funcs.c 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617
  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 1.95 2018/05/24 18:09:17 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. fprintf(stderr, "vasprintf failed (%s)", strerror(errno));
  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. if (ms->o.buf && *ms->o.buf)
  104. file_printf(ms, " ");
  105. file_vprintf(ms, f, va);
  106. if (error > 0)
  107. file_printf(ms, " (%s)", strerror(error));
  108. ms->event_flags |= EVENT_HAD_ERR;
  109. ms->error = error;
  110. }
  111. /*VARARGS*/
  112. protected void
  113. file_error(struct magic_set *ms, int error, const char *f, ...)
  114. {
  115. va_list va;
  116. va_start(va, f);
  117. file_error_core(ms, error, f, va, 0);
  118. va_end(va);
  119. }
  120. /*
  121. * Print an error with magic line number.
  122. */
  123. /*VARARGS*/
  124. protected void
  125. file_magerror(struct magic_set *ms, const char *f, ...)
  126. {
  127. va_list va;
  128. va_start(va, f);
  129. file_error_core(ms, 0, f, va, ms->line);
  130. va_end(va);
  131. }
  132. protected void
  133. file_oomem(struct magic_set *ms, size_t len)
  134. {
  135. file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
  136. len);
  137. }
  138. protected void
  139. file_badseek(struct magic_set *ms)
  140. {
  141. file_error(ms, errno, "error seeking");
  142. }
  143. protected void
  144. file_badread(struct magic_set *ms)
  145. {
  146. file_error(ms, errno, "error reading");
  147. }
  148. #ifndef COMPILE_ONLY
  149. static int
  150. checkdone(struct magic_set *ms, int *rv)
  151. {
  152. if ((ms->flags & MAGIC_CONTINUE) == 0)
  153. return 1;
  154. if (file_printf(ms, "\n- ") == -1)
  155. *rv = -1;
  156. return 0;
  157. }
  158. /*ARGSUSED*/
  159. protected int
  160. file_buffer(struct magic_set *ms, int fd, const char *inname __attribute__ ((__unused__)),
  161. const void *buf, size_t nb)
  162. {
  163. int m = 0, rv = 0, looks_text = 0;
  164. const char *code = NULL;
  165. const char *code_mime = "binary";
  166. const char *type = "application/octet-stream";
  167. const char *def = "data";
  168. const char *ftype = NULL;
  169. char *rbuf = NULL;
  170. struct buffer b;
  171. buffer_init(&b, fd, buf, nb);
  172. ms->mode = b.st.st_mode;
  173. if (nb == 0) {
  174. def = "empty";
  175. type = "application/x-empty";
  176. goto simple;
  177. } else if (nb == 1) {
  178. def = "very short file (no magic)";
  179. goto simple;
  180. }
  181. if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
  182. looks_text = file_encoding(ms, &b, NULL, 0,
  183. &code, &code_mime, &ftype);
  184. }
  185. #ifdef __EMX__
  186. if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
  187. m = file_os2_apptype(ms, inname, &b);
  188. if ((ms->flags & MAGIC_DEBUG) != 0)
  189. (void)fprintf(stderr, "[try os2_apptype %d]\n", m);
  190. switch (m) {
  191. case -1:
  192. return -1;
  193. case 0:
  194. break;
  195. default:
  196. return 1;
  197. }
  198. }
  199. #endif
  200. #if HAVE_FORK
  201. /* try compression stuff */
  202. if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0) {
  203. m = file_zmagic(ms, &b, inname);
  204. if ((ms->flags & MAGIC_DEBUG) != 0)
  205. (void)fprintf(stderr, "[try zmagic %d]\n", m);
  206. if (m) {
  207. goto done_encoding;
  208. }
  209. }
  210. #endif
  211. /* Check if we have a tar file */
  212. if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0) {
  213. m = file_is_tar(ms, &b);
  214. if ((ms->flags & MAGIC_DEBUG) != 0)
  215. (void)fprintf(stderr, "[try tar %d]\n", m);
  216. if (m) {
  217. if (checkdone(ms, &rv))
  218. goto done;
  219. }
  220. }
  221. /* Check if we have a CDF file */
  222. if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0) {
  223. m = file_trycdf(ms, &b);
  224. if ((ms->flags & MAGIC_DEBUG) != 0)
  225. (void)fprintf(stderr, "[try cdf %d]\n", m);
  226. if (m) {
  227. if (checkdone(ms, &rv))
  228. goto done;
  229. }
  230. }
  231. #ifdef BUILTIN_ELF
  232. if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && nb > 5 && fd != -1) {
  233. file_pushbuf_t *pb;
  234. /*
  235. * We matched something in the file, so this
  236. * *might* be an ELF file, and the file is at
  237. * least 5 bytes long, so if it's an ELF file
  238. * it has at least one byte past the ELF magic
  239. * number - try extracting information from the
  240. * ELF headers that cannot easily be extracted
  241. * with rules in the magic file. We we don't
  242. * print the information yet.
