123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119112011211122112311241125112611271128112911301131113211331134113511361137113811391140114111421143114411451146114711481149115011511152115311541155115611571158115911601161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218121912201221122212231224122512261227122812291230123112321233123412351236123712381239124012411242124312441245124612471248124912501251125212531254125512561257125812591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317131813191320132113221323132413251326132713281329133013311332133313341335133613371338133913401341134213431344134513461347134813491350135113521353135413551356135713581359136013611362136313641365136613671368136913701371137213731374137513761377137813791380138113821383138413851386138713881389139013911392139313941395139613971398139914001401140214031404140514061407140814091410141114121413141414151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470 |
- /*
- +----------------------------------------------------------------------+
- | PHP Version 5 |
- +----------------------------------------------------------------------+
- | Copyright (c) 1997-2016 The PHP Group |
- +----------------------------------------------------------------------+
- | This source file is subject to version 3.01 of the PHP license, |
- | that is bundled with this package in the file LICENSE, and is |
- | available through the world-wide-web at the following url: |
- | http://www.php.net/license/3_01.txt |
- | If you did not receive a copy of the PHP license and are unable to |
- | obtain it through the world-wide-web, please send a note to |
- | license@php.net so we can mail you a copy immediately. |
- +----------------------------------------------------------------------+
- | Author: Tsukada Takuya <tsukada@fminn.nagano.nagano.jp> |
- +----------------------------------------------------------------------+
- */
- /* $Id$ */
- #ifdef HAVE_CONFIG_H
- #include "config.h"
- #endif
- #include "php.h"
- #include "php_ini.h"
- #if HAVE_MBREGEX
- #include "ext/standard/php_smart_str.h"
- #include "ext/standard/info.h"
- #include "php_mbregex.h"
- #include "mbstring.h"
- #include "php_onig_compat.h" /* must come prior to the oniguruma header */
- #include <oniguruma.h>
- #undef UChar
- ZEND_EXTERN_MODULE_GLOBALS(mbstring)
- struct _zend_mb_regex_globals {
- OnigEncoding default_mbctype;
- OnigEncoding current_mbctype;
- HashTable ht_rc;
- zval *search_str;
- zval *search_str_val;
- unsigned int search_pos;
- php_mb_regex_t *search_re;
- OnigRegion *search_regs;
- OnigOptionType regex_default_options;
- OnigSyntaxType *regex_default_syntax;
- };
- #define MBREX(g) (MBSTRG(mb_regex_globals)->g)
- /* {{{ static void php_mb_regex_free_cache() */
- static void php_mb_regex_free_cache(php_mb_regex_t **pre)
- {
- onig_free(*pre);
- }
- /* }}} */
- /* {{{ _php_mb_regex_globals_ctor */
- static int _php_mb_regex_globals_ctor(zend_mb_regex_globals *pglobals TSRMLS_DC)
- {
- pglobals->default_mbctype = ONIG_ENCODING_UTF8;
- pglobals->current_mbctype = ONIG_ENCODING_UTF8;
- zend_hash_init(&(pglobals->ht_rc), 0, NULL, (void (*)(void *)) php_mb_regex_free_cache, 1);
- pglobals->search_str = (zval*) NULL;
- pglobals->search_re = (php_mb_regex_t*)NULL;
- pglobals->search_pos = 0;
- pglobals->search_regs = (OnigRegion*)NULL;
- pglobals->regex_default_options = ONIG_OPTION_MULTILINE | ONIG_OPTION_SINGLELINE;
- pglobals->regex_default_syntax = ONIG_SYNTAX_RUBY;
- return SUCCESS;
- }
- /* }}} */
- /* {{{ _php_mb_regex_globals_dtor */
- static void _php_mb_regex_globals_dtor(zend_mb_regex_globals *pglobals TSRMLS_DC)
- {
- zend_hash_destroy(&pglobals->ht_rc);
- }
- /* }}} */
- /* {{{ php_mb_regex_globals_alloc */
- zend_mb_regex_globals *php_mb_regex_globals_alloc(TSRMLS_D)
- {
- zend_mb_regex_globals *pglobals = pemalloc(
- sizeof(zend_mb_regex_globals), 1);
- if (!pglobals) {
- return NULL;
- }
- if (SUCCESS != _php_mb_regex_globals_ctor(pglobals TSRMLS_CC)) {
- pefree(pglobals, 1);
- return NULL;
- }
- return pglobals;
- }
- /* }}} */
- /* {{{ php_mb_regex_globals_free */
- void php_mb_regex_globals_free(zend_mb_regex_globals *pglobals TSRMLS_DC)
- {
- if (!pglobals) {
- return;
- }
- _php_mb_regex_globals_dtor(pglobals TSRMLS_CC);
- pefree(pglobals, 1);
- }
- /* }}} */
- /* {{{ PHP_MINIT_FUNCTION(mb_regex) */
- PHP_MINIT_FUNCTION(mb_regex)
- {
- onig_init();
- return SUCCESS;
- }
- /* }}} */
- /* {{{ PHP_MSHUTDOWN_FUNCTION(mb_regex) */
- PHP_MSHUTDOWN_FUNCTION(mb_regex)
- {
- onig_end();
- return SUCCESS;
- }
- /* }}} */
- /* {{{ PHP_RINIT_FUNCTION(mb_regex) */
- PHP_RINIT_FUNCTION(mb_regex)
- {
- return MBSTRG(mb_regex_globals) ? SUCCESS: FAILURE;
- }
- /* }}} */
- /* {{{ PHP_RSHUTDOWN_FUNCTION(mb_regex) */
