funcs.c 11 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461
  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. #ifndef lint
  29. FILE_RCSID("@(#)$File: funcs.c,v 1.61 2012/10/30 23:11:51 christos Exp $")
  30. #endif /* lint */
  31. #include "magic.h"
  32. #include <stdarg.h>
  33. #include <stdlib.h>
  34. #include <string.h>
  35. #include <ctype.h>
  36. #if defined(HAVE_WCHAR_H)
  37. #include <wchar.h>
  38. #endif
  39. #if defined(HAVE_WCTYPE_H)
  40. #include <wctype.h>
  41. #endif
  42. #if defined(HAVE_LIMITS_H)
  43. #include <limits.h>
  44. #endif
  45. #ifndef SIZE_MAX
  46. #define SIZE_MAX ((size_t)~0)
  47. #endif
  48. /*
  49. * Like printf, only we append to a buffer.
  50. */
  51. protected int
  52. file_vprintf(struct magic_set *ms, const char *fmt, va_list ap)
  53. {
  54. int len;
  55. char *buf, *newstr;
  56. len = vasprintf(&buf, fmt, ap);
  57. if (len < 0)
  58. goto out;
  59. if (ms->o.buf != NULL) {
  60. len = asprintf(&newstr, "%s%s", ms->o.buf, buf);
  61. free(buf);
  62. if (len < 0)
  63. goto out;
  64. free(ms->o.buf);
  65. buf = newstr;
  66. }
  67. ms->o.buf = buf;
  68. return 0;
  69. out:
  70. file_error(ms, errno, "vasprintf failed");
  71. return -1;
  72. }
  73. protected int
  74. file_printf(struct magic_set *ms, const char *fmt, ...)
  75. {
  76. int rv;
  77. va_list ap;
  78. va_start(ap, fmt);
  79. rv = file_vprintf(ms, fmt, ap);
  80. va_end(ap);
  81. return rv;
  82. }
  83. /*
  84. * error - print best error message possible
  85. */
  86. /*VARARGS*/
  87. private void
  88. file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
  89. size_t lineno)
  90. {
  91. /* Only the first error is ok */
  92. if (ms->event_flags & EVENT_HAD_ERR)
  93. return;
  94. if (lineno != 0) {
  95. free(ms->o.buf);
  96. ms->o.buf = NULL;
  97. file_printf(ms, "line %" SIZE_T_FORMAT "u: ", lineno);
  98. }
  99. file_vprintf(ms, f, va);
  100. if (error > 0)
  101. file_printf(ms, " (%s)", strerror(error));
  102. ms->event_flags |= EVENT_HAD_ERR;
  103. ms->error = error;
  104. }
  105. /*VARARGS*/
  106. protected void
  107. file_error(struct magic_set *ms, int error, const char *f, ...)
  108. {
  109. va_list va;
  110. va_start(va, f);
  111. file_error_core(ms, error, f, va, 0);
  112. va_end(va);
  113. }
  114. /*
  115. * Print an error with magic line number.
  116. */
  117. /*VARARGS*/
  118. protected void
  119. file_magerror(struct magic_set *ms, const char *f, ...)
