funcs.c 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652
  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 1.104 2019/05/07 02:27:11 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. protected int
  148. file_separator(struct magic_set *ms)
  149. {
  150. return file_printf(ms, "\n- ");
  151. }
  152. static int
  153. checkdone(struct magic_set *ms, int *rv)
  154. {
  155. if ((ms->flags & MAGIC_CONTINUE) == 0)
  156. return 1;
  157. if (file_separator(ms) == -1)
  158. *rv = -1;
  159. return 0;
  160. }
  161. protected int
  162. file_default(struct magic_set *ms, size_t nb)
  163. {
  164. if (ms->flags & MAGIC_MIME) {
  165. if ((ms->flags & MAGIC_MIME_TYPE) &&
  166. file_printf(ms, "application/%s",
  167. nb ? "octet-stream" : "x-empty") == -1)
  168. return -1;
  169. return 1;
  170. }
  171. if (ms->flags & MAGIC_APPLE) {
  172. if (file_printf(ms, "UNKNUNKN") == -1)
  173. return -1;
  174. return 1;
  175. }
  176. if (ms->flags & MAGIC_EXTENSION) {
  177. if (file_printf(ms, "???") == -1)
  178. return -1;
  179. return 1;
  180. }
  181. return 0;
  182. }
  183. /*
  184. * The magic detection functions return:
  185. * 1: found
  186. * 0: not found
  187. * -1: error
  188. */
  189. /*ARGSUSED*/
  190. protected int
  191. file_buffer(struct magic_set *ms, int fd, struct stat *st,
  192. const char *inname __attribute__ ((__unused__)),
  193. const void *buf, size_t nb)
  194. {
  195. int m = 0, rv = 0, looks_text = 0;
  196. const char *code = NULL;
  197. const char *code_mime = "binary";
  198. const char *def = "data";
  199. const char *ftype = NULL;
  200. char *rbuf = NULL;
  201. struct buffer b;
  202. buffer_init(&b, fd, st, buf, nb);
  203. ms->mode = b.st.st_mode;
  204. if (nb == 0) {
  205. def = "empty";
  206. goto simple;
  207. } else if (nb == 1) {
  208. def = "very short file (no magic)";
  209. goto simple;
  210. }
  211. if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
  212. looks_text = file_encoding(ms, &b, NULL, 0,
  213. &code, &code_mime, &ftype);
  214. }
  215. #ifdef __EMX__
  216. if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
  217. m = file_os2_apptype(ms, inname, &b);
  218. if ((ms->flags & MAGIC_DEBUG) != 0)
  219. (void)fprintf(stderr, "[try os2_apptype %d]\n", m);
  220. switch (m) {
  221. case -1:
  222. return -1;
  223. case 0:
  224. break;
  225. default:
  226. return 1;
  227. }
  228. }
  229. #endif
  230. #if HAVE_FORK
  231. /* try compression stuff */
  232. if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0) {
  233. m = file_zmagic(ms, &b, inname);
  234. if ((ms->flags & MAGIC_DEBUG) != 0)
  235. (void)fprintf(stderr, "[try zmagic %d]\n", m);
  236. if (m) {
  237. goto done_encoding;
  238. }
  239. }
  240. #endif
  241. /* Check if we have a tar file */
  242. if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0) {
  243. m = file_is_tar(ms, &b);
  244. if ((ms->flags & MAGIC_DEBUG) != 0)
  245. (void)fprintf(stderr, "[try tar %d]\n", m);
  246. if (m) {
  247. if (checkdone(ms, &rv))
  248. goto done;
  249. }
  250. }
  251. /* Check if we have a JSON file */
  252. if ((ms->flags & MAGIC_NO_CHECK_JSON) == 0) {
  253. m = file_is_json(ms, &b);
  254. if ((ms->flags & MAGIC_DEBUG) != 0)
  255. (void)fprintf(stderr, "[try json %d]\n", m);
  256. if (m) {
  257. if (checkdone(ms, &rv))
  258. goto done;
  259. }
  260. }
  261. /* Check if we have a CDF file */
  262. if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0) {
  263. m = file_trycdf(ms, &b);
  264. if ((ms->flags & MAGIC_DEBUG) != 0)
  265. (void)fprintf(stderr, "[try cdf %d]\n", m);
  266. if (m) {
  267. if (checkdone(ms, &rv))
  268. goto done;
  269. }
  270. }
  271. #ifdef BUILTIN_ELF
  272. if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && nb > 5 && fd != -1) {
  273. file_pushbuf_t *pb;
  274. /*
  275. * We matched something in the file, so this
  276. * *might* be an ELF file, and the file is at
  277. * least 5 bytes long, so if it's an ELF file
  278. * it has at least one byte past the ELF magic
  279. * number - try extracting information from the
  280. * ELF headers that cannot easily be extracted
  281. * with rules in the magic file. We we don't
  282. * print the information yet.