- PHP_RSHUTDOWN_FUNCTION(mb_regex)
- {
- MBREX(current_mbctype) = MBREX(default_mbctype);
- if (MBREX(search_str) != NULL) {
- zval_ptr_dtor(&MBREX(search_str));
- MBREX(search_str) = (zval *)NULL;
- }
- MBREX(search_pos) = 0;
- if (MBREX(search_regs) != NULL) {
- onig_region_free(MBREX(search_regs), 1);
- MBREX(search_regs) = (OnigRegion *)NULL;
- }
- zend_hash_clean(&MBREX(ht_rc));
- return SUCCESS;
- }
- /* }}} */
- /* {{{ PHP_MINFO_FUNCTION(mb_regex) */
- PHP_MINFO_FUNCTION(mb_regex)
- {
- char buf[32];
- php_info_print_table_start();
- php_info_print_table_row(2, "Multibyte (japanese) regex support", "enabled");
- snprintf(buf, sizeof(buf), "%d.%d.%d",
- ONIGURUMA_VERSION_MAJOR,
- ONIGURUMA_VERSION_MINOR,
- ONIGURUMA_VERSION_TEENY);
- #ifdef PHP_ONIG_BUNDLED
- #ifdef USE_COMBINATION_EXPLOSION_CHECK
- php_info_print_table_row(2, "Multibyte regex (oniguruma) backtrack check", "On");
- #else /* USE_COMBINATION_EXPLOSION_CHECK */
- php_info_print_table_row(2, "Multibyte regex (oniguruma) backtrack check", "Off");
- #endif /* USE_COMBINATION_EXPLOSION_CHECK */
- #endif /* PHP_BUNDLED_ONIG */
- php_info_print_table_row(2, "Multibyte regex (oniguruma) version", buf);
- php_info_print_table_end();
- }
- /* }}} */
- /*
- * encoding name resolver
- */
- /* {{{ encoding name map */
- typedef struct _php_mb_regex_enc_name_map_t {
- const char *names;
- OnigEncoding code;
- } php_mb_regex_enc_name_map_t;
- php_mb_regex_enc_name_map_t enc_name_map[] = {
- #ifdef ONIG_ENCODING_EUC_JP
- {
- "EUC-JP\0EUCJP\0X-EUC-JP\0UJIS\0EUCJP\0EUCJP-WIN\0",
- ONIG_ENCODING_EUC_JP
- },
- #endif
- #ifdef ONIG_ENCODING_UTF8
- {
- "UTF-8\0UTF8\0",
- ONIG_ENCODING_UTF8
- },
- #endif
- #ifdef ONIG_ENCODING_UTF16_BE
- {
- "UTF-16\0UTF-16BE\0",
- ONIG_ENCODING_UTF16_BE
- },
- #endif
- #ifdef ONIG_ENCODING_UTF16_LE
- {
- "UTF-16LE\0",
- ONIG_ENCODING_UTF16_LE
- },
- #endif
- #ifdef ONIG_ENCODING_UTF32_BE
- {
- "UCS-4\0UTF-32\0UTF-32BE\0",
- ONIG_ENCODING_UTF32_BE
- },
- #endif
- #ifdef ONIG_ENCODING_UTF32_LE
- {
- "UCS-4LE\0UTF-32LE\0",
- ONIG_ENCODING_UTF32_LE
- },
- #endif
- #ifdef ONIG_ENCODING_SJIS
- {
- "SJIS\0CP932\0MS932\0SHIFT_JIS\0SJIS-WIN\0WINDOWS-31J\0",
- ONIG_ENCODING_SJIS
- },
- #endif
- #ifdef ONIG_ENCODING_BIG5
- {
- "BIG5\0BIG-5\0BIGFIVE\0CN-BIG5\0BIG-FIVE\0",
- ONIG_ENCODING_BIG5
- },
- #endif
- #ifdef ONIG_ENCODING_EUC_CN
- {
- "EUC-CN\0EUCCN\0EUC_CN\0GB-2312\0GB2312\0",
- ONIG_ENCODING_EUC_CN
- },
- #endif
- #ifdef ONIG_ENCODING_EUC_TW
- {
- "EUC-TW\0EUCTW\0EUC_TW\0",
- ONIG_ENCODING_EUC_TW
- },
- #endif
- #ifdef ONIG_ENCODING_EUC_KR
- {
- "EUC-KR\0EUCKR\0EUC_KR\0",
- ONIG_ENCODING_EUC_KR
- },
- #endif
- #if defined(ONIG_ENCODING_KOI8) && !PHP_ONIG_BAD_KOI8_ENTRY
- {
- "KOI8\0KOI-8\0",
- ONIG_ENCODING_KOI8
- },
- #endif
- #ifdef ONIG_ENCODING_KOI8_R
- {
- "KOI8R\0KOI8-R\0KOI-8R\0",
- ONIG_ENCODING_KOI8_R
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_1
- {
- "ISO-8859-1\0ISO8859-1\0ISO_8859_1\0ISO8859_1\0",
- ONIG_ENCODING_ISO_8859_1
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_2
- {
- "ISO-8859-2\0ISO8859-2\0ISO_8859_2\0ISO8859_2\0",
- ONIG_ENCODING_ISO_8859_2
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_3
- {
- "ISO-8859-3\0ISO8859-3\0ISO_8859_3\0ISO8859_3\0",
- ONIG_ENCODING_ISO_8859_3
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_4
- {
- "ISO-8859-4\0ISO8859-4\0ISO_8859_4\0ISO8859_4\0",
- ONIG_ENCODING_ISO_8859_4
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_5
- {
- "ISO-8859-5\0ISO8859-5\0ISO_8859_5\0ISO8859_5\0",
- ONIG_ENCODING_ISO_8859_5
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_6
- {
- "ISO-8859-6\0ISO8859-6\0ISO_8859_6\0ISO8859_6\0",
- ONIG_ENCODING_ISO_8859_6
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_7
- {
- "ISO-8859-7\0ISO8859-7\0ISO_8859_7\0ISO8859_7\0",
- ONIG_ENCODING_ISO_8859_7
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_8
- {
- "ISO-8859-8\0ISO8859-8\0ISO_8859_8\0ISO8859_8\0",
- ONIG_ENCODING_ISO_8859_8
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_9
- {
- "ISO-8859-9\0ISO8859-9\0ISO_8859_9\0ISO8859_9\0",
- ONIG_ENCODING_ISO_8859_9
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_10
- {
- "ISO-8859-10\0ISO8859-10\0ISO_8859_10\0ISO8859_10\0",
- ONIG_ENCODING_ISO_8859_10
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_11
- {
- "ISO-8859-11\0ISO8859-11\0ISO_8859_11\0ISO8859_11\0",
- ONIG_ENCODING_ISO_8859_11
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_13
- {
- "ISO-8859-13\0ISO8859-13\0ISO_8859_13\0ISO8859_13\0",
- ONIG_ENCODING_ISO_8859_13
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_14
- {
- "ISO-8859-14\0ISO8859-14\0ISO_8859_14\0ISO8859_14\0",
- ONIG_ENCODING_ISO_8859_14
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_15
- {
- "ISO-8859-15\0ISO8859-15\0ISO_8859_15\0ISO8859_15\0",
