sqstdrex.cpp 19 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667
  1. /* see copyright notice in squirrel.h */
  2. #include <squirrel.h>
  3. #include <string.h>
  4. #include <ctype.h>
  5. #include <setjmp.h>
  6. #include <sqstdstring.h>
  7. #ifdef _UINCODE
  8. #define scisprint iswprint
  9. #else
  10. #define scisprint isprint
  11. #endif
  12. #ifdef _DEBUG
  13. #include <stdio.h>
  14. static const SQChar *g_nnames[] =
  15. {
  16. _SC("NONE"),_SC("OP_GREEDY"), _SC("OP_OR"),
  17. _SC("OP_EXPR"),_SC("OP_NOCAPEXPR"),_SC("OP_DOT"), _SC("OP_CLASS"),
  18. _SC("OP_CCLASS"),_SC("OP_NCLASS"),_SC("OP_RANGE"),_SC("OP_CHAR"),
  19. _SC("OP_EOL"),_SC("OP_BOL"),_SC("OP_WB"),_C("OP_MB")
  20. };
  21. #endif
  22. #define OP_GREEDY (MAX_CHAR+1) // * + ? {n}
  23. #define OP_OR (MAX_CHAR+2)
  24. #define OP_EXPR (MAX_CHAR+3) //parentesis ()
  25. #define OP_NOCAPEXPR (MAX_CHAR+4) //parentesis (?:)
  26. #define OP_DOT (MAX_CHAR+5)
  27. #define OP_CLASS (MAX_CHAR+6)
  28. #define OP_CCLASS (MAX_CHAR+7)
  29. #define OP_NCLASS (MAX_CHAR+8) //negates class the [^
  30. #define OP_RANGE (MAX_CHAR+9)
  31. #define OP_CHAR (MAX_CHAR+10)
  32. #define OP_EOL (MAX_CHAR+11)
  33. #define OP_BOL (MAX_CHAR+12)
  34. #define OP_WB (MAX_CHAR+13)
  35. #define OP_MB (MAX_CHAR+14) //match balanced
  36. #define SQREX_SYMBOL_ANY_CHAR ('.')
  37. #define SQREX_SYMBOL_GREEDY_ONE_OR_MORE ('+')
  38. #define SQREX_SYMBOL_GREEDY_ZERO_OR_MORE ('*')
  39. #define SQREX_SYMBOL_GREEDY_ZERO_OR_ONE ('?')
  40. #define SQREX_SYMBOL_BRANCH ('|')
  41. #define SQREX_SYMBOL_END_OF_STRING ('$')
  42. #define SQREX_SYMBOL_BEGINNING_OF_STRING ('^')
  43. #define SQREX_SYMBOL_ESCAPE_CHAR ('\\')
  44. typedef int SQRexNodeType;
  45. typedef struct tagSQRexNode{
  46. SQRexNodeType type;
  47. SQInteger left;
  48. SQInteger right;
  49. SQInteger next;
  50. }SQRexNode;
  51. struct SQRex{
  52. const SQChar *_eol;
  53. const SQChar *_bol;
  54. const SQChar *_p;
  55. SQInteger _first;
  56. SQInteger _op;
  57. SQRexNode *_nodes;
  58. SQInteger _nallocated;
  59. SQInteger _nsize;
  60. SQInteger _nsubexpr;
  61. SQRexMatch *_matches;
  62. SQInteger _currsubexp;
  63. void *_jmpbuf;
  64. const SQChar **_error;
  65. };
  66. static SQInteger sqstd_rex_list(SQRex *exp);
  67. static SQInteger sqstd_rex_newnode(SQRex *exp, SQRexNodeType type)
  68. {
  69. SQRexNode n;
  70. n.type = type;
  71. n.next = n.right = n.left = -1;
  72. if(type == OP_EXPR)
  73. n.right = exp->_nsubexpr++;
  74. if(exp->_nallocated < (exp->_nsize + 1)) {
  75. SQInteger oldsize = exp->_nallocated;
  76. exp->_nallocated *= 2;
  77. exp->_nodes = (SQRexNode *)sq_realloc(exp->_nodes, oldsize * sizeof(SQRexNode) ,exp->_nallocated * sizeof(SQRexNode));
  78. }
  79. exp->_nodes[exp->_nsize++] = n;
  80. SQInteger newid = exp->_nsize - 1;
  81. return (SQInteger)newid;
  82. }
  83. static void sqstd_rex_error(SQRex *exp,const SQChar *error)
  84. {
  85. if(exp->_error) *exp->_error = error;
