funcs.c 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645
  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 1.100 2018/10/01 18:45:39 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. #include <limits.h>
  44. #ifndef SIZE_MAX
  45. #define SIZE_MAX ((size_t)~0)
  46. #endif
  47. /*
  48. * Like printf, only we append to a buffer.
  49. */
  50. protected int
  51. file_vprintf(struct magic_set *ms, const char *fmt, va_list ap)
  52. {
  53. int len;
  54. char *buf, *newstr;
  55. if (ms->event_flags & EVENT_HAD_ERR)
  56. return 0;
  57. len = vasprintf(&buf, fmt, ap);
  58. if (len < 0)
  59. goto out;
  60. if (ms->o.buf != NULL) {
  61. len = asprintf(&newstr, "%s%s", ms->o.buf, buf);
  62. free(buf);
  63. if (len < 0)
  64. goto out;
  65. free(ms->o.buf);
  66. buf = newstr;
  67. }
  68. ms->o.buf = buf;
  69. return 0;
  70. out:
  71. fprintf(stderr, "vasprintf failed (%s)", strerror(errno));
  72. return -1;
  73. }
  74. protected int
  75. file_printf(struct magic_set *ms, const char *fmt, ...)
  76. {
  77. int rv;
  78. va_list ap;
  79. va_start(ap, fmt);
  80. rv = file_vprintf(ms, fmt, ap);
  81. va_end(ap);
  82. return rv;
  83. }
  84. /*
  85. * error - print best error message possible
  86. */
  87. /*VARARGS*/
  88. __attribute__((__format__(__printf__, 3, 0)))
  89. private void
  90. file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
  91. size_t lineno)
  92. {
  93. /* Only the first error is ok */
  94. if (ms->event_flags & EVENT_HAD_ERR)
  95. return;
  96. if (lineno != 0) {
  97. free(ms->o.buf);
  98. ms->o.buf = NULL;
  99. (void)file_printf(ms, "line %" SIZE_T_FORMAT "u:", lineno);
  100. }
  101. if (ms->o.buf && *ms->o.buf)
  102. (void)file_printf(ms, " ");
  103. (void)file_vprintf(ms, f, va);
  104. if (error > 0)
  105. (void)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. static int
  148. checkdone(struct magic_set *ms, int *rv)
  149. {
  150. if ((ms->flags & MAGIC_CONTINUE) == 0)
  151. return 1;
  152. if (file_printf(ms, "\n- ") == -1)
  153. *rv = -1;
  154. return 0;
  155. }
  156. protected int
  157. file_default(struct magic_set *ms, size_t nb)
  158. {
  159. if (ms->flags & MAGIC_MIME) {
  160. if ((ms->flags & MAGIC_MIME_TYPE) &&
  161. file_printf(ms, "application/%s",
  162. nb ? "octet-stream" : "x-empty") == -1)
  163. return -1;
  164. return 1;
  165. }
  166. if (ms->flags & MAGIC_APPLE) {
  167. if (file_printf(ms, "UNKNUNKN") == -1)
  168. return -1;
  169. return 1;
  170. }
  171. if (ms->flags & MAGIC_EXTENSION) {
  172. if (file_printf(ms, "???") == -1)
  173. return -1;
  174. return 1;
  175. }
  176. return 0;
  177. }
  178. /*
  179. * The magic detection functions return:
  180. * 1: found
  181. * 0: not found
  182. * -1: error
  183. */
  184. /*ARGSUSED*/
  185. protected int
  186. file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((__unused__)),
  187. const void *buf, size_t nb)
  188. {
  189. int m = 0, rv = 0, looks_text = 0;
  190. const char *code = NULL;
  191. const char *code_mime = "binary";
  192. const char *def = "data";
  193. const char *ftype = NULL;
  194. char *rbuf = NULL;
  195. struct buffer b;
  196. buffer_init(&b, fd, buf, nb);
  197. ms->mode = b.st.st_mode;
  198. if (nb == 0) {
  199. def = "empty";
  200. goto simple;
  201. } else if (nb == 1) {
  202. def = "very short file (no magic)";
  203. goto simple;
  204. }
  205. if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
  206. looks_text = file_encoding(ms, &b, NULL, 0,
  207. &code, &code_mime, &ftype);
  208. }
  209. #ifdef __EMX__