- ONIG_ENCODING_ISO_8859_15
- },
- #endif
- #ifdef ONIG_ENCODING_ISO_8859_16
- {
- "ISO-8859-16\0ISO8859-16\0ISO_8859_16\0ISO8859_16\0",
- ONIG_ENCODING_ISO_8859_16
- },
- #endif
- #ifdef ONIG_ENCODING_ASCII
- {
- "ASCII\0US-ASCII\0US_ASCII\0ISO646\0",
- ONIG_ENCODING_ASCII
- },
- #endif
- { NULL, ONIG_ENCODING_UNDEF }
- };
- /* }}} */
- /* {{{ php_mb_regex_name2mbctype */
- static OnigEncoding _php_mb_regex_name2mbctype(const char *pname)
- {
- const char *p;
- php_mb_regex_enc_name_map_t *mapping;
- if (pname == NULL || !*pname) {
- return ONIG_ENCODING_UNDEF;
- }
- for (mapping = enc_name_map; mapping->names != NULL; mapping++) {
- for (p = mapping->names; *p != '\0'; p += (strlen(p) + 1)) {
- if (strcasecmp(p, pname) == 0) {
- return mapping->code;
- }
- }
- }
- return ONIG_ENCODING_UNDEF;
- }
- /* }}} */
- /* {{{ php_mb_regex_mbctype2name */
- static const char *_php_mb_regex_mbctype2name(OnigEncoding mbctype)
- {
- php_mb_regex_enc_name_map_t *mapping;
- for (mapping = enc_name_map; mapping->names != NULL; mapping++) {
- if (mapping->code == mbctype) {
- return mapping->names;
- }
- }
- return NULL;
- }
- /* }}} */
- /* {{{ php_mb_regex_set_mbctype */
- int php_mb_regex_set_mbctype(const char *encname TSRMLS_DC)
- {
- OnigEncoding mbctype = _php_mb_regex_name2mbctype(encname);
- if (mbctype == ONIG_ENCODING_UNDEF) {
- return FAILURE;
- }
- MBREX(current_mbctype) = mbctype;
- return SUCCESS;
- }
- /* }}} */
- /* {{{ php_mb_regex_set_default_mbctype */
- int php_mb_regex_set_default_mbctype(const char *encname TSRMLS_DC)
- {
- OnigEncoding mbctype = _php_mb_regex_name2mbctype(encname);
- if (mbctype == ONIG_ENCODING_UNDEF) {
- return FAILURE;
- }
- MBREX(default_mbctype) = mbctype;
- return SUCCESS;
- }
- /* }}} */
- /* {{{ php_mb_regex_get_mbctype */
- const char *php_mb_regex_get_mbctype(TSRMLS_D)
- {
- return _php_mb_regex_mbctype2name(MBREX(current_mbctype));
- }
- /* }}} */
- /* {{{ php_mb_regex_get_default_mbctype */
- const char *php_mb_regex_get_default_mbctype(TSRMLS_D)
- {
- return _php_mb_regex_mbctype2name(MBREX(default_mbctype));
- }
- /* }}} */
- /*
- * regex cache
- */
- /* {{{ php_mbregex_compile_pattern */
- static php_mb_regex_t *php_mbregex_compile_pattern(const char *pattern, int patlen, OnigOptionType options, OnigEncoding enc, OnigSyntaxType *syntax TSRMLS_DC)
- {
- int err_code = 0;
- int found = 0;
- php_mb_regex_t *retval = NULL, **rc = NULL;
- OnigErrorInfo err_info;
- OnigUChar err_str[ONIG_MAX_ERROR_MESSAGE_LEN];
- found = zend_hash_find(&MBREX(ht_rc), (char *)pattern, patlen+1, (void **) &rc);
- if (found == FAILURE || (*rc)->options != options || (*rc)->enc != enc || (*rc)->syntax != syntax) {
- if ((err_code = onig_new(&retval, (OnigUChar *)pattern, (OnigUChar *)(pattern + patlen), options, enc, syntax, &err_info)) != ONIG_NORMAL) {
- onig_error_code_to_str(err_str, err_code, &err_info);
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "mbregex compile err: %s", err_str);
- retval = NULL;
- goto out;
- }
- zend_hash_update(&MBREX(ht_rc), (char *) pattern, patlen + 1, (void *) &retval, sizeof(retval), NULL);
- } else if (found == SUCCESS) {
- retval = *rc;
- }
- out:
- return retval;
- }
- /* }}} */
- /* {{{ _php_mb_regex_get_option_string */
- static size_t _php_mb_regex_get_option_string(char *str, size_t len, OnigOptionType option, OnigSyntaxType *syntax)
- {
- size_t len_left = len;
- size_t len_req = 0;
- char *p = str;
- char c;
- if ((option & ONIG_OPTION_IGNORECASE) != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 'i';
- }
- ++len_req;
- }
- if ((option & ONIG_OPTION_EXTEND) != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 'x';
- }
- ++len_req;
- }
- if ((option & (ONIG_OPTION_MULTILINE | ONIG_OPTION_SINGLELINE)) ==
- (ONIG_OPTION_MULTILINE | ONIG_OPTION_SINGLELINE)) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 'p';
- }
- ++len_req;
- } else {
- if ((option & ONIG_OPTION_MULTILINE) != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 'm';
- }
- ++len_req;
- }
- if ((option & ONIG_OPTION_SINGLELINE) != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 's';
- }
- ++len_req;
- }
- }
- if ((option & ONIG_OPTION_FIND_LONGEST) != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 'l';
- }
- ++len_req;
- }
- if ((option & ONIG_OPTION_FIND_NOT_EMPTY) != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = 'n';
- }
- ++len_req;
- }
- c = 0;
- if (syntax == ONIG_SYNTAX_JAVA) {
- c = 'j';
- } else if (syntax == ONIG_SYNTAX_GNU_REGEX) {
- c = 'u';
- } else if (syntax == ONIG_SYNTAX_GREP) {
- c = 'g';
- } else if (syntax == ONIG_SYNTAX_EMACS) {
- c = 'c';
- } else if (syntax == ONIG_SYNTAX_RUBY) {
- c = 'r';
- } else if (syntax == ONIG_SYNTAX_PERL) {
- c = 'z';
- } else if (syntax == ONIG_SYNTAX_POSIX_BASIC) {