  86. longjmp(*((jmp_buf*)exp->_jmpbuf),-1);
  87. }
  88. static void sqstd_rex_expect(SQRex *exp, SQInteger n){
  89. if((*exp->_p) != n)
  90. sqstd_rex_error(exp, _SC("expected paren"));
  91. exp->_p++;
  92. }
  93. static SQChar sqstd_rex_escapechar(SQRex *exp)
  94. {
  95. if(*exp->_p == SQREX_SYMBOL_ESCAPE_CHAR){
  96. exp->_p++;
  97. switch(*exp->_p) {
  98. case 'v': exp->_p++; return '\v';
  99. case 'n': exp->_p++; return '\n';
  100. case 't': exp->_p++; return '\t';
  101. case 'r': exp->_p++; return '\r';
  102. case 'f': exp->_p++; return '\f';
  103. default: return (*exp->_p++);
  104. }
  105. } else if(!scisprint(*exp->_p)) sqstd_rex_error(exp,_SC("letter expected"));
  106. return (*exp->_p++);
  107. }
  108. static SQInteger sqstd_rex_charclass(SQRex *exp,SQInteger classid)
  109. {
  110. SQInteger n = sqstd_rex_newnode(exp,OP_CCLASS);
  111. exp->_nodes[n].left = classid;
  112. return n;
  113. }
  114. static SQInteger sqstd_rex_charnode(SQRex *exp,SQBool isclass)
  115. {
  116. SQChar t;
  117. if(*exp->_p == SQREX_SYMBOL_ESCAPE_CHAR) {
  118. exp->_p++;
  119. switch(*exp->_p) {
  120. case 'n': exp->_p++; return sqstd_rex_newnode(exp,'\n');
  121. case 't': exp->_p++; return sqstd_rex_newnode(exp,'\t');
  122. case 'r': exp->_p++; return sqstd_rex_newnode(exp,'\r');
  123. case 'f': exp->_p++; return sqstd_rex_newnode(exp,'\f');
  124. case 'v': exp->_p++; return sqstd_rex_newnode(exp,'\v');
  125. case 'a': case 'A': case 'w': case 'W': case 's': case 'S':
  126. case 'd': case 'D': case 'x': case 'X': case 'c': case 'C':
  127. case 'p': case 'P': case 'l': case 'u':
  128. {
  129. t = *exp->_p; exp->_p++;
  130. return sqstd_rex_charclass(exp,t);
  131. }
  132. case 'm':
  133. {
  134. SQChar cb, ce; //cb = character begin match ce = character end match
  135. cb = *++exp->_p; //skip 'm'
  136. ce = *++exp->_p;
  137. exp->_p++; //points to the next char to be parsed
  138. if ((!cb) || (!ce)) sqstd_rex_error(exp,_SC("balanced chars expected"));
  139. if ( cb == ce ) sqstd_rex_error(exp,_SC("open/close char can't be the same"));
  140. SQInteger node = sqstd_rex_newnode(exp,OP_MB);
  141. exp->_nodes[node].left = cb;
  142. exp->_nodes[node].right = ce;
  143. return node;
  144. }
  145. case 'b':
  146. case 'B':
  147. if(!isclass) {
  148. SQInteger node = sqstd_rex_newnode(exp,OP_WB);
  149. exp->_nodes[node].left = *exp->_p;
  150. exp->_p++;
  151. return node;
  152. } //else default
  153. default:
  154. t = *exp->_p; exp->_p++;
  155. return sqstd_rex_newnode(exp,t);
  156. }
  157. }
  158. else if(!scisprint(*exp->_p)) {
  159. sqstd_rex_error(exp,_SC("letter expected"));
  160. }
  161. t = *exp->_p; exp->_p++;
  162. return sqstd_rex_newnode(exp,t);
  163. }
  164. static SQInteger sqstd_rex_class(SQRex *exp)
  165. {
  166. SQInteger ret = -1;
  167. SQInteger first = -1,chain;
  168. if(*exp->_p == SQREX_SYMBOL_BEGINNING_OF_STRING){
  169. ret = sqstd_rex_newnode(exp,OP_NCLASS);
  170. exp->_p++;
