funcs.c 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559
  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 1.79 2014/12/16 20:52:49 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. #ifndef SIZE_MAX
  47. #define SIZE_MAX ((size_t)~0)
  48. #endif
  49. /*
  50. * Like printf, only we append to a buffer.
  51. */
  52. protected int
  53. file_vprintf(struct magic_set *ms, const char *fmt, va_list ap)
  54. {
  55. int len;
  56. char *buf, *newstr;
  57. if (ms->event_flags & EVENT_HAD_ERR)
  58. return 0;
  59. len = vasprintf(&buf, fmt, ap);
  60. if (len < 0)
  61. goto out;
  62. if (ms->o.buf != NULL) {
  63. len = asprintf(&newstr, "%s%s", ms->o.buf, buf);
  64. free(buf);
  65. if (len < 0)
  66. goto out;
  67. free(ms->o.buf);
  68. buf = newstr;
  69. }
  70. ms->o.buf = buf;
  71. return 0;
  72. out:
  73. file_error(ms, errno, "vasprintf failed");
  74. return -1;
  75. }
  76. protected int
  77. file_printf(struct magic_set *ms, const char *fmt, ...)
  78. {
  79. int rv;
  80. va_list ap;
  81. va_start(ap, fmt);
  82. rv = file_vprintf(ms, fmt, ap);
  83. va_end(ap);
  84. return rv;
  85. }
  86. /*
  87. * error - print best error message possible
  88. */
  89. /*VARARGS*/
  90. __attribute__((__format__(__printf__, 3, 0)))
  91. private void
  92. file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
  93. size_t lineno)
  94. {
  95. /* Only the first error is ok */
  96. if (ms->event_flags & EVENT_HAD_ERR)
  97. return;
  98. if (lineno != 0) {
  99. free(ms->o.buf);
  100. ms->o.buf = NULL;
  101. file_printf(ms, "line %" SIZE_T_FORMAT "u: ", lineno);
  102. }
  103. file_vprintf(ms, f, va);
  104. if (error > 0)
  105. file_printf(ms, " (%s)", strerror(error));
  106. ms->event_flags |= EVENT_HAD_ERR;
  107. ms->error = error;
  108. }
  109. /*VARARGS*/
  110. protected void
  111. file_error(struct magic_set *ms, int error, const char *f, ...)
  112. {
  113. va_list va;
  114. va_start(va, f);
  115. file_error_core(ms, error, f, va, 0);
  116. va_end(va);
  117. }
  118. /*
  119. * Print an error with magic line number.
  120. */
  121. /*VARARGS*/
  122. protected void
  123. file_magerror(struct magic_set *ms, const char *f, ...)
  124. {
  125. va_list va;
  126. va_start(va, f);
  127. file_error_core(ms, 0, f, va, ms->line);
  128. va_end(va);
  129. }
  130. protected void
  131. file_oomem(struct magic_set *ms, size_t len)
  132. {
  133. file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
  134. len);
  135. }
  136. protected void
  137. file_badseek(struct magic_set *ms)
  138. {
  139. file_error(ms, errno, "error seeking");
  140. }
  141. protected void
  142. file_badread(struct magic_set *ms)
  143. {
  144. file_error(ms, errno, "error reading");
  145. }
  146. #ifndef COMPILE_ONLY
  147. /*ARGSUSED*/
  148. protected int
  149. file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((__unused__)),
  150. const void *buf, size_t nb)
  151. {
  152. int m = 0, rv = 0, looks_text = 0;
  153. int mime = ms->flags & MAGIC_MIME;
  154. const unsigned char *ubuf = CAST(const unsigned char *, buf);
  155. unichar *u8buf = NULL;
  156. size_t ulen;
  157. const char *code = NULL;
  158. const char *code_mime = "binary";
  159. const char *type = "application/octet-stream";
  160. const char *def = "data";
  161. const char *ftype = NULL;
  162. if (nb == 0) {
  163. def = "empty";
  164. type = "application/x-empty";
  165. goto simple;
  166. } else if (nb == 1) {
  167. def = "very short file (no magic)";
  168. goto simple;
  169. }
  170. if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
  171. looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen,
  172. &code, &code_mime, &ftype);
  173. }
  174. #ifdef __EMX__
  175. if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
  176. switch (file_os2_apptype(ms, inname, buf, nb)) {
  177. case -1:
  178. return -1;
  179. case 0:
  180. break;
  181. default:
  182. return 1;
  183. }
  184. }
  185. #endif
  186. #if HAVE_FORK
  187. /* try compression stuff */
  188. if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0)
  189. if ((m = file_zmagic(ms, fd, inname, ubuf, nb)) != 0) {
  190. if ((ms->flags & MAGIC_DEBUG) != 0)
  191. (void)fprintf(stderr, "zmagic %d\n", m);
  192. goto done_encoding;
  193. }
  194. #endif
  195. /* Check if we have a tar file */
  196. if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0)
  197. if ((m = file_is_tar(ms, ubuf, nb)) != 0) {
  198. if ((ms->flags & MAGIC_DEBUG) != 0)
  199. (void)fprintf(stderr, "tar %d\n", m);
  200. goto done;
  201. }
  202. /* Check if we have a CDF file */
  203. if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0)
  204. if ((m = file_trycdf(ms, fd, ubuf, nb)) != 0) {
  205. if ((ms->flags & MAGIC_DEBUG) != 0)
  206. (void)fprintf(stderr, "cdf %d\n", m);
  207. goto done;
  208. }
  209. /* try soft magic tests */
  210. if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0)
  211. if ((m = file_softmagic(ms, ubuf, nb, 0, NULL, BINTEST,
  212. looks_text)) != 0) {
  213. if ((ms->flags & MAGIC_DEBUG) != 0)
  214. (void)fprintf(stderr, "softmagic %d\n", m);
  215. #ifdef BUILTIN_ELF
  216. if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
  217. nb > 5 && fd != -1) {
  218. /*
  219. * We matched something in the file, so this
  220. * *might* be an ELF file, and the file is at
  221. * least 5 bytes long, so if it's an ELF file
  222. * it has at least one byte past the ELF magic
  223. * number - try extracting information from the
  224. * ELF headers that cannot easily * be
  225. * extracted with rules in the magic file.
