transform.c 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567
  1. /* This file is part of GNU tar.
  2. Copyright (C) 2006, 2007, 2008 Free Software Foundation, Inc.
  3. This program is free software; you can redistribute it and/or modify it
  4. under the terms of the GNU General Public License as published by the
  5. Free Software Foundation; either version 3, or (at your option) any later
  6. version.
  7. This program is distributed in the hope that it will be useful, but
  8. WITHOUT ANY WARRANTY; without even the implied warranty of
  9. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
  10. Public License for more details.
  11. You should have received a copy of the GNU General Public License along
  12. with this program; if not, write to the Free Software Foundation, Inc.,
  13. 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
  14. #include <system.h>
  15. #include <regex.h>
  16. #include "common.h"
  17. enum transform_type
  18. {
  19. transform_first,
  20. transform_global
  21. };
  22. enum replace_segm_type
  23. {
  24. segm_literal, /* Literal segment */
  25. segm_backref, /* Back-reference segment */
  26. segm_case_ctl /* Case control segment (GNU extension) */
  27. };
  28. enum case_ctl_type
  29. {
  30. ctl_stop, /* Stop case conversion */
  31. ctl_upcase_next,/* Turn the next character to uppercase */
  32. ctl_locase_next,/* Turn the next character to lowercase */
  33. ctl_upcase, /* Turn the replacement to uppercase until ctl_stop */
  34. ctl_locase /* Turn the replacement to lowercase until ctl_stop */
  35. };
  36. struct replace_segm
  37. {
  38. struct replace_segm *next;
  39. enum replace_segm_type type;
  40. union
  41. {
  42. struct
  43. {
  44. char *ptr;
  45. size_t size;
  46. } literal; /* type == segm_literal */
  47. size_t ref; /* type == segm_backref */
  48. enum case_ctl_type ctl; /* type == segm_case_ctl */
  49. } v;
  50. };
  51. struct transform
  52. {
  53. struct transform *next;
  54. enum transform_type transform_type;
  55. unsigned match_number;
  56. regex_t regex;
  57. /* Compiled replacement expression */
  58. struct replace_segm *repl_head, *repl_tail;
  59. size_t segm_count; /* Number of elements in the above list */
  60. };
  61. static struct transform *transform_head, *transform_tail;
  62. static struct transform *
  63. new_transform ()
  64. {
  65. struct transform *p = xzalloc (sizeof *p);
  66. if (transform_tail)
  67. transform_tail->next = p;
  68. else
  69. transform_head = p;
  70. transform_tail = p;
  71. return p;
  72. }
  73. static struct replace_segm *
  74. add_segment (struct transform *tf)
  75. {
  76. struct replace_segm *segm = xmalloc (sizeof *segm);
  77. segm->next = NULL;
  78. if (tf->repl_tail)
  79. tf->repl_tail->next = segm;
  80. else
  81. tf->repl_head = segm;
  82. tf->repl_tail = segm;
  83. tf->segm_count++;
  84. return segm;
  85. }
  86. static void
  87. add_literal_segment (struct transform *tf, char *str, char *end)
  88. {
  89. size_t len = end - str;
  90. if (len)
  91. {
  92. struct replace_segm *segm = add_segment (tf);
  93. segm->type = segm_literal;
  94. segm->v.literal.ptr = xmalloc (len + 1);
  95. memcpy (segm->v.literal.ptr, str, len);
  96. segm->v.literal.ptr[len] = 0;
  97. segm->v.literal.size = len;
  98. }
  99. }
  100. static void
  101. add_char_segment (struct transform *tf, int chr)
  102. {
  103. struct replace_segm *segm = add_segment (tf);