- c = 'b';
- } else if (syntax == ONIG_SYNTAX_POSIX_EXTENDED) {
- c = 'd';
- }
- if (c != 0) {
- if (len_left > 0) {
- --len_left;
- *(p++) = c;
- }
- ++len_req;
- }
- if (len_left > 0) {
- --len_left;
- *(p++) = '\0';
- }
- ++len_req;
- if (len < len_req) {
- return len_req;
- }
- return 0;
- }
- /* }}} */
- /* {{{ _php_mb_regex_init_options */
- static void
- _php_mb_regex_init_options(const char *parg, int narg, OnigOptionType *option, OnigSyntaxType **syntax, int *eval)
- {
- int n;
- char c;
- int optm = 0;
- *syntax = ONIG_SYNTAX_RUBY;
- if (parg != NULL) {
- n = 0;
- while(n < narg) {
- c = parg[n++];
- switch (c) {
- case 'i':
- optm |= ONIG_OPTION_IGNORECASE;
- break;
- case 'x':
- optm |= ONIG_OPTION_EXTEND;
- break;
- case 'm':
- optm |= ONIG_OPTION_MULTILINE;
- break;
- case 's':
- optm |= ONIG_OPTION_SINGLELINE;
- break;
- case 'p':
- optm |= ONIG_OPTION_MULTILINE | ONIG_OPTION_SINGLELINE;
- break;
- case 'l':
- optm |= ONIG_OPTION_FIND_LONGEST;
- break;
- case 'n':
- optm |= ONIG_OPTION_FIND_NOT_EMPTY;
- break;
- case 'j':
- *syntax = ONIG_SYNTAX_JAVA;
- break;
- case 'u':
- *syntax = ONIG_SYNTAX_GNU_REGEX;
- break;
- case 'g':
- *syntax = ONIG_SYNTAX_GREP;
- break;
- case 'c':
- *syntax = ONIG_SYNTAX_EMACS;
- break;
- case 'r':
- *syntax = ONIG_SYNTAX_RUBY;
- break;
- case 'z':
- *syntax = ONIG_SYNTAX_PERL;
- break;
- case 'b':
- *syntax = ONIG_SYNTAX_POSIX_BASIC;
- break;
- case 'd':
- *syntax = ONIG_SYNTAX_POSIX_EXTENDED;
- break;
- case 'e':
- if (eval != NULL) *eval = 1;
- break;
- default:
- break;
- }
- }
- if (option != NULL) *option|=optm;
- }
- }
- /* }}} */
- /*
- * php functions
- */
- /* {{{ proto string mb_regex_encoding([string encoding])
- Returns the current encoding for regex as a string. */
- PHP_FUNCTION(mb_regex_encoding)
- {
- size_t argc = ZEND_NUM_ARGS();
- char *encoding;
- int encoding_len;
- OnigEncoding mbctype;
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|s", &encoding, &encoding_len) == FAILURE) {
- return;
- }
- if (argc == 0) {
- const char *retval = _php_mb_regex_mbctype2name(MBREX(current_mbctype));
- if (retval == NULL) {
- RETURN_FALSE;
- }
- RETURN_STRING((char *)retval, 1);
- } else if (argc == 1) {
- mbctype = _php_mb_regex_name2mbctype(encoding);
- if (mbctype == ONIG_ENCODING_UNDEF) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Unknown encoding \"%s\"", encoding);
- RETURN_FALSE;
- }
- MBREX(current_mbctype) = mbctype;
- RETURN_TRUE;
- }
- }
- /* }}} */
- /* {{{ _php_mb_regex_ereg_exec */
- static void _php_mb_regex_ereg_exec(INTERNAL_FUNCTION_PARAMETERS, int icase)
- {
- zval **arg_pattern, *array;
- char *string;
- int string_len;
- php_mb_regex_t *re;
- OnigRegion *regs = NULL;
- int i, match_len, beg, end;
- OnigOptionType options;
- char *str;
- array = NULL;
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "Zs|z", &arg_pattern, &string, &string_len, &array) == FAILURE) {
- RETURN_FALSE;
- }
- options = MBREX(regex_default_options);
- if (icase) {
- options |= ONIG_OPTION_IGNORECASE;
- }
- /* compile the regular expression from the supplied regex */
- if (Z_TYPE_PP(arg_pattern) != IS_STRING) {
- /* we convert numbers to integers and treat them as a string */
- if (Z_TYPE_PP(arg_pattern) == IS_DOUBLE) {
- convert_to_long_ex(arg_pattern); /* get rid of decimal places */
- }
- convert_to_string_ex(arg_pattern);
- /* don't bother doing an extended regex with just a number */
- }
- if (!Z_STRVAL_PP(arg_pattern) || Z_STRLEN_PP(arg_pattern) == 0) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "empty pattern");
- RETVAL_FALSE;
- goto out;
- }
- re = php_mbregex_compile_pattern(Z_STRVAL_PP(arg_pattern), Z_STRLEN_PP(arg_pattern), options, MBREX(current_mbctype), MBREX(regex_default_syntax) TSRMLS_CC);
- if (re == NULL) {
- RETVAL_FALSE;
- goto out;
- }
- regs = onig_region_new();
- /* actually execute the regular expression */
- if (onig_search(re, (OnigUChar *)string, (OnigUChar *)(string + string_len), (OnigUChar *)string, (OnigUChar *)(string + string_len), regs, 0) < 0) {
- RETVAL_FALSE;
- goto out;
- }
- match_len = 1;
- str = string;
- if (array != NULL) {
- match_len = regs->end[0] - regs->beg[0];
- zval_dtor(array);
- array_init(array);
- for (i = 0; i < regs->num_regs; i++) {
- beg = regs->beg[i];
- end = regs->end[i];
- if (beg >= 0 && beg < end && end <= string_len) {
- add_index_stringl(array, i, (char *)&str[beg], end - beg, 1);
- } else {
- add_index_bool(array, i, 0);
- }
- }
- }
- if (match_len == 0) {
- match_len = 1;
- }
- RETVAL_LONG(match_len);
- out:
- if (regs != NULL) {
- onig_region_free(regs, 1);
- }
- }
- /* }}} */
- /* {{{ proto int mb_ereg(string pattern, string string [, array registers])
- Regular expression match for multibyte string */