  243. */
  244. if ((pb = file_push_buffer(ms)) == NULL)
  245. return -1;
  246. rv = file_tryelf(ms, &b);
  247. rbuf = file_pop_buffer(ms, pb);
  248. if (rv != 1) {
  249. free(rbuf);
  250. rbuf = NULL;
  251. }
  252. if ((ms->flags & MAGIC_DEBUG) != 0)
  253. (void)fprintf(stderr, "[try elf %d]\n", m);
  254. }
  255. #endif
  256. /* try soft magic tests */
  257. if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
  258. m = file_softmagic(ms, &b, NULL, NULL, BINTEST, looks_text);
  259. if ((ms->flags & MAGIC_DEBUG) != 0)
  260. (void)fprintf(stderr, "[try softmagic %d]\n", m);
  261. if (m == 1 && rbuf) {
  262. if (file_printf(ms, "%s", rbuf) == -1)
  263. goto done;
  264. }
  265. if (m) {
  266. if (checkdone(ms, &rv))
  267. goto done;
  268. }
  269. }
  270. /* try text properties */
  271. if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
  272. m = file_ascmagic(ms, &b, looks_text);
  273. if ((ms->flags & MAGIC_DEBUG) != 0)
  274. (void)fprintf(stderr, "[try ascmagic %d]\n", m);
  275. if (m) {
  276. if (checkdone(ms, &rv))
  277. goto done;
  278. }
  279. }
  280. simple:
  281. /* give up */
  282. m = 1;
  283. if (ms->flags & MAGIC_MIME) {
  284. if ((ms->flags & MAGIC_MIME_TYPE) &&
  285. file_printf(ms, "%s", type) == -1)
  286. rv = -1;
  287. } else if (ms->flags & MAGIC_APPLE) {
  288. if (file_printf(ms, "UNKNUNKN") == -1)
  289. rv = -1;
  290. } else if (ms->flags & MAGIC_EXTENSION) {
  291. if (file_printf(ms, "???") == -1)
  292. rv = -1;
  293. } else {
  294. if (file_printf(ms, "%s", def) == -1)
  295. rv = -1;
  296. }
  297. done:
  298. if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
  299. if (ms->flags & MAGIC_MIME_TYPE)
  300. if (file_printf(ms, "; charset=") == -1)
  301. rv = -1;
  302. if (file_printf(ms, "%s", code_mime) == -1)
  303. rv = -1;
  304. }
  305. #if HAVE_FORK
  306. done_encoding:
  307. #endif
  308. free(rbuf);
  309. buffer_fini(&b);
  310. if (rv)
  311. return rv;
  312. return m;
  313. }
  314. #endif
  315. protected int
  316. file_reset(struct magic_set *ms, int checkloaded)
  317. {
  318. if (checkloaded && ms->mlist[0] == NULL) {
  319. file_error(ms, 0, "no magic files loaded");
  320. return -1;
  321. }
  322. if (ms->o.buf) {
  323. free(ms->o.buf);
  324. ms->o.buf = NULL;
  325. }
  326. if (ms->o.pbuf) {
  327. free(ms->o.pbuf);
  328. ms->o.pbuf = NULL;
  329. }
  330. ms->event_flags &= ~EVENT_HAD_ERR;
  331. ms->error = -1;
  332. return 0;
  333. }
  334. #define OCTALIFY(n, o) \
  335. /*LINTED*/ \
  336. (void)(*(n)++ = '\\', \
  337. *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
  338. *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
  339. *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
  340. (o)++)
  341. protected const char *
  342. file_getbuffer(struct magic_set *ms)
  343. {
  344. char *pbuf, *op, *np;
  345. size_t psize, len;
  346. if (ms->event_flags & EVENT_HAD_ERR)
  347. return NULL;