  120. {
  121. va_list va;
  122. va_start(va, f);
  123. file_error_core(ms, 0, f, va, ms->line);
  124. va_end(va);
  125. }
  126. protected void
  127. file_oomem(struct magic_set *ms, size_t len)
  128. {
  129. file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
  130. len);
  131. }
  132. protected void
  133. file_badseek(struct magic_set *ms)
  134. {
  135. file_error(ms, errno, "error seeking");
  136. }
  137. protected void
  138. file_badread(struct magic_set *ms)
  139. {
  140. file_error(ms, errno, "error reading");
  141. }
  142. #ifndef COMPILE_ONLY
  143. protected int
  144. file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((unused)),
  145. const void *buf, size_t nb)
  146. {
  147. int m = 0, rv = 0, looks_text = 0;
  148. int mime = ms->flags & MAGIC_MIME;
  149. const unsigned char *ubuf = CAST(const unsigned char *, buf);
  150. unichar *u8buf = NULL;
  151. size_t ulen;
  152. const char *code = NULL;
  153. const char *code_mime = "binary";
  154. const char *type = NULL;
  155. if (nb == 0) {
  156. if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
  157. file_printf(ms, mime ? "application/x-empty" :
  158. "empty") == -1)
  159. return -1;
  160. return 1;
  161. } else if (nb == 1) {
  162. if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
  163. file_printf(ms, mime ? "application/octet-stream" :
  164. "very short file (no magic)") == -1)
  165. return -1;
  166. return 1;
  167. }
  168. if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
  169. looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen,
  170. &code, &code_mime, &type);
  171. }
  172. #ifdef __EMX__
  173. if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
  174. switch (file_os2_apptype(ms, inname, buf, nb)) {
  175. case -1:
  176. return -1;
  177. case 0:
  178. break;
  179. default:
  180. return 1;
  181. }
  182. }
  183. #endif
  184. #if HAVE_FORK
  185. /* try compression stuff */
  186. if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0)
  187. if ((m = file_zmagic(ms, fd, inname, ubuf, nb)) != 0) {
  188. if ((ms->flags & MAGIC_DEBUG) != 0)
  189. (void)fprintf(stderr, "zmagic %d\n", m);
  190. goto done;
  191. }
  192. #endif
  193. /* Check if we have a tar file */
  194. if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0)
  195. if ((m = file_is_tar(ms, ubuf, nb)) != 0) {
  196. if ((ms->flags & MAGIC_DEBUG) != 0)
  197. (void)fprintf(stderr, "tar %d\n", m);
  198. goto done;
  199. }
  200. /* Check if we have a CDF file */
  201. if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0)
  202. if ((m = file_trycdf(ms, fd, ubuf, nb)) != 0) {
  203. if ((ms->flags & MAGIC_DEBUG) != 0)
  204. (void)fprintf(stderr, "cdf %d\n", m);
  205. goto done;
  206. }
  207. /* try soft magic tests */
  208. if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0)
  209. if ((m = file_softmagic(ms, ubuf, nb, BINTEST,
  210. looks_text)) != 0) {
  211. if ((ms->flags & MAGIC_DEBUG) != 0)
  212. (void)fprintf(stderr, "softmagic %d\n", m);
  213. #ifdef BUILTIN_ELF
  214. if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
  215. nb > 5 && fd != -1) {
  216. /*
  217. * We matched something in the file, so this
  218. * *might* be an ELF file, and the file is at
  219. * least 5 bytes long, so if it's an ELF file
  220. * it has at least one byte past the ELF magic
  221. * number - try extracting information from the
  222. * ELF headers that cannot easily * be
  223. * extracted with rules in the magic file.
  224. */
  225. if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0)
  226. if ((ms->flags & MAGIC_DEBUG) != 0)
  227. (void)fprintf(stderr,
  228. "elf %d\n", m);
  229. }
  230. #endif
  231. goto done;
  232. }
  233. /* try text properties */
  234. if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
  235. if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) {
  236. if ((ms->flags & MAGIC_DEBUG) != 0)
  237. (void)fprintf(stderr, "ascmagic %d\n", m);
  238. goto done;
  239. }
  240. /* try to discover text encoding */
  241. if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
  242. if (looks_text == 0)
  243. if ((m = file_ascmagic_with_encoding( ms, ubuf,
  244. nb, u8buf, ulen, code, type, looks_text))
  245. != 0) {
  246. if ((ms->flags & MAGIC_DEBUG) != 0)
  247. (void)fprintf(stderr,
  248. "ascmagic/enc %d\n", m);
  249. goto done;
  250. }
  251. }
  252. }
  253. /* give up */
  254. m = 1;
  255. if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