- PHP_FUNCTION(mb_ereg)
- {
- _php_mb_regex_ereg_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 0);
- }
- /* }}} */
- /* {{{ proto int mb_eregi(string pattern, string string [, array registers])
- Case-insensitive regular expression match for multibyte string */
- PHP_FUNCTION(mb_eregi)
- {
- _php_mb_regex_ereg_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 1);
- }
- /* }}} */
- /* {{{ _php_mb_regex_ereg_replace_exec */
- static void _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAMETERS, OnigOptionType options, int is_callable)
- {
- zval **arg_pattern_zval;
- char *arg_pattern;
- int arg_pattern_len;
- char *replace;
- int replace_len;
- zend_fcall_info arg_replace_fci;
- zend_fcall_info_cache arg_replace_fci_cache;
- char *string;
- int string_len;
- char *p;
- php_mb_regex_t *re;
- OnigSyntaxType *syntax;
- OnigRegion *regs = NULL;
- smart_str out_buf = { 0 };
- smart_str eval_buf = { 0 };
- smart_str *pbuf;
- int i, err, eval, n;
- OnigUChar *pos;
- OnigUChar *string_lim;
- char *description = NULL;
- char pat_buf[6];
- const mbfl_encoding *enc;
- {
- const char *current_enc_name;
- current_enc_name = _php_mb_regex_mbctype2name(MBREX(current_mbctype));
- if (current_enc_name == NULL ||
- (enc = mbfl_name2encoding(current_enc_name)) == NULL) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Unknown error");
- RETURN_FALSE;
- }
- }
- eval = 0;
- {
- char *option_str = NULL;
- int option_str_len = 0;
- if (!is_callable) {
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "Zss|s",
- &arg_pattern_zval,
- &replace, &replace_len,
- &string, &string_len,
- &option_str, &option_str_len) == FAILURE) {
- RETURN_FALSE;
- }
- } else {
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "Zfs|s",
- &arg_pattern_zval,
- &arg_replace_fci, &arg_replace_fci_cache,
- &string, &string_len,
- &option_str, &option_str_len) == FAILURE) {
- RETURN_FALSE;
- }
- }
- if (option_str != NULL) {
- _php_mb_regex_init_options(option_str, option_str_len, &options, &syntax, &eval);
- } else {
- options |= MBREX(regex_default_options);
- syntax = MBREX(regex_default_syntax);
- }
- }
- if (Z_TYPE_PP(arg_pattern_zval) == IS_STRING) {
- arg_pattern = Z_STRVAL_PP(arg_pattern_zval);
- arg_pattern_len = Z_STRLEN_PP(arg_pattern_zval);
- } else {
- /* FIXME: this code is not multibyte aware! */
- convert_to_long_ex(arg_pattern_zval);
- pat_buf[0] = (char)Z_LVAL_PP(arg_pattern_zval);
- pat_buf[1] = '\0';
- pat_buf[2] = '\0';
- pat_buf[3] = '\0';
- pat_buf[4] = '\0';
- pat_buf[5] = '\0';
- arg_pattern = pat_buf;
- arg_pattern_len = 1;
- }
- /* create regex pattern buffer */
- re = php_mbregex_compile_pattern(arg_pattern, arg_pattern_len, options, MBREX(current_mbctype), syntax TSRMLS_CC);
- if (re == NULL) {
- RETURN_FALSE;
- }
- if (eval || is_callable) {
- pbuf = &eval_buf;
- description = zend_make_compiled_string_description("mbregex replace" TSRMLS_CC);
- } else {
- pbuf = &out_buf;
- description = NULL;
- }
- if (is_callable) {
- if (eval) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Option 'e' cannot be used with replacement callback");
- RETURN_FALSE;
- }
- }
- /* do the actual work */
- err = 0;
- pos = (OnigUChar *)string;
- string_lim = (OnigUChar*)(string + string_len);
- regs = onig_region_new();
- while (err >= 0) {
- err = onig_search(re, (OnigUChar *)string, (OnigUChar *)string_lim, pos, (OnigUChar *)string_lim, regs, 0);
- if (err <= -2) {
- OnigUChar err_str[ONIG_MAX_ERROR_MESSAGE_LEN];
- onig_error_code_to_str(err_str, err);
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "mbregex search failure in php_mbereg_replace_exec(): %s", err_str);
- break;
- }
- if (err >= 0) {
- #if moriyoshi_0
- if (regs->beg[0] == regs->end[0]) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Empty regular expression");
- break;
- }
- #endif
- /* copy the part of the string before the match */
- smart_str_appendl(&out_buf, pos, (size_t)((OnigUChar *)(string + regs->beg[0]) - pos));
- if (!is_callable) {
- /* copy replacement and backrefs */
- i = 0;
- p = replace;
- while (i < replace_len) {
- int fwd = (int) php_mb_mbchar_bytes_ex(p, enc);
- n = -1;
- if ((replace_len - i) >= 2 && fwd == 1 &&
- p[0] == '\\' && p[1] >= '0' && p[1] <= '9') {
- n = p[1] - '0';
- }
- if (n >= 0 && n < regs->num_regs) {
- if (regs->beg[n] >= 0 && regs->beg[n] < regs->end[n] && regs->end[n] <= string_len) {
- smart_str_appendl(pbuf, string + regs->beg[n], regs->end[n] - regs->beg[n]);
- }
- p += 2;
- i += 2;
- } else {
- smart_str_appendl(pbuf, p, fwd);
- p += fwd;
- i += fwd;
- }
- }
- }
- if (eval) {
- zval v;
- /* null terminate buffer */
- smart_str_0(&eval_buf);
- /* do eval */
- if (zend_eval_stringl(eval_buf.c, eval_buf.len, &v, description TSRMLS_CC) == FAILURE) {
- efree(description);