  348. if (ms->flags & MAGIC_RAW)
  349. return ms->o.buf;
  350. if (ms->o.buf == NULL)
  351. return NULL;
  352. /* * 4 is for octal representation, + 1 is for NUL */
  353. len = strlen(ms->o.buf);
  354. if (len > (SIZE_MAX - 1) / 4) {
  355. file_oomem(ms, len);
  356. return NULL;
  357. }
  358. psize = len * 4 + 1;
  359. if ((pbuf = CAST(char *, realloc(ms->o.pbuf, psize))) == NULL) {
  360. file_oomem(ms, psize);
  361. return NULL;
  362. }
  363. ms->o.pbuf = pbuf;
  364. #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
  365. {
  366. mbstate_t state;
  367. wchar_t nextchar;
  368. int mb_conv = 1;
  369. size_t bytesconsumed;
  370. char *eop;
  371. (void)memset(&state, 0, sizeof(mbstate_t));
  372. np = ms->o.pbuf;
  373. op = ms->o.buf;
  374. eop = op + len;
  375. while (op < eop) {
  376. bytesconsumed = mbrtowc(&nextchar, op,
  377. (size_t)(eop - op), &state);
  378. if (bytesconsumed == (size_t)(-1) ||
  379. bytesconsumed == (size_t)(-2)) {
  380. mb_conv = 0;
  381. break;
  382. }
  383. if (iswprint(nextchar)) {
  384. (void)memcpy(np, op, bytesconsumed);
  385. op += bytesconsumed;
  386. np += bytesconsumed;
  387. } else {
  388. while (bytesconsumed-- > 0)
  389. OCTALIFY(np, op);
  390. }
  391. }
  392. *np = '\0';
  393. /* Parsing succeeded as a multi-byte sequence */
  394. if (mb_conv != 0)
  395. return ms->o.pbuf;
  396. }
  397. #endif
  398. for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
  399. if (isprint((unsigned char)*op)) {
  400. *np++ = *op++;
  401. } else {
  402. OCTALIFY(np, op);
  403. }
  404. }
  405. *np = '\0';
  406. return ms->o.pbuf;
  407. }
  408. protected int
  409. file_check_mem(struct magic_set *ms, unsigned int level)
  410. {
  411. size_t len;
  412. if (level >= ms->c.len) {
  413. len = (ms->c.len = 20 + level) * sizeof(*ms->c.li);
  414. ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
  415. malloc(len) :
  416. realloc(ms->c.li, len));
  417. if (ms->c.li == NULL) {
  418. file_oomem(ms, len);
  419. return -1;
  420. }
  421. }
  422. ms->c.li[level].got_match = 0;
  423. #ifdef ENABLE_CONDITIONALS
  424. ms->c.li[level].last_match = 0;
  425. ms->c.li[level].last_cond = COND_NONE;
  426. #endif /* ENABLE_CONDITIONALS */
  427. return 0;
  428. }
  429. protected size_t
  430. file_printedlen(const struct magic_set *ms)
  431. {
  432. return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
  433. }
  434. protected int
  435. file_replace(struct magic_set *ms, const char *pat, const char *rep)
  436. {
  437. file_regex_t rx;
  438. int rc, rv = -1;
  439. rc = file_regcomp(&rx, pat, REG_EXTENDED);
  440. if (rc) {
  441. file_regerror(&rx, rc, ms);
  442. } else {
  443. regmatch_t rm;
  444. int nm = 0;
  445. while (file_regexec(&rx, ms->o.buf, 1, &rm, 0) == 0) {
  446. ms->o.buf[rm.rm_so] = '\0';
  447. if (file_printf(ms, "%s%s", rep,
  448. rm.rm_eo != 0 ? ms->o.buf + rm.rm_eo : "") == -1)
  449. goto out;
  450. nm++;
  451. }
  452. rv = nm;