  283. */
  284. if ((pb = file_push_buffer(ms)) == NULL)
  285. return -1;
  286. rv = file_tryelf(ms, &b);
  287. rbuf = file_pop_buffer(ms, pb);
  288. if (rv == -1) {
  289. free(rbuf);
  290. rbuf = NULL;
  291. }
  292. if ((ms->flags & MAGIC_DEBUG) != 0)
  293. (void)fprintf(stderr, "[try elf %d]\n", m);
  294. }
  295. #endif
  296. /* try soft magic tests */
  297. if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
  298. m = file_softmagic(ms, &b, NULL, NULL, BINTEST, looks_text);
  299. if ((ms->flags & MAGIC_DEBUG) != 0)
  300. (void)fprintf(stderr, "[try softmagic %d]\n", m);
  301. if (m == 1 && rbuf) {
  302. if (file_printf(ms, "%s", rbuf) == -1)
  303. goto done;
  304. }
  305. if (m) {
  306. if (checkdone(ms, &rv))
  307. goto done;
  308. }
  309. }
  310. /* try text properties */
  311. if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
  312. m = file_ascmagic(ms, &b, looks_text);
  313. if ((ms->flags & MAGIC_DEBUG) != 0)
  314. (void)fprintf(stderr, "[try ascmagic %d]\n", m);
  315. if (m) {
  316. goto done;
  317. }
  318. }
  319. simple:
  320. /* give up */
  321. if (m == 0) {
  322. m = 1;
  323. rv = file_default(ms, nb);
  324. if (rv == 0)
  325. if (file_printf(ms, "%s", def) == -1)
  326. rv = -1;
  327. }
  328. done:
  329. if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
  330. if (ms->flags & MAGIC_MIME_TYPE)
  331. if (file_printf(ms, "; charset=") == -1)
  332. rv = -1;
  333. if (file_printf(ms, "%s", code_mime) == -1)
  334. rv = -1;
  335. }
  336. #if HAVE_FORK
  337. done_encoding:
  338. #endif
  339. free(rbuf);
  340. buffer_fini(&b);
  341. if (rv)
  342. return rv;
  343. return m;
  344. }
  345. #endif
  346. protected int
  347. file_reset(struct magic_set *ms, int checkloaded)
  348. {
  349. if (checkloaded && ms->mlist[0] == NULL) {
  350. file_error(ms, 0, "no magic files loaded");
  351. return -1;
  352. }
  353. if (ms->o.buf) {
  354. free(ms->o.buf);
  355. ms->o.buf = NULL;
  356. }
  357. if (ms->o.pbuf) {
  358. free(ms->o.pbuf);
  359. ms->o.pbuf = NULL;
  360. }
  361. ms->event_flags &= ~EVENT_HAD_ERR;
  362. ms->error = -1;
  363. return 0;
  364. }
  365. #define OCTALIFY(n, o) \
  366. /*LINTED*/ \
  367. (void)(*(n)++ = '\\', \
  368. *(n)++ = ((CAST(uint32_t, *(o)) >> 6) & 3) + '0', \
  369. *(n)++ = ((CAST(uint32_t, *(o)) >> 3) & 7) + '0', \
  370. *(n)++ = ((CAST(uint32_t, *(o)) >> 0) & 7) + '0', \
  371. (o)++)
  372. protected const char *
  373. file_getbuffer(struct magic_set *ms)
  374. {
  375. char *pbuf, *op, *np;
  376. size_t psize, len;
  377. if (ms->event_flags & EVENT_HAD_ERR)
  378. return NULL;
  379. if (ms->flags & MAGIC_RAW)
  380. return ms->o.buf;
  381. if (ms->o.buf == NULL)
  382. return NULL;
  383. /* * 4 is for octal representation, + 1 is for NUL */
  384. len = strlen(ms->o.buf);
  385. if (len > (SIZE_MAX - 1) / 4) {
  386. file_oomem(ms, len);
  387. return NULL;
  388. }
  389. psize = len * 4 + 1;
  390. if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
  391. file_oomem(ms, psize);
  392. return NULL;
  393. }
  394. ms->o.pbuf = pbuf;
  395. #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
  396. {
  397. mbstate_t state;
  398. wchar_t nextchar;
  399. int mb_conv = 1;
  400. size_t bytesconsumed;
  401. char *eop;
  402. (void)memset(&state, 0, sizeof(mbstate_t));
  403. np = ms->o.pbuf;
  404. op = ms->o.buf;
  405. eop = op + len;
  406. while (op < eop) {
  407. bytesconsumed = mbrtowc(&nextchar, op,
  408. CAST(size_t, eop - op), &state);
  409. if (bytesconsumed == CAST(size_t, -1) ||
  410. bytesconsumed == CAST(size_t, -2)) {
  411. mb_conv = 0;
  412. break;
  413. }
  414. if (iswprint(nextchar)) {
  415. (void)memcpy(np, op, bytesconsumed);
  416. op += bytesconsumed;
  417. np += bytesconsumed;
  418. } else {
  419. while (bytesconsumed-- > 0)
  420. OCTALIFY(np, op);
  421. }
  422. }
  423. *np = '\0';
  424. /* Parsing succeeded as a multi-byte sequence */
  425. if (mb_conv != 0)
  426. return ms->o.pbuf;
  427. }
  428. #endif
  429. for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
  430. if (isprint(CAST(unsigned char, *op))) {
  431. *np++ = *op++;
  432. } else {
  433. OCTALIFY(np, op);
  434. }
  435. }
  436. *np = '\0';
  437. return ms->o.pbuf;
  438. }
  439. protected int
  440. file_check_mem(struct magic_set *ms, unsigned int level)
