xref: /dragonfly/contrib/file/src/funcs.c (revision ef3ac1d1)
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 1.70 2014/03/14 19:02:37 christos Exp $")
31 #endif	/* lint */
32 
33 #include "magic.h"
34 #include <assert.h>
35 #include <stdarg.h>
36 #include <stdlib.h>
37 #include <string.h>
38 #include <ctype.h>
39 #if defined(HAVE_WCHAR_H)
40 #include <wchar.h>
41 #endif
42 #if defined(HAVE_WCTYPE_H)
43 #include <wctype.h>
44 #endif
45 #if defined(HAVE_LIMITS_H)
46 #include <limits.h>
47 #endif
48 #if defined(HAVE_LOCALE_H)
49 #include <locale.h>
50 #endif
51 
52 #ifndef SIZE_MAX
53 #define SIZE_MAX	((size_t)~0)
54 #endif
55 
56 /*
57  * Like printf, only we append to a buffer.
58  */
59 protected int
60 file_vprintf(struct magic_set *ms, const char *fmt, va_list ap)
61 {
62 	int len;
63 	char *buf, *newstr;
64 
65 	if (ms->event_flags & EVENT_HAD_ERR)
66 		return 0;
67 	len = vasprintf(&buf, fmt, ap);
68 	if (len < 0)
69 		goto out;
70 
71 	if (ms->o.buf != NULL) {
72 		len = asprintf(&newstr, "%s%s", ms->o.buf, buf);
73 		free(buf);
74 		if (len < 0)
75 			goto out;
76 		free(ms->o.buf);
77 		buf = newstr;
78 	}
79 	ms->o.buf = buf;
80 	return 0;
81 out:
82 	file_error(ms, errno, "vasprintf failed");
83 	return -1;
84 }
85 
86 protected int
87 file_printf(struct magic_set *ms, const char *fmt, ...)
88 {
89 	int rv;
90 	va_list ap;
91 
92 	va_start(ap, fmt);
93 	rv = file_vprintf(ms, fmt, ap);
94 	va_end(ap);
95 	return rv;
96 }
97 
98 /*
99  * error - print best error message possible
100  */
101 /*VARARGS*/
102 __attribute__((__format__(__printf__, 3, 0)))
103 private void
104 file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
105     size_t lineno)
106 {
107 	/* Only the first error is ok */
108 	if (ms->event_flags & EVENT_HAD_ERR)
109 		return;
110 	if (lineno != 0) {
111 		free(ms->o.buf);
112 		ms->o.buf = NULL;
113 		file_printf(ms, "line %" SIZE_T_FORMAT "u: ", lineno);
114 	}
115 	file_vprintf(ms, f, va);
116 	if (error > 0)
117 		file_printf(ms, " (%s)", strerror(error));
118 	ms->event_flags |= EVENT_HAD_ERR;
119 	ms->error = error;
120 }
121 
122 /*VARARGS*/
123 protected void
124 file_error(struct magic_set *ms, int error, const char *f, ...)
125 {
126 	va_list va;
127 	va_start(va, f);
128 	file_error_core(ms, error, f, va, 0);
129 	va_end(va);
130 }
131 
132 /*
133  * Print an error with magic line number.
134  */
135 /*VARARGS*/
136 protected void
137 file_magerror(struct magic_set *ms, const char *f, ...)