  226. */
  227. if ((m = file_tryelf(ms, fd, ubuf, nb)) != 0)
  228. if ((ms->flags & MAGIC_DEBUG) != 0)
  229. (void)fprintf(stderr,
  230. "elf %d\n", m);
  231. }
  232. #endif
  233. goto done;
  234. }
  235. /* try text properties */
  236. if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
  237. if ((m = file_ascmagic(ms, ubuf, nb, looks_text)) != 0) {
  238. if ((ms->flags & MAGIC_DEBUG) != 0)
  239. (void)fprintf(stderr, "ascmagic %d\n", m);
  240. goto done;
  241. }
  242. }
  243. simple:
  244. /* give up */
  245. m = 1;
  246. if ((!mime || (mime & MAGIC_MIME_TYPE)) &&
  247. file_printf(ms, "%s", mime ? type : def) == -1) {
  248. rv = -1;
  249. }
  250. done:
  251. if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
  252. if (ms->flags & MAGIC_MIME_TYPE)
  253. if (file_printf(ms, "; charset=") == -1)
  254. rv = -1;
  255. if (file_printf(ms, "%s", code_mime) == -1)
  256. rv = -1;
  257. }
  258. #if HAVE_FORK
  259. done_encoding:
  260. #endif
  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. file_regex_t rx;
  390. int rc, rv = -1;
  391. rc = file_regcomp(&rx, pat, REG_EXTENDED);
  392. if (rc) {
  393. file_regerror(&rx, rc, ms);
  394. } else {
  395. regmatch_t rm;
  396. int nm = 0;
  397. while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
  398. ms->o.buf[rm.rm_so] = '\0';
  399. if (file_printf(ms, "%s%s", rep,
  400. rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
  401. goto out;
  402. nm++;
  403. }
  404. rv = nm;
  405. }
  406. out:
  407. file_regfree(&rx);
  408. return rv;
  409. }
  410. protected int
  411. file_regcomp(file_regex_t *rx, const char *pat, int flags)
  412. {
  413. #ifdef USE_C_LOCALE
  414. rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0);
  415. assert(rx->c_lc_ctype != NULL);
  416. rx->old_lc_ctype = uselocale(rx->c_lc_ctype);
  417. assert(rx->old_lc_ctype != NULL);
  418. #endif
  419. rx->pat = pat;
  420. return rx->rc = regcomp(&rx->rx, pat, flags);
  421. }
  422. protected int
  423. file_regexec(file_regex_t *rx, const char *str, size_t nmatch,
  424. regmatch_t* pmatch, int eflags)
  425. {
  426. assert(rx->rc == 0);
  427. return regexec(&rx->rx, str, nmatch, pmatch, eflags);
  428. }
  429. protected void
  430. file_regfree(file_regex_t *rx)
  431. {
  432. if (rx->rc == 0)
  433. regfree(&rx->rx);
  434. #ifdef USE_C_LOCALE
  435. (void)uselocale(rx->old_lc_ctype);
  436. freelocale(rx->c_lc_ctype);
  437. #endif
  438. }
  439. protected void
  440. file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
  441. {
  442. char errmsg[512];
  443. (void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
  444. file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
  445. errmsg);
  446. }
  447. protected file_pushbuf_t *
  448. file_push_buffer(struct magic_set *ms)
  449. {
  450. file_pushbuf_t *pb;
  451. if (ms->event_flags & EVENT_HAD_ERR)
  452. return NULL;
  453. if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL)
  454. return NULL;
  455. pb->buf = ms->o.buf;
  456. pb->offset = ms->offset;
  457. ms->o.buf = NULL;
  458. ms->offset = 0;
  459. return pb;
  460. }
  461. protected char *
  462. file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
  463. {
  464. char *rbuf;
  465. if (ms->event_flags & EVENT_HAD_ERR) {
  466. free(pb->buf);
  467. free(pb);
  468. return NULL;
  469. }
  470. rbuf = ms->o.buf;
  471. ms->o.buf = pb->buf;
  472. ms->offset = pb->offset;
  473. free(pb);
  474. return rbuf;
  475. }
  476. /*
  477. * convert string to ascii printable format.
  478. */
  479. protected char *
  480. file_printable(char *buf, size_t bufsiz, const char *str)
  481. {
  482. char *ptr, *eptr;
  483. const unsigned char *s = (const unsigned char *)str;
  484. for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) {
  485. if (isprint(*s)) {
  486. *ptr++ = *s;
  487. continue;
  488. }
  489. if (ptr >= eptr - 3)
  490. break;
  491. *ptr++ = '\\';
  492. *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0';
  493. *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0';
  494. *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0';
  495. }
  496. *ptr = '\0';
  497. return buf;
  498. }