  210. if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
  211. m = file_os2_apptype(ms, inname, &b);
  212. if ((ms->flags & MAGIC_DEBUG) != 0)
  213. (void)fprintf(stderr, "[try os2_apptype %d]\n", m);
  214. switch (m) {
  215. case -1:
  216. return -1;
  217. case 0:
  218. break;
  219. default:
  220. return 1;
  221. }
  222. }
  223. #endif
  224. #if HAVE_FORK
  225. /* try compression stuff */
  226. if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0) {
  227. m = file_zmagic(ms, &b, inname);
  228. if ((ms->flags & MAGIC_DEBUG) != 0)
  229. (void)fprintf(stderr, "[try zmagic %d]\n", m);
  230. if (m) {
  231. goto done_encoding;
  232. }
  233. }
  234. #endif
  235. /* Check if we have a tar file */
  236. if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0) {
  237. m = file_is_tar(ms, &b);
  238. if ((ms->flags & MAGIC_DEBUG) != 0)
  239. (void)fprintf(stderr, "[try tar %d]\n", m);
  240. if (m) {
  241. if (checkdone(ms, &rv))
  242. goto done;
  243. }
  244. }
  245. /* Check if we have a JSON file */
  246. if ((ms->flags & MAGIC_NO_CHECK_JSON) == 0) {
  247. m = file_is_json(ms, &b);
  248. if ((ms->flags & MAGIC_DEBUG) != 0)
  249. (void)fprintf(stderr, "[try json %d]\n", m);
  250. if (m) {
  251. if (checkdone(ms, &rv))
  252. goto done;
  253. }
  254. }
  255. /* Check if we have a CDF file */
  256. if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0) {
  257. m = file_trycdf(ms, &b);
  258. if ((ms->flags & MAGIC_DEBUG) != 0)
  259. (void)fprintf(stderr, "[try cdf %d]\n", m);
  260. if (m) {
  261. if (checkdone(ms, &rv))
  262. goto done;
  263. }
  264. }
  265. #ifdef BUILTIN_ELF
  266. if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && nb > 5 && fd != -1) {
  267. file_pushbuf_t *pb;
  268. /*
  269. * We matched something in the file, so this
  270. * *might* be an ELF file, and the file is at
  271. * least 5 bytes long, so if it's an ELF file
  272. * it has at least one byte past the ELF magic
  273. * number - try extracting information from the
  274. * ELF headers that cannot easily be extracted
  275. * with rules in the magic file. We we don't
  276. * print the information yet.
  277. */
  278. if ((pb = file_push_buffer(ms)) == NULL)
  279. return -1;
  280. rv = file_tryelf(ms, &b);
  281. rbuf = file_pop_buffer(ms, pb);
  282. if (rv == -1) {
  283. free(rbuf);
  284. rbuf = NULL;
  285. }
  286. if ((ms->flags & MAGIC_DEBUG) != 0)
  287. (void)fprintf(stderr, "[try elf %d]\n", m);
  288. }
  289. #endif
  290. /* try soft magic tests */
  291. if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
  292. m = file_softmagic(ms, &b, NULL, NULL, BINTEST, looks_text);
  293. if ((ms->flags & MAGIC_DEBUG) != 0)
  294. (void)fprintf(stderr, "[try softmagic %d]\n", m);
  295. if (m == 1 && rbuf) {
  296. if (file_printf(ms, "%s", rbuf) == -1)
  297. goto done;
  298. }
  299. if (m) {
  300. if (checkdone(ms, &rv))
  301. goto done;
  302. }
  303. }
  304. /* try text properties */
  305. if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
  306. m = file_ascmagic(ms, &b, looks_text);
  307. if ((ms->flags & MAGIC_DEBUG) != 0)
  308. (void)fprintf(stderr, "[try ascmagic %d]\n", m);
  309. if (m) {
  310. if (checkdone(ms, &rv))
  311. goto done;
  312. }
  313. }
  314. simple:
  315. /* give up */
  316. if (m == 0) {
  317. m = 1;
  318. rv = file_default(ms, nb);
  319. if (rv == 0)
  320. if (file_printf(ms, "%s", def) == -1)
  321. rv = -1;
  322. }
  323. done:
  324. if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
  325. if (ms->flags & MAGIC_MIME_TYPE)
