funcs.c revision ae66b4b7f2cfa501c0f1c60e016d70d68170f7d5
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 1.72 2014/05/14 23:15:42 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 */
59protected int
60file_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;
81out:
82	file_error(ms, errno, "vasprintf failed");
83	return -1;
84}
85
86protected int
87file_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)))
103private void
104file_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*/
123protected void
124file_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*/
136protected void
137file_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
145protected void
146file_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
152protected void
153file_badseek(struct magic_set *ms)
154{
155	file_error(ms, errno, "error seeking");
156}
157
158protected void
159file_badread(struct magic_set *ms)
160{
161	file_error(ms, errno, "error reading");
162}
163
164#ifndef COMPILE_ONLY
165protected int
166file_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
268simple:
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#if HAVE_FORK
284 done_encoding:
285#endif
286	free(u8buf);
287	if (rv)
288		return rv;
289
290	return m;
291}
292#endif
293
294protected int
295file_reset(struct magic_set *ms)
296{
297	if (ms->mlist[0] == NULL) {
298		file_error(ms, 0, "no magic files loaded");
299		return -1;
300	}
301	if (ms->o.buf) {
302		free(ms->o.buf);
303		ms->o.buf = NULL;
304	}
305	if (ms->o.pbuf) {
306		free(ms->o.pbuf);
307		ms->o.pbuf = NULL;
308	}
309	ms->event_flags &= ~EVENT_HAD_ERR;
310	ms->error = -1;
311	return 0;
312}
313
314#define OCTALIFY(n, o)	\
315	/*LINTED*/ \
316	(void)(*(n)++ = '\\', \
317	*(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
318	*(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
319	*(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
320	(o)++)
321
322protected const char *
323file_getbuffer(struct magic_set *ms)
324{
325	char *pbuf, *op, *np;
326	size_t psize, len;
327
328	if (ms->event_flags & EVENT_HAD_ERR)
329		return NULL;
330
331	if (ms->flags & MAGIC_RAW)
332		return ms->o.buf;
333
334	if (ms->o.buf == NULL)
335		return NULL;
336
337	/* * 4 is for octal representation, + 1 is for NUL */
338	len = strlen(ms->o.buf);
339	if (len > (SIZE_MAX - 1) / 4) {
340		file_oomem(ms, len);
341		return NULL;
342	}
343	psize = len * 4 + 1;
344	if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
345		file_oomem(ms, psize);
346		return NULL;
347	}
348	ms->o.pbuf = pbuf;
349
350#if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
351	{
352		mbstate_t state;
353		wchar_t nextchar;
354		int mb_conv = 1;
355		size_t bytesconsumed;
356		char *eop;
357		(void)memset(&state, 0, sizeof(mbstate_t));
358
359		np = ms->o.pbuf;
360		op = ms->o.buf;
361		eop = op + len;
362
363		while (op < eop) {
364			bytesconsumed = mbrtowc(&nextchar, op,
365			    (size_t)(eop - op), &state);
366			if (bytesconsumed == (size_t)(-1) ||
367			    bytesconsumed == (size_t)(-2)) {
368				mb_conv = 0;
369				break;
370			}
371
372			if (iswprint(nextchar)) {
373				(void)memcpy(np, op, bytesconsumed);
374				op += bytesconsumed;
375				np += bytesconsumed;
376			} else {
377				while (bytesconsumed-- > 0)
378					OCTALIFY(np, op);
379			}
380		}
381		*np = '\0';
382
383		/* Parsing succeeded as a multi-byte sequence */
384		if (mb_conv != 0)
385			return ms->o.pbuf;
386	}
387#endif
388
389	for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
390		if (isprint((unsigned char)*op)) {
391			*np++ = *op++;
392		} else {
393			OCTALIFY(np, op);
394		}
395	}
396	*np = '\0';
397	return ms->o.pbuf;
398}
399
400protected int
401file_check_mem(struct magic_set *ms, unsigned int level)
402{
403	size_t len;
404
405	if (level >= ms->c.len) {
406		len = (ms->c.len += 20) * sizeof(*ms->c.li);
407		ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
408		    malloc(len) :
409		    realloc(ms->c.li, len));
410		if (ms->c.li == NULL) {
411			file_oomem(ms, len);
412			return -1;
413		}
414	}
415	ms->c.li[level].got_match = 0;
416#ifdef ENABLE_CONDITIONALS
417	ms->c.li[level].last_match = 0;
418	ms->c.li[level].last_cond = COND_NONE;
419#endif /* ENABLE_CONDITIONALS */
420	return 0;
421}
422
423protected size_t
424file_printedlen(const struct magic_set *ms)
425{
426	return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
427}
428
429protected int
430file_replace(struct magic_set *ms, const char *pat, const char *rep)
431{
432	file_regex_t rx;
433	int rc, rv = -1;
434
435	rc = file_regcomp(&rx, pat, REG_EXTENDED);
436	if (rc) {
437		file_regerror(&rx, rc, ms);
438	} else {
439		regmatch_t rm;
440		int nm = 0;
441		while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
442			ms->o.buf[rm.rm_so] = '\0';
443			if (file_printf(ms, "%s%s", rep,
444			    rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
445				goto out;
446			nm++;
447		}
448		rv = nm;
449	}
450out:
451	file_regfree(&rx);
452	return rv;
453}
454
455protected int
456file_regcomp(file_regex_t *rx, const char *pat, int flags)
457{
458	rx->old_lc_ctype = setlocale(LC_CTYPE, NULL);
459	assert(rx->old_lc_ctype != NULL);
460	rx->old_lc_ctype = strdup(rx->old_lc_ctype);
461	assert(rx->old_lc_ctype != NULL);
462	rx->pat = pat;
463
464	(void)setlocale(LC_CTYPE, "C");
465	return rx->rc = regcomp(&rx->rx, pat, flags);
466}
467
468protected int
469file_regexec(file_regex_t *rx, const char *str, size_t nmatch,
470    regmatch_t* pmatch, int eflags)
471{
472	assert(rx->rc == 0);
473	return regexec(&rx->rx, str, nmatch, pmatch, eflags);
474}
475
476protected void
477file_regfree(file_regex_t *rx)
478{
479	if (rx->rc == 0)
480		regfree(&rx->rx);
481	(void)setlocale(LC_CTYPE, rx->old_lc_ctype);
482	free(rx->old_lc_ctype);
483}
484
485protected void
486file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
487{
488	char errmsg[512];
489
490	(void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
491	file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
492	    errmsg);
493}
494
495protected file_pushbuf_t *
496file_push_buffer(struct magic_set *ms)
497{
498	file_pushbuf_t *pb;
499
500	if (ms->event_flags & EVENT_HAD_ERR)
501		return NULL;
502
503	if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL)
504		return NULL;
505
506	pb->buf = ms->o.buf;
507	pb->offset = ms->offset;
508
509	ms->o.buf = NULL;
510	ms->offset = 0;
511
512	return pb;
513}
514
515protected char *
516file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
517{
518	char *rbuf;
519
520	if (ms->event_flags & EVENT_HAD_ERR) {
521		free(pb->buf);
522		free(pb);
523		return NULL;
524	}
525
526	rbuf = ms->o.buf;
527
528	ms->o.buf = pb->buf;
529	ms->offset = pb->offset;
530
531	free(pb);
532	return rbuf;
533}
534