- php_error_docref(NULL TSRMLS_CC,E_ERROR, "Failed evaluating code: %s%s", PHP_EOL, eval_buf.c);
- /* zend_error() does not return in this case */
- }
- /* result of eval */
- convert_to_string(&v);
- smart_str_appendl(&out_buf, Z_STRVAL(v), Z_STRLEN(v));
- /* Clean up */
- eval_buf.len = 0;
- zval_dtor(&v);
- } else if (is_callable) {
- zval *retval_ptr = NULL;
- zval **args[1];
- zval *subpats;
- int i;
- MAKE_STD_ZVAL(subpats);
- array_init(subpats);
- for (i = 0; i < regs->num_regs; i++) {
- add_next_index_stringl(subpats, string + regs->beg[i], regs->end[i] - regs->beg[i], 1);
- }
- args[0] = &subpats;
- /* null terminate buffer */
- smart_str_0(&eval_buf);
- arg_replace_fci.param_count = 1;
- arg_replace_fci.params = args;
- arg_replace_fci.retval_ptr_ptr = &retval_ptr;
- if (zend_call_function(&arg_replace_fci, &arg_replace_fci_cache TSRMLS_CC) == SUCCESS && arg_replace_fci.retval_ptr_ptr && retval_ptr) {
- convert_to_string_ex(&retval_ptr);
- smart_str_appendl(&out_buf, Z_STRVAL_P(retval_ptr), Z_STRLEN_P(retval_ptr));
- eval_buf.len = 0;
- zval_ptr_dtor(&retval_ptr);
- } else {
- if (!EG(exception)) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Unable to call custom replacement function");
- }
- }
- zval_ptr_dtor(&subpats);
- }
- n = regs->end[0];
- if ((pos - (OnigUChar *)string) < n) {
- pos = (OnigUChar *)string + n;
- } else {
- if (pos < string_lim) {
- smart_str_appendl(&out_buf, pos, 1);
- }
- pos++;
- }
- } else { /* nomatch */
- /* stick that last bit of string on our output */
- if (string_lim - pos > 0) {
- smart_str_appendl(&out_buf, pos, string_lim - pos);
- }
- }
- onig_region_free(regs, 0);
- }
- if (description) {
- efree(description);
- }
- if (regs != NULL) {
- onig_region_free(regs, 1);
- }
- smart_str_free(&eval_buf);
- if (err <= -2) {
- smart_str_free(&out_buf);
- RETVAL_FALSE;
- } else {
- smart_str_appendc(&out_buf, '\0');
- RETVAL_STRINGL((char *)out_buf.c, out_buf.len - 1, 0);
- }
- }
- /* }}} */
- /* {{{ proto string mb_ereg_replace(string pattern, string replacement, string string [, string option])
- Replace regular expression for multibyte string */
- PHP_FUNCTION(mb_ereg_replace)
- {
- _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 0, 0);
- }
- /* }}} */
- /* {{{ proto string mb_eregi_replace(string pattern, string replacement, string string)
- Case insensitive replace regular expression for multibyte string */
- PHP_FUNCTION(mb_eregi_replace)
- {
- _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, ONIG_OPTION_IGNORECASE, 0);
- }
- /* }}} */
- /* {{{ proto string mb_ereg_replace_callback(string pattern, string callback, string string [, string option])
- regular expression for multibyte string using replacement callback */
- PHP_FUNCTION(mb_ereg_replace_callback)
- {
- _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 0, 1);
- }
- /* }}} */
- /* {{{ proto array mb_split(string pattern, string string [, int limit])
- split multibyte string into array by regular expression */
- PHP_FUNCTION(mb_split)
- {
- char *arg_pattern;
- int arg_pattern_len;
- php_mb_regex_t *re;
- OnigRegion *regs = NULL;
- char *string;
- OnigUChar *pos, *chunk_pos;
- int string_len;
- int n, err;
- long count = -1;
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "ss|l", &arg_pattern, &arg_pattern_len, &string, &string_len, &count) == FAILURE) {
- RETURN_FALSE;
- }
- if (count > 0) {
- count--;
- }
- /* create regex pattern buffer */
- if ((re = php_mbregex_compile_pattern(arg_pattern, arg_pattern_len, MBREX(regex_default_options), MBREX(current_mbctype), MBREX(regex_default_syntax) TSRMLS_CC)) == NULL) {
- RETURN_FALSE;
- }
- array_init(return_value);
- chunk_pos = pos = (OnigUChar *)string;
- err = 0;
- regs = onig_region_new();
- /* churn through str, generating array entries as we go */
- while (count != 0 && (pos - (OnigUChar *)string) < string_len) {
- int beg, end;
- err = onig_search(re, (OnigUChar *)string, (OnigUChar *)(string + string_len), pos, (OnigUChar *)(string + string_len), regs, 0);
- if (err < 0) {
- break;
- }
- beg = regs->beg[0], end = regs->end[0];
- /* add it to the array */
- if ((pos - (OnigUChar *)string) < end) {
- if (beg < string_len && beg >= (chunk_pos - (OnigUChar *)string)) {
- add_next_index_stringl(return_value, (char *)chunk_pos, ((OnigUChar *)(string + beg) - chunk_pos), 1);
- --count;
- } else {
- err = -2;
- break;
- }
- /* point at our new starting point */
- chunk_pos = pos = (OnigUChar *)string + end;
- } else {
- pos++;
- }
- onig_region_free(regs, 0);
- }
- onig_region_free(regs, 1);
- /* see if we encountered an error */
- if (err <= -2) {
- OnigUChar err_str[ONIG_MAX_ERROR_MESSAGE_LEN];
- onig_error_code_to_str(err_str, err);
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "mbregex search failure in mbsplit(): %s", err_str);
