting information from the
243 				 * ELF headers that cannot easily * be
244 				 * extracted with rules in the magic file.
245 				 */
246 				if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0)
247 					if ((ms->flags & MAGIC_DEBUG) != 0)
248 						(void)fprintf(stderr,
249 						    "elf %d\n", m);
250 			}
251 #endif
252 			goto done;
253 		}
254 
255 	/* try text properties (and possibly text tokens) */
256 	if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
257 
258 		if ((m = file_ascmagic(ms, ubuf, nb)) != 0) {
259 			if ((ms->flags & MAGIC_DEBUG) != 0)
260 				(void)fprintf(stderr, "ascmagic %d\n", m);
261 			goto done;
262 		}
263 
264 		/* try to discover text encoding */
265 		if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
266 			if (looks_text == 0)
267 				if ((m = file_ascmagic_with_encoding( ms, ubuf,
268 				    nb, u8buf, ulen, code, type)) != 0) {
269 					if ((ms->flags & MAGIC_DEBUG) != 0)
270 						(void)fprintf(stderr,
271 						    "ascmagic/enc %d\n", m);
272 					goto done;
273 				}
274 		}
275 	}
276 
277 	/* give up */
278 	m = 1;
279 	if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
280 	    file_printf(ms, mime ? "application/octet-stream" : "data") == -1) {
281 	    rv = -1;
282 	}
283  done:
284 	if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
285 		if (ms->flags & MAGIC_MIME_TYPE)
286 			if (file_printf(ms, "; charset=") == -1)
287 				rv = -1;
288 		if (file_printf(ms, "%s", code_mime) == -1)
289 			rv = -1;
290 	}
291 	if (u8buf)
292 		free(u8buf);
293 	if (rv)
294 		return rv;
295 
296 	return m;
297 }
298 #endif
299 
300 protected int
301 file_reset(struct magic_set *ms)
302 {
303 	if (ms->mlist == NULL) {
304 		file_error(ms, 0, "no magic files loaded");
305 		return -1;
306 	}
307 	if (ms->o.buf) {
308 		free(ms->o.buf);
309 		ms->o.buf = NULL;
310 	}
311 	if (ms->o.pbuf) {
312 		free(ms->o.pbuf);
313 		ms->o.pbuf = NULL;
314 	}
315 	ms->event_flags &= ~EVENT_HAD_ERR;
316 	ms->error = -1;
317 	return 0;
318 }
319 
320 #define OCTALIFY(n, o)	\
321 	/*LINTED*/ \
322 	(void)(*(n)++ = '\\', \
323 	*(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
324 	*(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
325 	*(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
326 	(o)++)
327 
328 protected const char *
329 file_getbuffer(struct magic_set *ms)
330 {
331 	char *pbuf, *op, *np;
332 	size_t psize, len;
333 
334 	if (ms->event_flags & EVENT_HAD_ERR)
335 		return NULL;
336 
337 	if (ms->flags & MAGIC_RAW)
338 		return ms->o.buf;
339 
340 	if (ms->o.buf == NULL)
341 		return NULL;
342 
343 	/* * 4 is for octal representation, + 1 is for NUL */
344 	len = strlen(ms->o.buf);
345 	if (len > (SIZE_MAX - 1) / 4) {
346 		file_oomem(ms, len);
347 		return NULL;
348 	}
349 	psize = len * 4 + 1;
350 	if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
351 		file_oomem(ms, psize);
352 		return NULL;
353 	}
354 	ms->o.pbuf = pbuf;
355 
356 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
357 	{
358 		mbstate_t state;
359 		wchar_t nextchar;
360 		int mb_conv = 1;
361 		size_t bytesconsumed;
362 		char *eop;
363 		(void)memset(&state, 0, sizeof(mbstate_t));
364 
365 		np = ms->o.pbuf;
366 		op = ms->o.buf;
367 		eop = op + len;
368 
369 		while (op < eop) {
370 			bytesconsumed = mbrtowc(&nextchar, op,
371 			    (size_t)(eop - op), &state);
372 			if (bytesconsumed == (size_t)(-1) ||
373 			    bytesconsumed == (size_t)(-2)) {
374 				mb_conv = 0;
375 				break;
376 			}
377 
378 			if (iswprint(nextchar)) {
379 				(void)memcpy(np, op, bytesconsumed);
380 				op += bytesconsumed;
381 				np += bytesconsumed;
382 			} else {
383 				while (bytesconsumed-- > 0)
384 					OCTALIFY(np, op);
385 			}
386 		}
387 		*np = '\0';
388 
389 		/* Parsing succeeded as a multi-byte sequence */
390 		if (mb_conv != 0)
391 			return ms->o.pbuf;
392 	}
393 #endif
394 
395 	for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
396 		if (isprint((unsigned char)*op)) {
397 			*np++ = *op++;
398 		} else {
399 			OCTALIFY(np, op);
400 		}
401 	}
402 	*np = '\0';
403 	return ms->o.pbuf;
404 }
405 
406 protected int
407 file_check_mem(struct magic_set *ms, unsigned int level)
408 {
409 	size_t len;
410 
411 	if (level >= ms->c.len) {
412 		len = (ms->c.len += 20) * sizeof(*ms->c.li);
413 		ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
414 		    malloc(len) :
415 		    realloc(ms->c.li, len));
416 		if (ms->c.li == NULL) {
417 			file_oomem(ms, len);
418 			return -1;
419 		}
420 	}
421 	ms->c.li[level].got_match = 0;
422 #ifdef ENABLE_CONDITIONALS
423 	ms->c.li[level].last_match = 0;
424 	ms->c.li[level].last_cond = COND_NONE;
425 #endif /* ENABLE_CONDITIONALS */
426 	return 0;
427 }
428 
429 protected size_t
430 file_printedlen(const struct magic_set *ms)
431 {
432 	return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
433 }
434 
435 protected int
436 file_replace(struct magic_set *ms, const char *pat, const char *rep)
437 {
438 	regex_t rx;
439 	int rc;
440 
441 	rc = regcomp(&rx, pat, REG_EXTENDED);
442 	if (rc) {
443 		char errmsg[512];
444 		(void)regerror(rc, &rx, errmsg, sizeof(errmsg));
445 		file_magerror(ms, "regex error %d, (%s)", rc, errmsg);
446 		return -1;
447 	} else {
448 		regmatch_t rm;
449 		int nm = 0;
450 		while (regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
451 			ms->o.buf[rm.rm_so] = '\0';
452 			if (file_printf(ms, "%s%s", rep,
453 			    rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
454 				return -1;
455 			nm++;
456 		}
457 		regfree(&rx);
458 		return nm;
459 	}
460 }
461