funcs.c revision b4f95287e7ad4605d8433474cac175e666872738
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
30FILE_RCSID("@(#)$File: funcs.c,v 1.83 2015/06/16 14:17:37 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
49#ifndef SIZE_MAX
50#define SIZE_MAX	((size_t)~0)
51#endif
52
53/*
54 * Like printf, only we append to a buffer.
55 */
56protected int
57file_vprintf(struct magic_set *ms, const char *fmt, va_list ap)
58{
59	int len;
60	char *buf, *newstr;
61
62	if (ms->event_flags & EVENT_HAD_ERR)
63		return 0;
64	len = vasprintf(&buf, fmt, ap);
65	if (len < 0)
66		goto out;
67
68	if (ms->o.buf != NULL) {
69		len = asprintf(&newstr, "%s%s", ms->o.buf, buf);
70		free(buf);
71		if (len < 0)
72			goto out;
73		free(ms->o.buf);
74		buf = newstr;
75	}
76	ms->o.buf = buf;
77	return 0;
78out:
79	file_error(ms, errno, "vasprintf failed");
80	return -1;
81}
82
83protected int
84file_printf(struct magic_set *ms, const char *fmt, ...)
85{
86	int rv;
87	va_list ap;
88
89	va_start(ap, fmt);
90	rv = file_vprintf(ms, fmt, ap);
91	va_end(ap);
92	return rv;
93}
94
95/*
96 * error - print best error message possible
97 */
98/*VARARGS*/
99__attribute__((__format__(__printf__, 3, 0)))
100private void
101file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
102    size_t lineno)
103{
104	/* Only the first error is ok */
105	if (ms->event_flags & EVENT_HAD_ERR)
106		return;
107	if (lineno != 0) {
108		free(ms->o.buf);
109		ms->o.buf = NULL;
110		file_printf(ms, "line %" SIZE_T_FORMAT "u:", lineno);
111	}
112	if (ms->o.buf && *ms->o.buf)
113		file_printf(ms, " ");
114	file_vprintf(ms, f, va);
115	if (error > 0)
116		file_printf(ms, " (%s)", strerror(error));
117	ms->event_flags |= EVENT_HAD_ERR;
118	ms->error = error;
119}
120
121/*VARARGS*/
122protected void
123file_error(struct magic_set *ms, int error, const char *f, ...)
124{
125	va_list va;
126	va_start(va, f);
127	file_error_core(ms, error, f, va, 0);
128	va_end(va);
129}
130
131/*
132 * Print an error with magic line number.
133 */
134/*VARARGS*/
135protected void
136file_magerror(struct magic_set *ms, const char *f, ...)
137{
138	va_list va;
139	va_start(va, f);
140	file_error_core(ms, 0, f, va, ms->line);
141	va_end(va);
142}
143
144protected void
145file_oomem(struct magic_set *ms, size_t len)
146{
147	file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
148	    len);
149}
150
151protected void
152file_badseek(struct magic_set *ms)
153{
154	file_error(ms, errno, "error seeking");
155}
156
157protected void
158file_badread(struct magic_set *ms)
159{
160	file_error(ms, errno, "error reading");
161}
162
163#ifndef COMPILE_ONLY
164
165static int
166checkdone(struct magic_set *ms, int *rv)
167{
168	if ((ms->flags & MAGIC_CONTINUE) == 0)
169		return 1;
170	if (file_printf(ms, "\n- ") == -1)
171		*rv = -1;
172	return 0;
173}
174
175/*ARGSUSED*/
176protected int
177file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((__unused__)),
178    const void *buf, size_t nb)
179{
180	int m = 0, rv = 0, looks_text = 0;
181	int mime = ms->flags & MAGIC_MIME;
182	const unsigned char *ubuf = CAST(const unsigned char *, buf);
183	unichar *u8buf = NULL;
184	size_t ulen;
185	const char *code = NULL;
186	const char *code_mime = "binary";
187	const char *type = "application/octet-stream";
188	const char *def = "data";
189	const char *ftype = NULL;
190
191	if (nb == 0) {
192		def = "empty";
193		type = "application/x-empty";
194		goto simple;
195	} else if (nb == 1) {
196		def = "very short file (no magic)";
197		goto simple;
198	}
199
200	if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
201		looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen,
202		    &code, &code_mime, &ftype);
203	}
204
205#ifdef __EMX__
206	if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
207		switch (file_os2_apptype(ms, inname, buf, nb)) {
208		case -1:
209			return -1;
210		case 0:
211			break;
212		default:
213			return 1;
214		}
215	}
216#endif
217#if HAVE_FORK
218	/* try compression stuff */
219	if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0)
220		if ((m = file_zmagic(ms, fd, inname, ubuf, nb)) != 0) {
221			if ((ms->flags & MAGIC_DEBUG) != 0)
222				(void)fprintf(stderr, "zmagic %d\n", m);
223			goto done_encoding;
224		}
225#endif
226	/* Check if we have a tar file */