  104. segm->type = segm_literal;
  105. segm->v.literal.ptr = xmalloc (2);
  106. segm->v.literal.ptr[0] = chr;
  107. segm->v.literal.ptr[1] = 0;
  108. segm->v.literal.size = 1;
  109. }
  110. static void
  111. add_backref_segment (struct transform *tf, size_t ref)
  112. {
  113. struct replace_segm *segm = add_segment (tf);
  114. segm->type = segm_backref;
  115. segm->v.ref = ref;
  116. }
  117. static void
  118. add_case_ctl_segment (struct transform *tf, enum case_ctl_type ctl)
  119. {
  120. struct replace_segm *segm = add_segment (tf);
  121. segm->type = segm_case_ctl;
  122. segm->v.ctl = ctl;
  123. }
  124. static const char *
  125. parse_transform_expr (const char *expr)
  126. {
  127. int delim;
  128. int i, j, rc;
  129. char *str, *beg, *cur;
  130. const char *p;
  131. int cflags = 0;
  132. struct transform *tf = new_transform ();
  133. if (expr[0] != 's')
  134. USAGE_ERROR ((0, 0, _("Invalid transform expression")));
  135. delim = expr[1];
  136. /* Scan regular expression */
  137. for (i = 2; expr[i] && expr[i] != delim; i++)
  138. if (expr[i] == '\\' && expr[i+1])
  139. i++;
  140. if (expr[i] != delim)
  141. USAGE_ERROR ((0, 0, _("Invalid transform expression")));
  142. /* Scan replacement expression */
  143. for (j = i + 1; expr[j] && expr[j] != delim; j++)
  144. if (expr[j] == '\\' && expr[j+1])
  145. j++;
  146. if (expr[j] != delim)
  147. USAGE_ERROR ((0, 0, _("Invalid transform expression")));
  148. /* Check flags */
  149. tf->transform_type = transform_first;
  150. for (p = expr + j + 1; *p && *p != ';'; p++)
  151. switch (*p)
  152. {
  153. case 'g':
  154. tf->transform_type = transform_global;
  155. break;
  156. case 'i':
  157. cflags |= REG_ICASE;
  158. break;
  159. case 'x':
  160. cflags |= REG_EXTENDED;
  161. break;
  162. case '0': case '1': case '2': case '3': case '4':
  163. case '5': case '6': case '7': case '8': case '9':
  164. tf->match_number = strtoul (p, (char**) &p, 0);
  165. p--;
  166. break;
  167. default:
  168. USAGE_ERROR ((0, 0, _("Unknown flag in transform expression: %c"),
  169. *p));
  170. }
  171. if (*p == ';')
  172. p++;
  173. /* Extract and compile regex */
  174. str = xmalloc (i - 1);
  175. memcpy (str, expr + 2, i - 2);
  176. str[i - 2] = 0;
  177. rc = regcomp (&tf->regex, str, cflags);
  178. if (rc)
  179. {
  180. char errbuf[512];
  181. regerror (rc, &tf->regex, errbuf, sizeof (errbuf));
  182. USAGE_ERROR ((0, 0, _("Invalid transform expression: %s"), errbuf));
  183. }
  184. if (str[0] == '^' || str[strlen (str) - 1] == '$')
  185. tf->transform_type = transform_first;
  186. free (str);
  187. /* Extract and compile replacement expr */
  188. i++;
  189. str = xmalloc (j - i + 1);
  190. memcpy (str, expr + i, j - i);
  191. str[j - i] = 0;
  192. for (cur = beg = str; *cur;)
  193. {
  194. if (*cur == '\\')
  195. {
  196. size_t n;
  197. add_literal_segment (tf, beg, cur);
  198. switch (*++cur)
  199. {
  200. case '0': case '1': case '2': case '3': case '4':
  201. case '5': case '6': case '7': case '8': case '9':
  202. n = strtoul (cur, &cur, 10);
  203. if (n > tf->regex.re_nsub)
  204. USAGE_ERROR ((0, 0, _("Invalid transform replacement: back reference out of range")));
  205. add_backref_segment (tf, n);
  206. break;
  207. case '\\':
  208. add_char_segment (tf, '\\');
  209. cur++;
  210. break;
  211. case 'a':
  212. add_char_segment (tf, '\a');
  213. cur++;
  214. break;
  215. case 'b':