- zval_dtor(return_value);
- RETURN_FALSE;
- }
- /* otherwise we just have one last element to add to the array */
- n = ((OnigUChar *)(string + string_len) - chunk_pos);
- if (n > 0) {
- add_next_index_stringl(return_value, (char *)chunk_pos, n, 1);
- } else {
- add_next_index_stringl(return_value, "", 0, 1);
- }
- }
- /* }}} */
- /* {{{ proto bool mb_ereg_match(string pattern, string string [,string option])
- Regular expression match for multibyte string */
- PHP_FUNCTION(mb_ereg_match)
- {
- char *arg_pattern;
- int arg_pattern_len;
- char *string;
- int string_len;
- php_mb_regex_t *re;
- OnigSyntaxType *syntax;
- OnigOptionType option = 0;
- int err;
- {
- char *option_str = NULL;
- int option_str_len = 0;
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "ss|s",
- &arg_pattern, &arg_pattern_len, &string, &string_len,
- &option_str, &option_str_len)==FAILURE) {
- RETURN_FALSE;
- }
- if (option_str != NULL) {
- _php_mb_regex_init_options(option_str, option_str_len, &option, &syntax, NULL);
- } else {
- option |= MBREX(regex_default_options);
- syntax = MBREX(regex_default_syntax);
- }
- }
- if ((re = php_mbregex_compile_pattern(arg_pattern, arg_pattern_len, option, MBREX(current_mbctype), syntax TSRMLS_CC)) == NULL) {
- RETURN_FALSE;
- }
- /* match */
- err = onig_match(re, (OnigUChar *)string, (OnigUChar *)(string + string_len), (OnigUChar *)string, NULL, 0);
- if (err >= 0) {
- RETVAL_TRUE;
- } else {
- RETVAL_FALSE;
- }
- }
- /* }}} */
- /* regex search */
- /* {{{ _php_mb_regex_ereg_search_exec */
- static void
- _php_mb_regex_ereg_search_exec(INTERNAL_FUNCTION_PARAMETERS, int mode)
- {
- size_t argc = ZEND_NUM_ARGS();
- char *arg_pattern, *arg_options;
- int arg_pattern_len, arg_options_len;
- int n, i, err, pos, len, beg, end;
- OnigOptionType option;
- OnigUChar *str;
- OnigSyntaxType *syntax;
- if (zend_parse_parameters(argc TSRMLS_CC, "|ss", &arg_pattern, &arg_pattern_len, &arg_options, &arg_options_len) == FAILURE) {
- return;
- }
- option = MBREX(regex_default_options);
- if (argc == 2) {
- option = 0;
- _php_mb_regex_init_options(arg_options, arg_options_len, &option, &syntax, NULL);
- }
- if (argc > 0) {
- /* create regex pattern buffer */
- if ((MBREX(search_re) = php_mbregex_compile_pattern(arg_pattern, arg_pattern_len, option, MBREX(current_mbctype), MBREX(regex_default_syntax) TSRMLS_CC)) == NULL) {
- RETURN_FALSE;
- }
- }
- pos = MBREX(search_pos);
- str = NULL;
- len = 0;
- if (MBREX(search_str) != NULL && Z_TYPE_P(MBREX(search_str)) == IS_STRING){
- str = (OnigUChar *)Z_STRVAL_P(MBREX(search_str));
- len = Z_STRLEN_P(MBREX(search_str));
- }
- if (MBREX(search_re) == NULL) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "No regex given");
- RETURN_FALSE;
- }
- if (str == NULL) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "No string given");
- RETURN_FALSE;
- }
- if (MBREX(search_regs)) {
- onig_region_free(MBREX(search_regs), 1);
- }
- MBREX(search_regs) = onig_region_new();
- err = onig_search(MBREX(search_re), str, str + len, str + pos, str + len, MBREX(search_regs), 0);
- if (err == ONIG_MISMATCH) {
- MBREX(search_pos) = len;
- RETVAL_FALSE;
- } else if (err <= -2) {
- OnigUChar err_str[ONIG_MAX_ERROR_MESSAGE_LEN];
- onig_error_code_to_str(err_str, err);
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "mbregex search failure in mbregex_search(): %s", err_str);
- RETVAL_FALSE;
- } else {
- switch (mode) {
- case 1:
- array_init(return_value);
- beg = MBREX(search_regs)->beg[0];
- end = MBREX(search_regs)->end[0];
- add_next_index_long(return_value, beg);
- add_next_index_long(return_value, end - beg);
- break;
- case 2:
- array_init(return_value);
- n = MBREX(search_regs)->num_regs;
- for (i = 0; i < n; i++) {
- beg = MBREX(search_regs)->beg[i];
- end = MBREX(search_regs)->end[i];
- if (beg >= 0 && beg <= end && end <= len) {
- add_index_stringl(return_value, i, (char *)&str[beg], end - beg, 1);
- } else {
- add_index_bool(return_value, i, 0);
- }
- }
- break;
- default:
- RETVAL_TRUE;
- break;
- }
- end = MBREX(search_regs)->end[0];
- if (pos <= end) {
- MBREX(search_pos) = end;
- } else {
- MBREX(search_pos) = pos + 1;
- }
- }
- if (err < 0) {
- onig_region_free(MBREX(search_regs), 1);
- MBREX(search_regs) = (OnigRegion *)NULL;
- }
- }
- /* }}} */
- /* {{{ proto bool mb_ereg_search([string pattern[, string option]])
- Regular expression search for multibyte string */
- PHP_FUNCTION(mb_ereg_search)
- {
- _php_mb_regex_ereg_search_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 0);
- }
- /* }}} */
- /* {{{ proto array mb_ereg_search_pos([string pattern[, string option]])
- Regular expression search for multibyte string */
- PHP_FUNCTION(mb_ereg_search_pos)
- {