227	if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0)
228		if ((m = file_is_tar(ms, ubuf, nb)) != 0) {
229			if ((ms->flags & MAGIC_DEBUG) != 0)
230				(void)fprintf(stderr, "tar %d\n", m);
231			if (checkdone(ms, &rv))
232				goto done;
233		}
234
235	/* Check if we have a CDF file */
236	if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0)
237		if ((m = file_trycdf(ms, fd, ubuf, nb)) != 0) {
238			if ((ms->flags & MAGIC_DEBUG) != 0)
239				(void)fprintf(stderr, "cdf %d\n", m);
240			if (checkdone(ms, &rv))
241				goto done;
242		}
243
244	/* try soft magic tests */
245	if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0)
246		if ((m = file_softmagic(ms, ubuf, nb, 0, NULL, BINTEST,
247		    looks_text)) != 0) {
248			if ((ms->flags & MAGIC_DEBUG) != 0)
249				(void)fprintf(stderr, "softmagic %d\n", m);
250#ifdef BUILTIN_ELF
251			if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
252			    nb > 5 && fd != -1) {
253				/*
254				 * We matched something in the file, so this
255				 * *might* be an ELF file, and the file is at
256				 * least 5 bytes long, so if it's an ELF file
257				 * it has at least one byte past the ELF magic
258				 * number - try extracting information from the
259				 * ELF headers that cannot easily * be
260				 * extracted with rules in the magic file.
261				 */
262				if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0)
263					if ((ms->flags & MAGIC_DEBUG) != 0)
264						(void)fprintf(stderr,
265						    "elf %d\n", m);
266			}
267#endif
268			if (checkdone(ms, &rv))
269				goto done;
270		}
271
272	/* try text properties */
273	if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
274
275		if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) {
276			if ((ms->flags & MAGIC_DEBUG) != 0)
277				(void)fprintf(stderr, "ascmagic %d\n", m);
278			if (checkdone(ms, &rv))
279				goto done;
280		}
281	}
282
283simple:
284	/* give up */
285	m = 1;
286	if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
287	    file_printf(ms, "%s", mime ? type : def) == -1) {
288	    rv = -1;
289	}
290 done:
291	if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
292		if (ms->flags & MAGIC_MIME_TYPE)
293			if (file_printf(ms, "; charset=") == -1)
294				rv = -1;
295		if (file_printf(ms, "%s", code_mime) == -1)
296			rv = -1;
297	}
298#if HAVE_FORK
299 done_encoding:
300#endif
301	free(u8buf);
302	if (rv)
303		return rv;
304
305	return m;
306}
307#endif
308
309protected int
310file_reset(struct magic_set *ms)
311{
312	if (ms->mlist[0] == NULL) {
313		file_error(ms, 0, "no magic files loaded");
314		return -1;
315	}
316	if (ms->o.buf) {
317		free(ms->o.buf);
318		ms->o.buf = NULL;
319	}
320	if (ms->o.pbuf) {
321		free(ms->o.pbuf);
322		ms->o.pbuf = NULL;
323	}
324	ms->event_flags &= ~EVENT_HAD_ERR;
325	ms->error = -1;
326	return 0;
327}
328
329#define OCTALIFY(n, o)	\
330	/*LINTED*/ \
331	(void)(*(n)++ = '\\', \
332	*(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
333	*(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
334	*(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
335	(o)++)
336
337protected const char *
338file_getbuffer(struct magic_set *ms)
339{
340	char *pbuf, *op, *np;
341	size_t psize, len;
342
343	if (ms->event_flags & EVENT_HAD_ERR)
344		return NULL;
345
346	if (ms->flags & MAGIC_RAW)
347		return ms->o.buf;
348
349	if (ms->o.buf == NULL)
350		return NULL;
351
352	/* * 4 is for octal representation, + 1 is for NUL */
353	len = strlen(ms->o.buf);
354	if (len > (SIZE_MAX - 1) / 4) {
355		file_oomem(ms, len);
356		return NULL;
357	}
358	psize = len * 4 + 1;
359	if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
360		file_oomem(ms, psize);
361		return NULL;
362	}
363	ms->o.pbuf = pbuf;
364
365#if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
366	{
367		mbstate_t state;
368		wchar_t nextchar;
369		int mb_conv = 1;
370		size_t bytesconsumed;
371		char *eop;
372		(void)memset(&state, 0, sizeof(mbstate_t));
373
374		np = ms->o.pbuf;
375		op = ms->o.buf;
376		eop = op + len;
377
378		while (op < eop) {
379			bytesconsumed = mbrtowc(&nextchar, op,
380			    (size_t)(eop - op), &state);
381			if (bytesconsumed == (size_t)(-1) ||
382			    bytesconsumed == (size_t)(-2)) {
383				mb_conv = 0;
384				break;
385			}
386
387			if (iswprint(nextchar)) {
388				(void)memcpy(np, op, bytesconsumed);