  216. add_char_segment (tf, '\b');
  217. cur++;
  218. break;
  219. case 'f':
  220. add_char_segment (tf, '\f');
  221. cur++;
  222. break;
  223. case 'n':
  224. add_char_segment (tf, '\n');
  225. cur++;
  226. break;
  227. case 'r':
  228. add_char_segment (tf, '\r');
  229. cur++;
  230. break;
  231. case 't':
  232. add_char_segment (tf, '\t');
  233. cur++;
  234. break;
  235. case 'v':
  236. add_char_segment (tf, '\v');
  237. cur++;
  238. break;
  239. case '&':
  240. add_char_segment (tf, '&');
  241. cur++;
  242. break;
  243. case 'L':
  244. /* Turn the replacement to lowercase until a `\U' or `\E'
  245. is found, */
  246. add_case_ctl_segment (tf, ctl_locase);
  247. cur++;
  248. break;
  249. case 'l':
  250. /* Turn the next character to lowercase, */
  251. add_case_ctl_segment (tf, ctl_locase_next);
  252. cur++;
  253. break;
  254. case 'U':
  255. /* Turn the replacement to uppercase until a `\L' or `\E'
  256. is found, */
  257. add_case_ctl_segment (tf, ctl_upcase);
  258. cur++;
  259. break;
  260. case 'u':
  261. /* Turn the next character to uppercase, */
  262. add_case_ctl_segment (tf, ctl_upcase_next);
  263. cur++;
  264. break;
  265. case 'E':
  266. /* Stop case conversion started by `\L' or `\U'. */
  267. add_case_ctl_segment (tf, ctl_stop);
  268. cur++;
  269. break;
  270. default:
  271. /* Try to be nice */
  272. {
  273. char buf[2];
  274. buf[0] = '\\';
  275. buf[1] = *cur;
  276. add_literal_segment (tf, buf, buf + 2);
  277. }
  278. cur++;
  279. break;
  280. }
  281. beg = cur;
  282. }
  283. else if (*cur == '&')
  284. {
  285. add_literal_segment (tf, beg, cur);
  286. add_backref_segment (tf, 0);
  287. beg = ++cur;
  288. }
  289. else
  290. cur++;
  291. }
  292. add_literal_segment (tf, beg, cur);
  293. return p;
  294. }
  295. void
  296. set_transform_expr (const char *expr)
  297. {
  298. while (*expr)
  299. expr = parse_transform_expr (expr);
  300. }
  301. /* Run case conversion specified by CASE_CTL on array PTR of SIZE
  302. characters. Returns pointer to statically allocated storage. */
  303. static char *
  304. run_case_conv (enum case_ctl_type case_ctl, char *ptr, size_t size)
  305. {
  306. static char *case_ctl_buffer;
  307. static size_t case_ctl_bufsize;
  308. char *p;
  309. if (case_ctl_bufsize < size)
  310. {
  311. case_ctl_bufsize = size;
  312. case_ctl_buffer = xrealloc (case_ctl_buffer, case_ctl_bufsize);
  313. }
  314. memcpy (case_ctl_buffer, ptr, size);
  315. switch (case_ctl)
  316. {
  317. case ctl_upcase_next:
  318. case_ctl_buffer[0] = toupper (case_ctl_buffer[0]);
  319. break;
  320. case ctl_locase_next:
  321. case_ctl_buffer[0] = tolower (case_ctl_buffer[0]);
  322. break;
  323. case ctl_upcase:
  324. for (p = case_ctl_buffer; p < case_ctl_buffer + size; p++)
  325. *p = toupper (*p);
  326. break;
  327. case ctl_locase:
  328. for (p = case_ctl_buffer; p < case_ctl_buffer + size; p++)
  329. *p = tolower (*p);
  330. break;
  331. case ctl_stop:
  332. break;
  333. }
  334. return case_ctl_buffer;
  335. }
  336. static struct obstack stk;
  337. static bool stk_init;
  338. void
  339. _single_transform_name_to_obstack (struct transform *tf, char *input)
  340. {
  341. regmatch_t *rmp;
  342. int rc;
  343. size_t nmatches = 0;
  344. enum case_ctl_type case_ctl = ctl_stop, /* Current case conversion op */
  345. save_ctl = ctl_stop; /* Saved case_ctl for \u and \l */