  326. if (file_printf(ms, "; charset=") == -1)
  327. rv = -1;
  328. if (file_printf(ms, "%s", code_mime) == -1)
  329. rv = -1;
  330. }
  331. #if HAVE_FORK
  332. done_encoding:
  333. #endif
  334. free(rbuf);
  335. buffer_fini(&b);
  336. if (rv)
  337. return rv;
  338. return m;
  339. }
  340. #endif
  341. protected int
  342. file_reset(struct magic_set *ms, int checkloaded)
  343. {
  344. if (checkloaded && ms->mlist[0] == NULL) {
  345. file_error(ms, 0, "no magic files loaded");
  346. return -1;
  347. }
  348. if (ms->o.buf) {
  349. free(ms->o.buf);
  350. ms->o.buf = NULL;
  351. }
  352. if (ms->o.pbuf) {
  353. free(ms->o.pbuf);
  354. ms->o.pbuf = NULL;
  355. }
  356. ms->event_flags &= ~EVENT_HAD_ERR;
  357. ms->error = -1;
  358. return 0;
  359. }
  360. #define OCTALIFY(n, o) \
  361. /*LINTED*/ \
  362. (void)(*(n)++ = '\\', \
  363. *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
  364. *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
  365. *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
  366. (o)++)
  367. protected const char *
  368. file_getbuffer(struct magic_set *ms)
  369. {
  370. char *pbuf, *op, *np;
  371. size_t psize, len;
  372. if (ms->event_flags & EVENT_HAD_ERR)
  373. return NULL;
  374. if (ms->flags & MAGIC_RAW)
  375. return ms->o.buf;
  376. if (ms->o.buf == NULL)
  377. return NULL;
  378. /* * 4 is for octal representation, + 1 is for NUL */
  379. len = strlen(ms->o.buf);
  380. if (len > (SIZE_MAX - 1) / 4) {
  381. file_oomem(ms, len);
  382. return NULL;
  383. }
  384. psize = len * 4 + 1;
  385. if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
  386. file_oomem(ms, psize);
  387. return NULL;
  388. }
  389. ms->o.pbuf = pbuf;
  390. #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
  391. {
  392. mbstate_t state;
  393. wchar_t nextchar;
  394. int mb_conv = 1;
  395. size_t bytesconsumed;
  396. char *eop;
  397. (void)memset(&state, 0, sizeof(mbstate_t));
  398. np = ms->o.pbuf;
  399. op = ms->o.buf;
  400. eop = op + len;
  401. while (op < eop) {
  402. bytesconsumed = mbrtowc(&nextchar, op,
  403. (size_t)(eop - op), &state);
  404. if (bytesconsumed == (size_t)(-1) ||
  405. bytesconsumed == (size_t)(-2)) {
  406. mb_conv = 0;
  407. break;
  408. }
  409. if (iswprint(nextchar)) {
  410. (void)memcpy(np, op, bytesconsumed);
  411. op += bytesconsumed;
  412. np += bytesconsumed;
  413. } else {
  414. while (bytesconsumed-- > 0)
  415. OCTALIFY(np, op);
  416. }
  417. }
  418. *np = '\0';
  419. /* Parsing succeeded as a multi-byte sequence */
  420. if (mb_conv != 0)
  421. return ms->o.pbuf;
  422. }
  423. #endif
  424. for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
  425. if (isprint((unsigned char)*op)) {
  426. *np++ = *op++;
  427. } else {
  428. OCTALIFY(np, op);
  429. }
  430. }
  431. *np = '\0';
  432. return ms->o.pbuf;
  433. }
  434. protected int
  435. file_check_mem(struct magic_set *ms, unsigned int level)
  436. {
  437. size_t len;
  438. if (level >= ms->c.len) {
  439. len = (ms->c.len = 20 + level) * sizeof(*ms->c.li);
  440. ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
  441. malloc(len) :
  442. realloc(ms->c.li, len));
  443. if (ms->c.li == NULL) {
  444. file_oomem(ms, len);
  445. return -1;
  446. }
  447. }
  448. ms->c.li[level].got_match = 0;
  449. #ifdef ENABLE_CONDITIONALS
  450. ms->c.li[level].last_match = 0;
  451. ms->c.li[level].last_cond = COND_NONE;
  452. #endif /* ENABLE_CONDITIONALS */
  453. return 0;
  454. }
  455. protected size_t
  456. file_printedlen(const struct magic_set *ms)
  457. {