  441. {
  442. size_t len;
  443. if (level >= ms->c.len) {
  444. len = (ms->c.len = 20 + level) * sizeof(*ms->c.li);
  445. ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
  446. malloc(len) :
  447. realloc(ms->c.li, len));
  448. if (ms->c.li == NULL) {
  449. file_oomem(ms, len);
  450. return -1;
  451. }
  452. }
  453. ms->c.li[level].got_match = 0;
  454. #ifdef ENABLE_CONDITIONALS
  455. ms->c.li[level].last_match = 0;
  456. ms->c.li[level].last_cond = COND_NONE;
  457. #endif /* ENABLE_CONDITIONALS */
  458. return 0;
  459. }
  460. protected size_t
  461. file_printedlen(const struct magic_set *ms)
  462. {
  463. return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
  464. }
  465. protected int
  466. file_replace(struct magic_set *ms, const char *pat, const char *rep)
  467. {
  468. file_regex_t rx;
  469. int rc, rv = -1;
  470. rc = file_regcomp(&rx, pat, REG_EXTENDED);
  471. if (rc) {
  472. file_regerror(&rx, rc, ms);
  473. } else {
  474. regmatch_t rm;
  475. int nm = 0;
  476. while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
  477. ms->o.buf[rm.rm_so] = '\0';
  478. if (file_printf(ms, "%s%s", rep,
  479. rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
  480. goto out;
  481. nm++;
  482. }
  483. rv = nm;
  484. }
  485. out:
  486. file_regfree(&rx);
  487. return rv;
  488. }
  489. protected int
  490. file_regcomp(file_regex_t *rx, const char *pat, int flags)
  491. {
  492. #ifdef USE_C_LOCALE
  493. rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0);
  494. assert(rx->c_lc_ctype != NULL);
  495. rx->old_lc_ctype = uselocale(rx->c_lc_ctype);
  496. assert(rx->old_lc_ctype != NULL);
  497. #else
  498. rx->old_lc_ctype = setlocale(LC_CTYPE, "C");
  499. #endif
  500. rx->pat = pat;
  501. return rx->rc = regcomp(&rx->rx, pat, flags);
  502. }
  503. protected int
  504. file_regexec(file_regex_t *rx, const char *str, size_t nmatch,
  505. regmatch_t* pmatch, int eflags)
  506. {
  507. assert(rx->rc == 0);
  508. /* XXX: force initialization because glibc does not always do this */
  509. memset(pmatch, 0, nmatch * sizeof(*pmatch));
  510. return regexec(&rx->rx, str, nmatch, pmatch, eflags);
  511. }
  512. protected void
  513. file_regfree(file_regex_t *rx)
  514. {
  515. if (rx->rc == 0)
  516. regfree(&rx->rx);
  517. #ifdef USE_C_LOCALE
  518. (void)uselocale(rx->old_lc_ctype);
  519. freelocale(rx->c_lc_ctype);
  520. #else
  521. (void)setlocale(LC_CTYPE, rx->old_lc_ctype);
  522. #endif
  523. }
  524. protected void
  525. file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
  526. {
  527. char errmsg[512];
  528. (void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
  529. file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
  530. errmsg);
  531. }
  532. protected file_pushbuf_t *
  533. file_push_buffer(struct magic_set *ms)
  534. {
  535. file_pushbuf_t *pb;
  536. if (ms->event_flags & EVENT_HAD_ERR)
  537. return NULL;
  538. if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL)
  539. return NULL;
  540. pb->buf = ms->o.buf;
  541. pb->offset = ms->offset;
  542. ms->o.buf = NULL;
  543. ms->offset = 0;
  544. return pb;
  545. }
  546. protected char *
  547. file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
  548. {
  549. char *rbuf;
  550. if (ms->event_flags & EVENT_HAD_ERR) {
  551. free(pb->buf);
  552. free(pb);
  553. return NULL;
  554. }
  555. rbuf = ms->o.buf;
  556. ms->o.buf = pb->buf;
  557. ms->offset = pb->offset;
  558. free(pb);
  559. return rbuf;
  560. }
  561. /*
  562. * convert string to ascii printable format.
  563. */
  564. protected char *
  565. file_printable(char *buf, size_t bufsiz, const char *str, size_t slen)
  566. {
  567. char *ptr, *eptr = buf + bufsiz - 1;
  568. const unsigned char *s = RCAST(const unsigned char *, str);
  569. const unsigned char *es = s + slen;
  570. for (ptr = buf; ptr < eptr && s < es && *s; s++) {
  571. if (isprint(*s)) {
  572. *ptr++ = *s;
  573. continue;
  574. }
  575. if (ptr >= eptr - 3)
  576. break;
  577. *ptr++ = '\\';
  578. *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0';
  579. *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0';
  580. *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0';
  581. }
  582. *ptr = '\0';
  583. return buf;
  584. }