138 {
139 	va_list va;
140 	va_start(va, f);
141 	file_error_core(ms, 0, f, va, ms->line);
142 	va_end(va);
143 }
144 
145 protected void
146 file_oomem(struct magic_set *ms, size_t len)
147 {
148 	file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
149 	    len);
150 }
151 
152 protected void
153 file_badseek(struct magic_set *ms)
154 {
155 	file_error(ms, errno, "error seeking");
156 }
157 
158 protected void
159 file_badread(struct magic_set *ms)
160 {
161 	file_error(ms, errno, "error reading");
162 }
163 
164 #ifndef COMPILE_ONLY
165 protected int
166 file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((unused)),
167     const void *buf, size_t nb)
168 {
169 	int m = 0, rv = 0, looks_text = 0;
170 	int mime = ms->flags & MAGIC_MIME;
171 	const unsigned char *ubuf = CAST(const unsigned char *, buf);
172 	unichar *u8buf = NULL;
173 	size_t ulen;
174 	const char *code = NULL;
175 	const char *code_mime = "binary";
176 	const char *type = "application/octet-stream";
177 	const char *def = "data";
178 	const char *ftype = NULL;
179 
180 	if (nb == 0) {
181 		def = "empty";
182 		type = "application/x-empty";
183 		goto simple;
184 	} else if (nb == 1) {
185 		def = "very short file (no magic)";
186 		goto simple;
187 	}
188 
189 	if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
190 		looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen,
191 		    &code, &code_mime, &ftype);
192 	}
193 
194 #ifdef __EMX__
195 	if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
196 		switch (file_os2_apptype(ms, inname, buf, nb)) {
197 		case -1:
198 			return -1;
199 		case 0:
200 			break;
201 		default:
202 			return 1;
203 		}
204 	}
205 #endif
206 #if HAVE_FORK
207 	/* try compression stuff */
208 	if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0)
209 		if ((m = file_zmagic(ms, fd, inname, ubuf, nb)) != 0) {
210 			if ((ms->flags & MAGIC_DEBUG) != 0)
211 				(void)fprintf(stderr, "zmagic %d\n", m);
212 			goto done_encoding;
213 		}
214 #endif
215 	/* Check if we have a tar file */
216 	if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0)
217 		if ((m = file_is_tar(ms, ubuf, nb)) != 0) {
218 			if ((ms->flags & MAGIC_DEBUG) != 0)
219 				(void)fprintf(stderr, "tar %d\n", m);
220 			goto done;
221 		}
222 
223 	/* Check if we have a CDF file */
224 	if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0)
225 		if ((m = file_trycdf(ms, fd, ubuf, nb)) != 0) {
226 			if ((ms->flags & MAGIC_DEBUG) != 0)
227 				(void)fprintf(stderr, "cdf %d\n", m);
228 			goto done;
229 		}
230 
231 	/* try soft magic tests */
232 	if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0)
233 		if ((m = file_softmagic(ms, ubuf, nb, 0, BINTEST,
234 		    looks_text)) != 0) {
235 			if ((ms->flags & MAGIC_DEBUG) != 0)
236 				(void)fprintf(stderr, "softmagic %d\n", m);
237 #ifdef BUILTIN_ELF
238 			if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
239 			    nb > 5 && fd != -1) {
240 				/*
241 				 * We matched something in the file, so this
242 				 * *might* be an ELF file, and the file is at
243 				 * least 5 bytes long, so if it's an ELF file
244 				 * it has at least one byte past the ELF magic
245 				 * number - try extracting information from the
246 				 * ELF headers that cannot easily * be
247 				 * extracted with rules in the magic file.
248 				 */
249 				if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0)
250 					if ((ms->flags & MAGIC_DEBUG) != 0)
251 						(void)fprintf(stderr,
252 						    "elf %d\n", m);
253 			}
254 #endif
255 			goto done;
256 		}
257 
258 	/* try text properties */
259 	if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
260 
261 		if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) {
262 			if ((ms->flags & MAGIC_DEBUG) != 0)
263 				(void)fprintf(stderr, "ascmagic %d\n", m);
264 			goto done;
265 		}
266 	}
267 
268 simple:
269 	/* give up */
270 	m = 1;
271 	if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
272 	    file_printf(ms, "%s", mime ? type : def) == -1) {
273 	    rv = -1;
274 	}
275  done:
276 	if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
277 		if (ms->flags & MAGIC_MIME_TYPE)
278 			if (file_printf(ms, "; charset=") == -1)
279 				rv = -1;
280 		if (file_printf(ms, "%s", code_mime) == -1)
281 			rv = -1;
282 	}
283  done_encoding:
284 	free(u8buf);
285 	if (rv)