  458. return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
  459. }
  460. protected int
  461. file_replace(struct magic_set *ms, const char *pat, const char *rep)
  462. {
  463. file_regex_t rx;
  464. int rc, rv = -1;
  465. rc = file_regcomp(&rx, pat, REG_EXTENDED);
  466. if (rc) {
  467. file_regerror(&rx, rc, ms);
  468. } else {
  469. regmatch_t rm;
  470. int nm = 0;
  471. while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
  472. ms->o.buf[rm.rm_so] = '\0';
  473. if (file_printf(ms, "%s%s", rep,
  474. rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
  475. goto out;
  476. nm++;
  477. }
  478. rv = nm;
  479. }
  480. out:
  481. file_regfree(&rx);
  482. return rv;
  483. }
  484. protected int
  485. file_regcomp(file_regex_t *rx, const char *pat, int flags)
  486. {
  487. #ifdef USE_C_LOCALE
  488. rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0);
  489. assert(rx->c_lc_ctype != NULL);
  490. rx->old_lc_ctype = uselocale(rx->c_lc_ctype);
  491. assert(rx->old_lc_ctype != NULL);
  492. #else
  493. rx->old_lc_ctype = setlocale(LC_CTYPE, "C");
  494. #endif
  495. rx->pat = pat;
  496. return rx->rc = regcomp(&rx->rx, pat, flags);
  497. }
  498. protected int
  499. file_regexec(file_regex_t *rx, const char *str, size_t nmatch,
  500. regmatch_t* pmatch, int eflags)
  501. {
  502. assert(rx->rc == 0);
  503. /* XXX: force initialization because glibc does not always do this */
  504. memset(pmatch, 0, nmatch * sizeof(*pmatch));
  505. return regexec(&rx->rx, str, nmatch, pmatch, eflags);
  506. }
  507. protected void
  508. file_regfree(file_regex_t *rx)
  509. {
  510. if (rx->rc == 0)
  511. regfree(&rx->rx);
  512. #ifdef USE_C_LOCALE
  513. (void)uselocale(rx->old_lc_ctype);
  514. freelocale(rx->c_lc_ctype);
  515. #else
  516. (void)setlocale(LC_CTYPE, rx->old_lc_ctype);
  517. #endif
  518. }
  519. protected void
  520. file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
  521. {
  522. char errmsg[512];
  523. (void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
  524. file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
  525. errmsg);
  526. }
  527. protected file_pushbuf_t *
  528. file_push_buffer(struct magic_set *ms)
  529. {
  530. file_pushbuf_t *pb;
  531. if (ms->event_flags & EVENT_HAD_ERR)
  532. return NULL;
  533. if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL)
  534. return NULL;
  535. pb->buf = ms->o.buf;
  536. pb->offset = ms->offset;
  537. ms->o.buf = NULL;
  538. ms->offset = 0;
  539. return pb;
  540. }
  541. protected char *
  542. file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
  543. {
  544. char *rbuf;
  545. if (ms->event_flags & EVENT_HAD_ERR) {
  546. free(pb->buf);
  547. free(pb);
  548. return NULL;
  549. }
  550. rbuf = ms->o.buf;
  551. ms->o.buf = pb->buf;
  552. ms->offset = pb->offset;
  553. free(pb);
  554. return rbuf;
  555. }
  556. /*
  557. * convert string to ascii printable format.
  558. */
  559. protected char *
  560. file_printable(char *buf, size_t bufsiz, const char *str)
  561. {
  562. char *ptr, *eptr;
  563. const unsigned char *s = (const unsigned char *)str;
  564. for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) {
  565. if (isprint(*s)) {
  566. *ptr++ = *s;
  567. continue;
  568. }
  569. if (ptr >= eptr - 3)
  570. break;
  571. *ptr++ = '\\';
  572. *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0';
  573. *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0';
  574. *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0';
  575. }
  576. *ptr = '\0';
  577. return buf;
  578. }