  171. }else ret = sqstd_rex_newnode(exp,OP_CLASS);
  172. if(*exp->_p == ']') sqstd_rex_error(exp,_SC("empty class"));
  173. chain = ret;
  174. while(*exp->_p != ']' && exp->_p != exp->_eol) {
  175. if(*exp->_p == '-' && first != -1){
  176. SQInteger r;
  177. if(*exp->_p++ == ']') sqstd_rex_error(exp,_SC("unfinished range"));
  178. r = sqstd_rex_newnode(exp,OP_RANGE);
  179. if(exp->_nodes[first].type>*exp->_p) sqstd_rex_error(exp,_SC("invalid range"));
  180. if(exp->_nodes[first].type == OP_CCLASS) sqstd_rex_error(exp,_SC("cannot use character classes in ranges"));
  181. exp->_nodes[r].left = exp->_nodes[first].type;
  182. SQInteger t = sqstd_rex_escapechar(exp);
  183. exp->_nodes[r].right = t;
  184. exp->_nodes[chain].next = r;
  185. chain = r;
  186. first = -1;
  187. }
  188. else{
  189. if(first!=-1){
  190. SQInteger c = first;
  191. exp->_nodes[chain].next = c;
  192. chain = c;
  193. first = sqstd_rex_charnode(exp,SQTrue);
  194. }
  195. else{
  196. first = sqstd_rex_charnode(exp,SQTrue);
  197. }
  198. }
  199. }
  200. if(first!=-1){
  201. SQInteger c = first;
  202. exp->_nodes[chain].next = c;
  203. chain = c;
  204. first = -1;
  205. }
  206. /* hack? */
  207. exp->_nodes[ret].left = exp->_nodes[ret].next;
  208. exp->_nodes[ret].next = -1;
  209. return ret;
  210. }
  211. static SQInteger sqstd_rex_parsenumber(SQRex *exp)
  212. {
  213. SQInteger ret = *exp->_p-'0';
  214. SQInteger positions = 10;
  215. exp->_p++;
  216. while(isdigit(*exp->_p)) {
  217. ret = ret*10+(*exp->_p++-'0');
  218. if(positions==1000000000) sqstd_rex_error(exp,_SC("overflow in numeric constant"));
  219. positions *= 10;
  220. };
  221. return ret;
  222. }
  223. static SQInteger sqstd_rex_element(SQRex *exp)
  224. {
  225. SQInteger ret = -1;
  226. switch(*exp->_p)
  227. {
  228. case '(': {
  229. SQInteger expr;
  230. exp->_p++;
  231. if(*exp->_p =='?') {
  232. exp->_p++;
  233. sqstd_rex_expect(exp,':');
  234. expr = sqstd_rex_newnode(exp,OP_NOCAPEXPR);
  235. }
  236. else
  237. expr = sqstd_rex_newnode(exp,OP_EXPR);
  238. SQInteger newn = sqstd_rex_list(exp);
  239. exp->_nodes[expr].left = newn;
  240. ret = expr;
  241. sqstd_rex_expect(exp,')');
  242. }
  243. break;
  244. case '[':
  245. exp->_p++;
  246. ret = sqstd_rex_class(exp);
  247. sqstd_rex_expect(exp,']');
  248. break;
  249. case SQREX_SYMBOL_END_OF_STRING: exp->_p++; ret = sqstd_rex_newnode(exp,OP_EOL);break;
  250. case SQREX_SYMBOL_ANY_CHAR: exp->_p++; ret = sqstd_rex_newnode(exp,OP_DOT);break;
  251. default:
  252. ret = sqstd_rex_charnode(exp,SQFalse);
  253. break;
  254. }
  255. SQBool isgreedy = SQFalse;
  256. unsigned short p0 = 0, p1 = 0;
  257. switch(*exp->_p){
  258. case SQREX_SYMBOL_GREEDY_ZERO_OR_MORE: p0 = 0; p1 = 0xFFFF; exp->_p++; isgreedy = SQTrue; break;
  259. case SQREX_SYMBOL_GREEDY_ONE_OR_MORE: p0 = 1; p1 = 0xFFFF; exp->_p++; isgreedy = SQTrue; break;
  260. case SQREX_SYMBOL_GREEDY_ZERO_OR_ONE: p0 = 0; p1 = 1; exp->_p++; isgreedy = SQTrue; break;