- _php_mb_regex_ereg_search_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 1);
- }
- /* }}} */
- /* {{{ proto array mb_ereg_search_regs([string pattern[, string option]])
- Regular expression search for multibyte string */
- PHP_FUNCTION(mb_ereg_search_regs)
- {
- _php_mb_regex_ereg_search_exec(INTERNAL_FUNCTION_PARAM_PASSTHRU, 2);
- }
- /* }}} */
- /* {{{ proto bool mb_ereg_search_init(string string [, string pattern[, string option]])
- Initialize string and regular expression for search. */
- PHP_FUNCTION(mb_ereg_search_init)
- {
- size_t argc = ZEND_NUM_ARGS();
- zval *arg_str;
- char *arg_pattern = NULL, *arg_options = NULL;
- int arg_pattern_len = 0, arg_options_len = 0;
- OnigSyntaxType *syntax = NULL;
- OnigOptionType option;
- if (zend_parse_parameters(argc TSRMLS_CC, "z|ss", &arg_str, &arg_pattern, &arg_pattern_len, &arg_options, &arg_options_len) == FAILURE) {
- return;
- }
- if (argc > 1 && arg_pattern_len == 0) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Empty pattern");
- RETURN_FALSE;
- }
- option = MBREX(regex_default_options);
- syntax = MBREX(regex_default_syntax);
- if (argc == 3) {
- option = 0;
- _php_mb_regex_init_options(arg_options, arg_options_len, &option, &syntax, NULL);
- }
- if (argc > 1) {
- /* create regex pattern buffer */
- if ((MBREX(search_re) = php_mbregex_compile_pattern(arg_pattern, arg_pattern_len, option, MBREX(current_mbctype), syntax TSRMLS_CC)) == NULL) {
- RETURN_FALSE;
- }
- }
- if (MBREX(search_str) != NULL) {
- zval_ptr_dtor(&MBREX(search_str));
- MBREX(search_str) = (zval *)NULL;
- }
- MBREX(search_str) = arg_str;
- Z_ADDREF_P(MBREX(search_str));
- SEPARATE_ZVAL_IF_NOT_REF(&MBREX(search_str));
- MBREX(search_pos) = 0;
- if (MBREX(search_regs) != NULL) {
- onig_region_free(MBREX(search_regs), 1);
- MBREX(search_regs) = (OnigRegion *) NULL;
- }
- RETURN_TRUE;
- }
- /* }}} */
- /* {{{ proto array mb_ereg_search_getregs(void)
- Get matched substring of the last time */
- PHP_FUNCTION(mb_ereg_search_getregs)
- {
- int n, i, len, beg, end;
- OnigUChar *str;
- if (MBREX(search_regs) != NULL && Z_TYPE_P(MBREX(search_str)) == IS_STRING && Z_STRVAL_P(MBREX(search_str)) != NULL) {
- array_init(return_value);
- str = (OnigUChar *)Z_STRVAL_P(MBREX(search_str));
- len = Z_STRLEN_P(MBREX(search_str));
- n = MBREX(search_regs)->num_regs;
- for (i = 0; i < n; i++) {
- beg = MBREX(search_regs)->beg[i];
- end = MBREX(search_regs)->end[i];
- if (beg >= 0 && beg <= end && end <= len) {
- add_index_stringl(return_value, i, (char *)&str[beg], end - beg, 1);
- } else {
- add_index_bool(return_value, i, 0);
- }
- }
- } else {
- RETVAL_FALSE;
- }
- }
- /* }}} */
- /* {{{ proto int mb_ereg_search_getpos(void)
- Get search start position */
- PHP_FUNCTION(mb_ereg_search_getpos)
- {
- RETVAL_LONG(MBREX(search_pos));
- }
- /* }}} */
- /* {{{ proto bool mb_ereg_search_setpos(int position)
- Set search start position */
- PHP_FUNCTION(mb_ereg_search_setpos)
- {
- long position;
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "l", &position) == FAILURE) {
- return;
- }
- if (position < 0 || (MBREX(search_str) != NULL && Z_TYPE_P(MBREX(search_str)) == IS_STRING && position > Z_STRLEN_P(MBREX(search_str)))) {
- php_error_docref(NULL TSRMLS_CC, E_WARNING, "Position is out of range");
- MBREX(search_pos) = 0;
- RETURN_FALSE;
- }
- MBREX(search_pos) = position;
- RETURN_TRUE;
- }
- /* }}} */
- /* {{{ php_mb_regex_set_options */
- static void _php_mb_regex_set_options(OnigOptionType options, OnigSyntaxType *syntax, OnigOptionType *prev_options, OnigSyntaxType **prev_syntax TSRMLS_DC)
- {
- if (prev_options != NULL) {
- *prev_options = MBREX(regex_default_options);
- }
- if (prev_syntax != NULL) {
- *prev_syntax = MBREX(regex_default_syntax);
- }
- MBREX(regex_default_options) = options;
- MBREX(regex_default_syntax) = syntax;
- }
- /* }}} */
- /* {{{ proto string mb_regex_set_options([string options])
- Set or get the default options for mbregex functions */
- PHP_FUNCTION(mb_regex_set_options)
- {
- OnigOptionType opt;
- OnigSyntaxType *syntax;
- char *string = NULL;
- int string_len;
- char buf[16];
- if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|s",
- &string, &string_len) == FAILURE) {
- RETURN_FALSE;
- }
- if (string != NULL) {
- opt = 0;
- syntax = NULL;
- _php_mb_regex_init_options(string, string_len, &opt, &syntax, NULL);
- _php_mb_regex_set_options(opt, syntax, NULL, NULL TSRMLS_CC);
- } else {
- opt = MBREX(regex_default_options);
- syntax = MBREX(regex_default_syntax);
- }
- _php_mb_regex_get_option_string(buf, sizeof(buf), opt, syntax);
- RETVAL_STRING(buf, 1);
- }
- /* }}} */
- #endif /* HAVE_MBREGEX */
- /*
- * Local variables:
- * tab-width: 4
- * c-basic-offset: 4
- * End:
- * vim600: fdm=marker
- * vim: noet sw=4 ts=4
- */
|