  256. file_printf(ms, mime ? "application/octet-stream" : "data") == -1) {
  257. rv = -1;
  258. }
  259. done:
  260. if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
  261. if (ms->flags & MAGIC_MIME_TYPE)
  262. if (file_printf(ms, "; charset=") == -1)
  263. rv = -1;
  264. if (file_printf(ms, "%s", code_mime) == -1)
  265. rv = -1;
  266. }
  267. free(u8buf);
  268. if (rv)
  269. return rv;
  270. return m;
  271. }
  272. #endif
  273. protected int
  274. file_reset(struct magic_set *ms)
  275. {
  276. if (ms->mlist[0] == NULL) {
  277. file_error(ms, 0, "no magic files loaded");
  278. return -1;
  279. }
  280. if (ms->o.buf) {
  281. free(ms->o.buf);
  282. ms->o.buf = NULL;
  283. }
  284. if (ms->o.pbuf) {
  285. free(ms->o.pbuf);
  286. ms->o.pbuf = NULL;
  287. }
  288. ms->event_flags &= ~EVENT_HAD_ERR;
  289. ms->error = -1;
  290. return 0;
  291. }
  292. #define OCTALIFY(n, o) \
  293. /*LINTED*/ \
  294. (void)(*(n)++ = '\\', \
  295. *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
  296. *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
  297. *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
  298. (o)++)
  299. protected const char *
  300. file_getbuffer(struct magic_set *ms)
  301. {
  302. char *pbuf, *op, *np;
  303. size_t psize, len;
  304. if (ms->event_flags & EVENT_HAD_ERR)
  305. return NULL;
  306. if (ms->flags & MAGIC_RAW)
  307. return ms->o.buf;
  308. if (ms->o.buf == NULL)
  309. return NULL;
  310. /* * 4 is for octal representation, + 1 is for NUL */
  311. len = strlen(ms->o.buf);
  312. if (len > (SIZE_MAX - 1) / 4) {
  313. file_oomem(ms, len);
  314. return NULL;
  315. }
  316. psize = len * 4 + 1;
  317. if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
  318. file_oomem(ms, psize);
  319. return NULL;
  320. }
  321. ms->o.pbuf = pbuf;
  322. #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
  323. {
  324. mbstate_t state;
  325. wchar_t nextchar;
  326. int mb_conv = 1;
  327. size_t bytesconsumed;
  328. char *eop;
  329. (void)memset(&state, 0, sizeof(mbstate_t));
  330. np = ms->o.pbuf;
  331. op = ms->o.buf;
  332. eop = op + len;
  333. while (op < eop) {
  334. bytesconsumed = mbrtowc(&nextchar, op,
  335. (size_t)(eop - op), &state);
  336. if (bytesconsumed == (size_t)(-1) ||
  337. bytesconsumed == (size_t)(-2)) {
  338. mb_conv = 0;
  339. break;
  340. }
  341. if (iswprint(nextchar)) {
  342. (void)memcpy(np, op, bytesconsumed);
  343. op += bytesconsumed;
  344. np += bytesconsumed;
  345. } else {
  346. while (bytesconsumed-- > 0)
  347. OCTALIFY(np, op);
  348. }
  349. }
  350. *np = '\0';
  351. /* Parsing succeeded as a multi-byte sequence */
  352. if (mb_conv != 0)
  353. return ms->o.pbuf;
  354. }
  355. #endif
  356. for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
  357. if (isprint((unsigned char)*op)) {
  358. *np++ = *op++;
  359. } else {
  360. OCTALIFY(np, op);
  361. }
  362. }
  363. *np = '\0';
  364. return ms->o.pbuf;
  365. }
  366. protected int
  367. file_check_mem(struct magic_set *ms, unsigned int level)
  368. {
  369. size_t len;
  370. if (level >= ms->c.len) {
  371. len = (ms->c.len += 20) * sizeof(*ms->c.li);
  372. ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
  373. malloc(len) :
  374. realloc(ms->c.li, len));
  375. if (ms->c.li == NULL) {
  376. file_oomem(ms, len);
  377. return -1;
  378. }
  379. }
  380. ms->c.li[level].got_match = 0;
  381. #ifdef ENABLE_CONDITIONALS
  382. ms->c.li[level].last_match = 0;
  383. ms->c.li[level].last_cond = COND_NONE;
  384. #endif /* ENABLE_CONDITIONALS */
  385. return 0;
  386. }
  387. protected size_t
  388. file_printedlen(const struct magic_set *ms)
  389. {
  390. return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
  391. }
  392. protected int
  393. file_replace(struct magic_set *ms, const char *pat, const char *rep)
  394. {
  395. regex_t rx;
  396. int rc;
  397. rc = regcomp(&rx, pat, REG_EXTENDED);
  398. if (rc) {
  399. char errmsg[512];
  400. (void)regerror(rc, &rx, errmsg, sizeof(errmsg));
  401. file_magerror(ms, "regex error %d, (%s)", rc, errmsg);
  402. return -1;
  403. } else {
  404. regmatch_t rm;
  405. int nm = 0;
  406. while (regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
  407. ms->o.buf[rm.rm_so] = '\0';
  408. if (file_printf(ms, "%s%s", rep,
  409. rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
  410. return -1;
  411. nm++;
  412. }
  413. regfree(&rx);
  414. return nm;
  415. }
  416. }