  346. /* Reset case conversion after a single-char operation */
  347. #define CASE_CTL_RESET() if (case_ctl == ctl_upcase_next \
  348. || case_ctl == ctl_locase_next) \
  349. { \
  350. case_ctl = save_ctl; \
  351. save_ctl = ctl_stop; \
  352. }
  353. rmp = xmalloc ((tf->regex.re_nsub + 1) * sizeof (*rmp));
  354. while (*input)
  355. {
  356. size_t disp;
  357. char *ptr;
  358. rc = regexec (&tf->regex, input, tf->regex.re_nsub + 1, rmp, 0);
  359. if (rc == 0)
  360. {
  361. struct replace_segm *segm;
  362. disp = rmp[0].rm_eo;
  363. if (rmp[0].rm_so)
  364. obstack_grow (&stk, input, rmp[0].rm_so);
  365. nmatches++;
  366. if (tf->match_number && nmatches < tf->match_number)
  367. {
  368. obstack_grow (&stk, input, disp);
  369. input += disp;
  370. continue;
  371. }
  372. for (segm = tf->repl_head; segm; segm = segm->next)
  373. {
  374. switch (segm->type)
  375. {
  376. case segm_literal: /* Literal segment */
  377. if (case_ctl == ctl_stop)
  378. ptr = segm->v.literal.ptr;
  379. else
  380. {
  381. ptr = run_case_conv (case_ctl,
  382. segm->v.literal.ptr,
  383. segm->v.literal.size);
  384. CASE_CTL_RESET();
  385. }
  386. obstack_grow (&stk, ptr, segm->v.literal.size);
  387. break;
  388. case segm_backref: /* Back-reference segment */
  389. if (rmp[segm->v.ref].rm_so != -1
  390. && rmp[segm->v.ref].rm_eo != -1)
  391. {
  392. size_t size = rmp[segm->v.ref].rm_eo
  393. - rmp[segm->v.ref].rm_so;
  394. ptr = input + rmp[segm->v.ref].rm_so;
  395. if (case_ctl != ctl_stop)
  396. {
  397. ptr = run_case_conv (case_ctl, ptr, size);
  398. CASE_CTL_RESET();
  399. }
  400. obstack_grow (&stk, ptr, size);
  401. }
  402. break;
  403. case segm_case_ctl:
  404. switch (segm->v.ctl)
  405. {
  406. case ctl_upcase_next:
  407. case ctl_locase_next:
  408. switch (save_ctl)
  409. {
  410. case ctl_stop:
  411. case ctl_upcase:
  412. case ctl_locase:
  413. save_ctl = case_ctl;
  414. default:
  415. break;
  416. }
  417. /*FALL THROUGH*/
  418. case ctl_upcase:
  419. case ctl_locase:
  420. case ctl_stop:
  421. case_ctl = segm->v.ctl;
  422. }
  423. }
  424. }
  425. }
  426. else
  427. {
  428. disp = strlen (input);
  429. obstack_grow (&stk, input, disp);
  430. }
  431. input += disp;
  432. if (tf->transform_type == transform_first)
  433. {
  434. obstack_grow (&stk, input, strlen (input));
  435. break;
  436. }
  437. }
  438. obstack_1grow (&stk, 0);
  439. free (rmp);
  440. }
  441. bool
  442. _transform_name_to_obstack (char *input, char **output)
  443. {
  444. struct transform *tf;
  445. if (!stk_init)
  446. {
  447. obstack_init (&stk);
  448. stk_init = true;
  449. }
  450. for (tf = transform_head; tf; tf = tf->next)
  451. {
  452. _single_transform_name_to_obstack (tf, input);
  453. input = obstack_finish (&stk);
  454. }
  455. *output = input;
  456. return transform_head != NULL;
  457. }
  458. bool
  459. transform_name_fp (char **pinput, char *(*fun)(char *, void *), void *dat)
  460. {
  461. char *str;
  462. bool ret = _transform_name_to_obstack (*pinput, &str);
  463. if (ret)
  464. {
  465. assign_string (pinput, fun ? fun (str, dat) : str);
  466. obstack_free (&stk, str);
  467. }
  468. else if (fun)
  469. {
  470. *pinput = NULL;
  471. assign_string (pinput, fun (str, dat));
  472. free (str);
  473. ret = true;
  474. }
  475. return ret;
  476. }
  477. bool
  478. transform_name (char **pinput)
  479. {
  480. return transform_name_fp (pinput, NULL, NULL);
  481. }