  453. }
  454. out:
  455. file_regfree(&rx);
  456. return rv;
  457. }
  458. protected int
  459. file_regcomp(file_regex_t *rx, const char *pat, int flags)
  460. {
  461. #ifdef USE_C_LOCALE
  462. rx->c_lc_ctype = newlocale(LC_CTYPE_MASK, "C", 0);
  463. assert(rx->c_lc_ctype != NULL);
  464. rx->old_lc_ctype = uselocale(rx->c_lc_ctype);
  465. assert(rx->old_lc_ctype != NULL);
  466. #else
  467. rx->old_lc_ctype = setlocale(LC_CTYPE, "C");
  468. #endif
  469. rx->pat = pat;
  470. return rx->rc = regcomp(&rx->rx, pat, flags);
  471. }
  472. protected int
  473. file_regexec(file_regex_t *rx, const char *str, size_t nmatch,
  474. regmatch_t* pmatch, int eflags)
  475. {
  476. assert(rx->rc == 0);
  477. /* XXX: force initialization because glibc does not always do this */
  478. memset(pmatch, 0, nmatch * sizeof(*pmatch));
  479. return regexec(&rx->rx, str, nmatch, pmatch, eflags);
  480. }
  481. protected void
  482. file_regfree(file_regex_t *rx)
  483. {
  484. if (rx->rc == 0)
  485. regfree(&rx->rx);
  486. #ifdef USE_C_LOCALE
  487. (void)uselocale(rx->old_lc_ctype);
  488. freelocale(rx->c_lc_ctype);
  489. #else
  490. (void)setlocale(LC_CTYPE, rx->old_lc_ctype);
  491. #endif
  492. }
  493. protected void
  494. file_regerror(file_regex_t *rx, int rc, struct magic_set *ms)
  495. {
  496. char errmsg[512];
  497. (void)regerror(rc, &rx->rx, errmsg, sizeof(errmsg));
  498. file_magerror(ms, "regex error %d for `%s', (%s)", rc, rx->pat,
  499. errmsg);
  500. }
  501. protected file_pushbuf_t *
  502. file_push_buffer(struct magic_set *ms)
  503. {
  504. file_pushbuf_t *pb;
  505. if (ms->event_flags & EVENT_HAD_ERR)
  506. return NULL;
  507. if ((pb = (CAST(file_pushbuf_t *, malloc(sizeof(*pb))))) == NULL)
  508. return NULL;
  509. pb->buf = ms->o.buf;
  510. pb->offset = ms->offset;
  511. ms->o.buf = NULL;
  512. ms->offset = 0;
  513. return pb;
  514. }
  515. protected char *
  516. file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
  517. {
  518. char *rbuf;
  519. if (ms->event_flags & EVENT_HAD_ERR) {
  520. free(pb->buf);
  521. free(pb);
  522. return NULL;
  523. }
  524. rbuf = ms->o.buf;
  525. ms->o.buf = pb->buf;
  526. ms->offset = pb->offset;
  527. free(pb);
  528. return rbuf;
  529. }
  530. /*
  531. * convert string to ascii printable format.
  532. */
  533. protected char *
  534. file_printable(char *buf, size_t bufsiz, const char *str)
  535. {
  536. char *ptr, *eptr;
  537. const unsigned char *s = (const unsigned char *)str;
  538. for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) {
  539. if (isprint(*s)) {
  540. *ptr++ = *s;
  541. continue;
  542. }
  543. if (ptr >= eptr - 3)
  544. break;
  545. *ptr++ = '\\';
  546. *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0';
  547. *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0';
  548. *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0';
  549. }
  550. *ptr = '\0';
  551. return buf;
  552. }