  261. case '{':
  262. exp->_p++;
  263. if(!isdigit(*exp->_p)) sqstd_rex_error(exp,_SC("number expected"));
  264. p0 = (unsigned short)sqstd_rex_parsenumber(exp);
  265. /*******************************/
  266. switch(*exp->_p) {
  267. case '}':
  268. p1 = p0; exp->_p++;
  269. break;
  270. case ',':
  271. exp->_p++;
  272. p1 = 0xFFFF;
  273. if(isdigit(*exp->_p)){
  274. p1 = (unsigned short)sqstd_rex_parsenumber(exp);
  275. }
  276. sqstd_rex_expect(exp,'}');
  277. break;
  278. default:
  279. sqstd_rex_error(exp,_SC(", or } expected"));
  280. }
  281. /*******************************/
  282. isgreedy = SQTrue;
  283. break;
  284. }
  285. if(isgreedy) {
  286. SQInteger nnode = sqstd_rex_newnode(exp,OP_GREEDY);
  287. exp->_nodes[nnode].left = ret;
  288. exp->_nodes[nnode].right = ((p0)<<16)|p1;
  289. ret = nnode;
  290. }
  291. if((*exp->_p != SQREX_SYMBOL_BRANCH) && (*exp->_p != ')') && (*exp->_p != SQREX_SYMBOL_GREEDY_ZERO_OR_MORE) && (*exp->_p != SQREX_SYMBOL_GREEDY_ONE_OR_MORE) && (*exp->_p != '\0')) {
  292. SQInteger nnode = sqstd_rex_element(exp);
  293. exp->_nodes[ret].next = nnode;
  294. }
  295. return ret;
  296. }
  297. static SQInteger sqstd_rex_list(SQRex *exp)
  298. {
  299. SQInteger ret=-1,e;
  300. if(*exp->_p == SQREX_SYMBOL_BEGINNING_OF_STRING) {
  301. exp->_p++;
  302. ret = sqstd_rex_newnode(exp,OP_BOL);
  303. }
  304. e = sqstd_rex_element(exp);
  305. if(ret != -1) {
  306. exp->_nodes[ret].next = e;
  307. }
  308. else ret = e;
  309. if(*exp->_p == SQREX_SYMBOL_BRANCH) {
  310. SQInteger temp,tright;
  311. exp->_p++;
  312. temp = sqstd_rex_newnode(exp,OP_OR);
  313. exp->_nodes[temp].left = ret;
  314. tright = sqstd_rex_list(exp);
  315. exp->_nodes[temp].right = tright;
  316. ret = temp;
  317. }
  318. return ret;
  319. }
  320. static SQBool sqstd_rex_matchcclass(SQInteger cclass,SQChar c)
  321. {
  322. switch(cclass) {
  323. case 'a': return isalpha(c)?SQTrue:SQFalse;
  324. case 'A': return !isalpha(c)?SQTrue:SQFalse;
  325. case 'w': return (isalnum(c) || c == '_')?SQTrue:SQFalse;
  326. case 'W': return (!isalnum(c) && c != '_')?SQTrue:SQFalse;
  327. case 's': return isspace(c)?SQTrue:SQFalse;
  328. case 'S': return !isspace(c)?SQTrue:SQFalse;
  329. case 'd': return isdigit(c)?SQTrue:SQFalse;
  330. case 'D': return !isdigit(c)?SQTrue:SQFalse;
  331. case 'x': return isxdigit(c)?SQTrue:SQFalse;
  332. case 'X': return !isxdigit(c)?SQTrue:SQFalse;
  333. case 'c': return iscntrl(c)?SQTrue:SQFalse;
  334. case 'C': return !iscntrl(c)?SQTrue:SQFalse;
  335. case 'p': return ispunct(c)?SQTrue:SQFalse;
  336. case 'P': return !ispunct(c)?SQTrue:SQFalse;
  337. case 'l': return islower(c)?SQTrue:SQFalse;
  338. case 'u': return isupper(c)?SQTrue:SQFalse;
  339. }
  340. return SQFalse; /*cannot happen*/
  341. }
  342. static SQBool sqstd_rex_matchclass(SQRex* exp,SQRexNode *node,SQChar c)
  343. {
  344. do {
  345. switch(node->type) {
  346. case OP_RANGE:
  347. if(c >= node->left && c <= node->right) return SQTrue;
  348. break;
  349. case OP_CCLASS:
  350. if(sqstd_rex_matchcclass(node->left,c)) return SQTrue;
  351. break;
  352. default:
  353. if(c == node->type)return SQTrue;
  354. }
  355. } while((node->next != -1) && (node = &exp->_nodes[node->next]));
  356. return SQFalse;
  357. }
  358. static const SQChar *sqstd_rex_matchnode(SQRex* exp,SQRexNode *node,const SQChar *str,SQRexNode *next)
  359. {
  360. SQRexNodeType type = node->type;
  361. switch(type) {
  362. case OP_GREEDY: {
  363. //SQRexNode *greedystop = (node->next != -1) ? &exp->_nodes[node->next] : NULL;
  364. SQRexNode *greedystop = NULL;
  365. SQInteger p0 = (node->right >> 16)&0x0000FFFF, p1 = node->right&0x0000FFFF, nmaches = 0;
  366. const SQChar *s=str, *good = str;
  367. if(node->next != -1) {
  368. greedystop = &exp->_nodes[node->next];
  369. }
  370. else {
  371. greedystop = next;
  372. }
  373. while((nmaches == 0xFFFF || nmaches < p1)) {
  374. const SQChar *stop;
  375. if(!(s = sqstd_rex_matchnode(exp,&exp->_nodes[node->left],s,greedystop)))
  376. break;
  377. nmaches++;
  378. good=s;
  379. if(greedystop) {
  380. //checks that 0 matches satisfy the expression(if so skips)
  381. //if not would always stop(for instance if is a '?')
  382. if(greedystop->type != OP_GREEDY ||
  383. (greedystop->type == OP_GREEDY && ((greedystop->right >> 16)&0x0000FFFF) != 0))
  384. {
  385. SQRexNode *gnext = NULL;
  386. if(greedystop->next != -1) {
  387. gnext = &exp->_nodes[greedystop->next];
  388. }else if(next && next->next != -1){
  389. gnext = &exp->_nodes[next->next];
  390. }
  391. stop = sqstd_rex_matchnode(exp,greedystop,s,gnext);
  392. if(stop) {
  393. //if satisfied stop it
  394. if(p0 == p1 && p0 == nmaches) break;
  395. else if(nmaches >= p0 && p1 == 0xFFFF) break;
  396. else if(nmaches >= p0 && nmaches <= p1) break;
  397. }
  398. }
  399. }
  400. if(s >= exp->_eol)
  401. break;
  402. }
  403. if(p0 == p1 && p0 == nmaches) return good;
  404. else if(nmaches >= p0 && p1 == 0xFFFF) return good;
  405. else if(nmaches >= p0 && nmaches <= p1) return good;
  406. return NULL;
  407. }
  408. case OP_OR: {
  409. const SQChar *asd = str;
  410. SQRexNode *temp=&exp->_nodes[node->left];
  411. while( (asd = sqstd_rex_matchnode(exp,temp,asd,NULL)) ) {
  412. if(temp->next != -1)
  413. temp = &exp->_nodes[temp->next];
  414. else
  415. return asd;
  416. }
  417. asd = str;
  418. temp = &exp->_nodes[node->right];
  419. while( (asd = sqstd_rex_matchnode(exp,temp,asd,NULL)) ) {
  420. if(temp->next != -1)
  421. temp = &exp->_nodes[temp->next];
  422. else
  423. return asd;
  424. }
  425. return NULL;
  426. break;
  427. }
  428. case OP_EXPR:
  429. case OP_NOCAPEXPR:{
  430. SQRexNode *n = &exp->_nodes[node->left];
  431. const SQChar *cur = str;
  432. SQInteger capture = -1;
  433. if(node->type != OP_NOCAPEXPR && node->right == exp->_currsubexp) {
  434. capture = exp->_currsubexp;