286 		return rv;
287 
288 	return m;
289 }
290 #endif
291 
292 protected int
293 file_reset(struct magic_set *ms)
294 {
295 	if (ms->mlist[0] == NULL) {
296 		file_error(ms, 0, "no magic files loaded");
297 		return -1;
298 	}
299 	if (ms->o.buf) {
300 		free(ms->o.buf);
301 		ms->o.buf = NULL;
302 	}
303 	if (ms->o.pbuf) {
304 		free(ms->o.pbuf);
305 		ms->o.pbuf = NULL;
306 	}
307 	ms->event_flags &= ~EVENT_HAD_ERR;
308 	ms->error = -1;
309 	return 0;
310 }
311 
312 #define OCTALIFY(n, o)	\
313 	/*LINTED*/ \
314 	(void)(*(n)++ = '\\', \
315 	*(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
316 	*(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
317 	*(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
318 	(o)++)
319 
320 protected const char *
321 file_getbuffer(struct magic_set *ms)
322 {
323 	char *pbuf, *op, *np;
324 	size_t psize, len;
325 
326 	if (ms->event_flags & EVENT_HAD_ERR)
327 		return NULL;
328 
329 	if (ms->flags & MAGIC_RAW)
330 		return ms->o.buf;
331 
332 	if (ms->o.buf == NULL)
333 		return NULL;
334 
335 	/* * 4 is for octal representation, + 1 is for NUL */
336 	len = strlen(ms->o.buf);
337 	if (len > (SIZE_MAX - 1) / 4) {
338 		file_oomem(ms, len);
339 		return NULL;
340 	}
341 	psize = len * 4 + 1;
342 	if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
343 		file_oomem(ms, psize);
344 		return NULL;
345 	}
346 	ms->o.pbuf = pbuf;
347 
348 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
349 	{
350 		mbstate_t state;
351 		wchar_t nextchar;
352 		int mb_conv = 1;
353 		size_t bytesconsumed;
354 		char *eop;
355 		(void)memset(&state, 0, sizeof(mbstate_t));
356 
357 		np = ms->o.pbuf;
358 		op = ms->o.buf;
359 		eop = op + len;
360 
361 		while (op < eop) {
362 			bytesconsumed = mbrtowc(&nextchar, op,
363 			    (size_t)(eop - op), &state);
364 			if (bytesconsumed == (size_t)(-1) ||
365 			    bytesconsumed == (size_t)(-2)) {
366 				mb_conv = 0;
367 				break;
368 			}
369 
370 			if (iswprint(nextchar)) {
371 				(void)memcpy(np, op, bytesconsumed);
372 				op += bytesconsumed;
373 				np += bytesconsumed;
374 			} else {
375 				while (bytesconsumed-- > 0)
376 					OCTALIFY(np, op);
377 			}
378 		}
379 		*np = '\0';
380 
381 		/* Parsing succeeded as a multi-byte sequence */
382 		if (mb_conv != 0)
383 			return ms->o.pbuf;
384 	}
385 #endif
386 
387 	for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
388 		if (isprint((unsigned char)*op)) {
389 			*np++ = *op++;
390 		} else {
391 			OCTALIFY(np, op);
392 		}
393 	}
394 	*np = '\0';
395 	return ms->o.pbuf;
396 }
397 
398 protected int
399 file_check_mem(struct magic_set *ms, unsigned int level)
400 {
401 	size_t len;
402 
403 	if (level >= ms->c.len) {
404 		len = (ms->c.len += 20) * sizeof(*ms->c.li);
405 		ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
406 		    malloc(len) :
407 		    realloc(ms->c.li, len));
408 		if (ms->c.li == NULL) {
409 			file_oomem(ms, len);
410 			return -1;
411 		}
412 	}
413 	ms->c.li[level].got_match = 0;
414 #ifdef ENABLE_CONDITIONALS
415 	ms->c.li[level].last_match = 0;
416 	ms->c.li[level].last_cond = COND_NONE;
417 #endif /* ENABLE_CONDITIONALS */
418 	return 0;
419 }
420 
421 protected size_t
422 file_printedlen(const struct magic_set *ms)
423 {
424 	return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
425 }
426 
427 protected int
428 file_replace(struct magic_set *ms, const char *pat, const char *rep)
429 {
430 	regex_t rx;
431 	int rc, rv = -1;
432 	char *old_lc_ctype;
433 
434 	old_lc_ctype = setlocale(LC_CTYPE, NULL);
435 	assert(old_lc_ctype != NULL);
436 	old_lc_ctype = strdup(old_lc_ctype);
437 	assert(old_lc_ctype != NULL);
438 	(void)setlocale(LC_CTYPE, "C");
439 	rc = regcomp(&rx, pat, REG_EXTENDED);
440 	if (rc) {
441 		char errmsg[512];
442 		(void)regerror(rc, &rx, errmsg, sizeof(errmsg));
443 		file_magerror(ms, "regex error %d, (%s)", rc, errmsg);
444 	} else {
445 		regmatch_t rm;
446 		int nm = 0;
447 		while (regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
448 			ms->o.buf[rm.rm_so] = '\0';
449 			if (file_printf(ms, "%s%s", rep,
450 			    rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
451 				goto out;
452 			nm++;
453 		}
454 		regfree(&rx);
455 		rv = nm;
456 	}
457 out:
458 	(void)setlocale(LC_CTYPE, old_lc_ctype);
459 	free(old_lc_ctype);
460 	return rv;
461 }
462