389				op += bytesconsumed;
390				np += bytesconsumed;
391			} else {
392				while (bytesconsumed-- > 0)
393					OCTALIFY(np, op);
394			}
395		}
396		*np = '\0';
397
398		/* Parsing succeeded as a multi-byte sequence */
399		if (mb_conv != 0)
400			return ms->o.pbuf;
401	}
402#endif
403
404	for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
405		if (isprint((unsigned char)*op)) {
406			*np++ = *op++;
407		} else {
408			OCTALIFY(np, op);
409		}
410	}
411	*np = '\0';
412	return ms->o.pbuf;
413}
414
415protected int
416file_check_mem(struct magic_set *ms, unsigned int level)
417{
418	size_t len;
419
420	if (level >= ms->c.len) {
421		len = (ms->c.len = 20 + level) * sizeof(*ms->c.li);
422		ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
423		    malloc(len) :
424		    realloc(ms->c.li, len));
425		if (ms->c.li == NULL) {
426			file_oomem(ms, len);
427			return -1;
428		}
429	}
430	ms->c.li[level].got_match = 0;
431#ifdef ENABLE_CONDITIONALS
432	ms->c.li[level].last_match = 0;
433	ms->c.li[level].last_cond = COND_NONE;
434#endif /* ENABLE_CONDITIONALS */
435	return 0;
436}
437
438protected size_t
439file_printedlen(const struct magic_set *ms)
440{
441	return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
442}
443
444protected int
445file_replace(struct magic_set *ms, const char *pat, const char *rep)
446{
447	file_regex_t rx;
448	int rc, rv = -1;
449
450	rc = file_regcomp(&rx, pat, REG_EXTENDED);
451	if (rc) {
452		file_regerror(&rx, rc, ms);
453	} else {
454		regmatch_t rm;
455		int nm = 0;
456		while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
457			ms->o.buf[rm.rm_so] = '\0';
458			if (file_printf(ms, "%s%s", rep,
459			    rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
460				goto out;
461			nm++;
462		}
463		rv = nm;
464	}
465out:
466	file_regfree(&rx);
467	return rv;
468}
469
470protected int
471file_regcomp(file_regex_t *rx, const char *pat, int flags)
472{
473#ifdef USE_C_LOCALE
474	rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0);
475	assert(rx->c_lc_ctype != NULL);
476	rx->old_lc_ctype = uselocale(rx->c_lc_ctype);
477	assert(rx->old_lc_ctype != NULL);
478#endif
479	rx->pat = pat;
480
481	return rx->rc = regcomp(&rx->rx, pat, flags);
482}
483
484protected int
485file_regexec(file_regex_t *rx, const char *str, size_t nmatch,
486    regmatch_t* pmatch, int eflags)
487{
488	assert(rx->rc == 0);
489	return regexec(&rx->rx, str, nmatch, pmatch, eflags);
490}
491
492protected void
493file_regfree(file_regex_t *rx)
494{
495	if (rx->rc == 0)
496		regfree(&rx->rx);
497#ifdef USE_C_LOCALE
498	(void)uselocale(rx->old_lc_ctype);
499	freelocale(rx->c_lc_ctype);
500#endif
501}
502
503protected void
504file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
505{
506	char errmsg[512];
507
508	(void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
509	file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
510	    errmsg);
511}
512
513protected file_pushbuf_t *
514file_push_buffer(struct magic_set *ms)
515{
516	file_pushbuf_t *pb;
517
518	if (ms->event_flags & EVENT_HAD_ERR)
519		return NULL;
520
521	if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL)
522		return NULL;
523
524	pb->buf = ms->o.buf;
525	pb->offset = ms->offset;
526
527	ms->o.buf = NULL;
528	ms->offset = 0;
529
530	return pb;
531}
532
533protected char *
534file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
535{
536	char *rbuf;
537
538	if (ms->event_flags & EVENT_HAD_ERR) {
539		free(pb->buf);
540		free(pb);
541		return NULL;
542	}
543
544	rbuf = ms->o.buf;
545
546	ms->o.buf = pb->buf;
547	ms->offset = pb->offset;
548
549	free(pb);
550	return rbuf;
551}
552
553/*
554 * convert string to ascii printable format.
555 */
556protected char *
557file_printable(char *buf, size_t bufsiz, const char *str)
558{
559	char *ptr, *eptr;
560	const unsigned char *s = (const unsigned char *)str;
561
562	for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) {
563		if (isprint(*s)) {
564			*ptr++ = *s;
565			continue;
566		}
567		if (ptr >= eptr - 3)
568			break;
569		*ptr++ = '\\';
570		*ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0';
571		*ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0';
572		*ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0';
573	}
574	*ptr = '\0';
575	return buf;
576}
577