  435. exp->_matches[capture].begin = cur;
  436. exp->_currsubexp++;
  437. }
  438. int tempcap = exp->_currsubexp;
  439. do {
  440. SQRexNode *subnext = NULL;
  441. if(n->next != -1) {
  442. subnext = &exp->_nodes[n->next];
  443. }else {
  444. subnext = next;
  445. }
  446. if(!(cur = sqstd_rex_matchnode(exp,n,cur,subnext))) {
  447. if(capture != -1){
  448. exp->_matches[capture].begin = 0;
  449. exp->_matches[capture].len = 0;
  450. }
  451. return NULL;
  452. }
  453. } while((n->next != -1) && (n = &exp->_nodes[n->next]));
  454. exp->_currsubexp = tempcap;
  455. if(capture != -1)
  456. exp->_matches[capture].len = cur - exp->_matches[capture].begin;
  457. return cur;
  458. }
  459. case OP_WB:
  460. if((str == exp->_bol && !isspace(*str))
  461. || (str == exp->_eol && !isspace(*(str-1)))
  462. || (!isspace(*str) && isspace(*(str+1)))
  463. || (isspace(*str) && !isspace(*(str+1))) ) {
  464. return (node->left == 'b')?str:NULL;
  465. }
  466. return (node->left == 'b')?NULL:str;
  467. case OP_BOL:
  468. if(str == exp->_bol) return str;
  469. return NULL;
  470. case OP_EOL:
  471. if(str == exp->_eol) return str;
  472. return NULL;
  473. case OP_DOT:{
  474. str++;
  475. }
  476. return str;
  477. case OP_NCLASS:
  478. case OP_CLASS:
  479. if(sqstd_rex_matchclass(exp,&exp->_nodes[node->left],*str)?(type == OP_CLASS?SQTrue:SQFalse):(type == OP_NCLASS?SQTrue:SQFalse)) {
  480. str++;
  481. return str;
  482. }
  483. return NULL;
  484. case OP_CCLASS:
  485. if(sqstd_rex_matchcclass(node->left,*str)) {
  486. str++;
  487. return str;
  488. }
  489. return NULL;
  490. case OP_MB:
  491. {
  492. int cb = node->left; //char that opens a balanced expression
  493. if(*str != cb) return NULL; // string doesnt start with open char
  494. int ce = node->right; //char that closes a balanced expression
  495. int cont = 1;
  496. const SQChar *strEol = exp->_eol;
  497. while (++str < strEol) {
  498. if (*str == ce) {
  499. if (--cont == 0) {
  500. return ++str;
  501. }
  502. }
  503. else if (*str == cb) cont++;
  504. }
  505. }
  506. return NULL; // string ends out of balance
  507. default: /* char */
  508. if(*str != node->type) return NULL;
  509. str++;
  510. return str;
  511. }
  512. return NULL;
  513. }
  514. /* public api */
  515. SQRex *sqstd_rex_compile(const SQChar *pattern,const SQChar **error)
  516. {
  517. SQRex *exp = (SQRex *)sq_malloc(sizeof(SQRex));
  518. exp->_eol = exp->_bol = NULL;
  519. exp->_p = pattern;
  520. exp->_nallocated = (SQInteger)scstrlen(pattern) * sizeof(SQChar);
  521. exp->_nodes = (SQRexNode *)sq_malloc(exp->_nallocated * sizeof(SQRexNode));
  522. exp->_nsize = 0;
  523. exp->_matches = 0;
  524. exp->_nsubexpr = 0;
  525. exp->_first = sqstd_rex_newnode(exp,OP_EXPR);
  526. exp->_error = error;
  527. exp->_jmpbuf = sq_malloc(sizeof(jmp_buf));
  528. if(setjmp(*((jmp_buf*)exp->_jmpbuf)) == 0) {
  529. SQInteger res = sqstd_rex_list(exp);
