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