  530. exp->_nodes[exp->_first].left = res;
  531. if(*exp->_p!='\0')
  532. sqstd_rex_error(exp,_SC("unexpected character"));
  533. #ifdef _DEBUG
  534. {
  535. SQInteger nsize,i;
  536. SQRexNode *t;
  537. nsize = exp->_nsize;
  538. t = &exp->_nodes[0];
  539. scprintf(_SC("\n"));
  540. for(i = 0;i < nsize; i++) {
  541. if(exp->_nodes[i].type>MAX_CHAR)
  542. scprintf(_SC("[%02d] %10s "),i,g_nnames[exp->_nodes[i].type-MAX_CHAR]);
  543. else
  544. scprintf(_SC("[%02d] %10c "),i,exp->_nodes[i].type);
  545. scprintf(_SC("left %02d right %02d next %02d\n"),exp->_nodes[i].left,exp->_nodes[i].right,exp->_nodes[i].next);
  546. }
  547. scprintf(_SC("\n"));
  548. }
  549. #endif
  550. exp->_matches = (SQRexMatch *) sq_malloc(exp->_nsubexpr * sizeof(SQRexMatch));
  551. memset(exp->_matches,0,exp->_nsubexpr * sizeof(SQRexMatch));
  552. }
  553. else{
  554. sqstd_rex_free(exp);
  555. return NULL;
  556. }
  557. return exp;
  558. }
  559. void sqstd_rex_free(SQRex *exp)
  560. {
  561. if(exp) {
  562. if(exp->_nodes) sq_free(exp->_nodes,exp->_nallocated * sizeof(SQRexNode));
  563. if(exp->_jmpbuf) sq_free(exp->_jmpbuf,sizeof(jmp_buf));
  564. if(exp->_matches) sq_free(exp->_matches,exp->_nsubexpr * sizeof(SQRexMatch));
  565. sq_free(exp,sizeof(SQRex));
  566. }
  567. }
  568. SQBool sqstd_rex_match(SQRex* exp,const SQChar* text)
  569. {
  570. const SQChar* res = NULL;
  571. exp->_bol = text;
  572. exp->_eol = text + scstrlen(text);
  573. exp->_currsubexp = 0;
  574. res = sqstd_rex_matchnode(exp,exp->_nodes,text,NULL);
  575. if(res == NULL || res != exp->_eol)
  576. return SQFalse;
  577. return SQTrue;
  578. }
  579. SQBool sqstd_rex_searchrange(SQRex* exp,const SQChar* text_begin,const SQChar* text_end,const SQChar** out_begin, const SQChar** out_end)
  580. {
  581. const SQChar *cur = NULL;
  582. SQInteger node = exp->_first;
  583. if(text_begin >= text_end) return SQFalse;
  584. exp->_bol = text_begin;
  585. exp->_eol = text_end;
  586. do {
  587. cur = text_begin;
  588. while(node != -1) {
  589. exp->_currsubexp = 0;
  590. cur = sqstd_rex_matchnode(exp,&exp->_nodes[node],cur,NULL);
  591. if(!cur)
  592. break;
  593. node = exp->_nodes[node].next;
  594. }
  595. text_begin++;
  596. } while(cur == NULL && text_begin != text_end);
  597. if(cur == NULL)
  598. return SQFalse;
  599. --text_begin;
  600. if(out_begin) *out_begin = text_begin;
  601. if(out_end) *out_end = cur;
  602. return SQTrue;
  603. }
  604. SQBool sqstd_rex_search(SQRex* exp,const SQChar* text, const SQChar** out_begin, const SQChar** out_end)
  605. {
  606. return sqstd_rex_searchrange(exp,text,text + scstrlen(text),out_begin,out_end);
  607. }
  608. SQInteger sqstd_rex_getsubexpcount(SQRex* exp)
  609. {
  610. return exp->_nsubexpr;
  611. }
  612. SQBool sqstd_rex_getsubexp(SQRex* exp, SQInteger n, SQRexMatch *subexp)
  613. {
  614. if( n<0 || n >= exp->_nsubexpr) return SQFalse;
  615. *subexp = exp->_matches[n];
  616. return SQTrue;
  617. }