Annotation of embedaddon/php/ext/intl/grapheme/grapheme_util.c, revision 1.1
1.1 ! misho 1: /*
! 2: +----------------------------------------------------------------------+
! 3: | PHP Version 5 |
! 4: +----------------------------------------------------------------------+
! 5: | This source file is subject to version 3.01 of the PHP license, |
! 6: | that is bundled with this package in the file LICENSE, and is |
! 7: | available through the world-wide-web at the following url: |
! 8: | http://www.php.net/license/3_01.txt |
! 9: | If you did not receive a copy of the PHP license and are unable to |
! 10: | obtain it through the world-wide-web, please send a note to |
! 11: | license@php.net so we can mail you a copy immediately. |
! 12: +----------------------------------------------------------------------+
! 13: | Author: Ed Batutis <ed@batutis.com> |
! 14: +----------------------------------------------------------------------+
! 15: */
! 16:
! 17: /* {{{ includes */
! 18: #ifdef HAVE_CONFIG_H
! 19: #include "config.h"
! 20: #endif
! 21:
! 22: #include <php.h>
! 23: #include "grapheme.h"
! 24: #include "grapheme_util.h"
! 25: #include "intl_common.h"
! 26:
! 27: #include <unicode/utypes.h>
! 28: #include <unicode/ucol.h>
! 29: #include <unicode/ustring.h>
! 30: #include <unicode/ubrk.h>
! 31:
! 32: #include "ext/standard/php_string.h"
! 33:
! 34: ZEND_EXTERN_MODULE_GLOBALS( intl )
! 35:
! 36: /* }}} */
! 37:
! 38: /* {{{ grapheme_close_global_iterator - clean up */
! 39: void
! 40: grapheme_close_global_iterator( TSRMLS_D )
! 41: {
! 42: UBreakIterator *global_break_iterator = INTL_G( grapheme_iterator );
! 43:
! 44: if ( NULL != global_break_iterator ) {
! 45: ubrk_close(global_break_iterator);
! 46: }
! 47: }
! 48: /* }}} */
! 49:
! 50: /* {{{ grapheme_intl_case_fold: convert string to lowercase */
! 51: void
! 52: grapheme_intl_case_fold(UChar** ptr_to_free, UChar **str, int32_t *str_len, UErrorCode *pstatus )
! 53: {
! 54: UChar *dest;
! 55: int32_t dest_len, size_required;
! 56:
! 57: /* allocate a destination string that is a bit larger than the src, hoping that is enough */
! 58: dest_len = (*str_len) + ( *str_len / 10 );
! 59: dest = (UChar*) eumalloc(dest_len);
! 60:
! 61: *pstatus = U_ZERO_ERROR;
! 62: size_required = u_strFoldCase(dest, dest_len, *str, *str_len, U_FOLD_CASE_DEFAULT, pstatus);
! 63:
! 64: dest_len = size_required;
! 65:
! 66: if ( U_BUFFER_OVERFLOW_ERROR == *pstatus ) {
! 67:
! 68: dest = (UChar*) eurealloc(dest, dest_len);
! 69:
! 70: *pstatus = U_ZERO_ERROR;
! 71: size_required = u_strFoldCase(dest, dest_len, *str, *str_len, U_FOLD_CASE_DEFAULT, pstatus);
! 72: }
! 73:
! 74: if ( U_FAILURE(*pstatus) ) {
! 75: return;
! 76: }
! 77:
! 78: if ( NULL != ptr_to_free) {
! 79: efree(*ptr_to_free);
! 80: *ptr_to_free = dest;
! 81: }
! 82:
! 83: *str = dest;
! 84: *str_len = dest_len;
! 85:
! 86: return;
! 87: }
! 88: /* }}} */
! 89:
! 90: /* {{{ grapheme_substr_ascii f='from' - starting point, l='length' */
! 91: void
! 92: grapheme_substr_ascii(char *str, int str_len, int f, int l, int argc, char **sub_str, int *sub_str_len)
! 93: {
! 94: *sub_str = NULL;
! 95:
! 96: if (argc > 2) {
! 97: if ((l < 0 && -l > str_len)) {
! 98: return;
! 99: } else if (l > str_len) {
! 100: l = str_len;
! 101: }
! 102: } else {
! 103: l = str_len;
! 104: }
! 105:
! 106: if (f > str_len || (f < 0 && -f > str_len)) {
! 107: return;
! 108: }
! 109:
! 110: if (l < 0 && (l + str_len - f) < 0) {
! 111: return;
! 112: }
! 113:
! 114: /* if "from" position is negative, count start position from the end
! 115: * of the string
! 116: */
! 117: if (f < 0) {
! 118: f = str_len + f;
! 119: if (f < 0) {
! 120: f = 0;
! 121: }
! 122: }
! 123:
! 124:
! 125: /* if "length" position is negative, set it to the length
! 126: * needed to stop that many chars from the end of the string
! 127: */
! 128: if (l < 0) {
! 129: l = (str_len - f) + l;
! 130: if (l < 0) {
! 131: l = 0;
! 132: }
! 133: }
! 134:
! 135: if (f >= str_len) {
! 136: return;
! 137: }
! 138:
! 139: if ((f + l) > str_len) {
! 140: l = str_len - f;
! 141: }
! 142:
! 143: *sub_str = str + f;
! 144: *sub_str_len = l;
! 145:
! 146: return;
! 147: }
! 148: /* }}} */
! 149:
! 150: /* {{{ grapheme_strrpos_utf16 - strrpos using utf16 */
! 151: int
! 152: grapheme_strrpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int f_ignore_case TSRMLS_DC)
! 153: {
! 154: UChar *uhaystack, *puhaystack, *uhaystack_end, *uneedle;
! 155: int32_t uhaystack_len, uneedle_len;
! 156: UErrorCode status;
! 157: unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
! 158: UBreakIterator* bi = NULL;
! 159: int ret_pos, pos;
! 160:
! 161: /* convert the strings to UTF-16. */
! 162: uhaystack = NULL;
! 163: uhaystack_len = 0;
! 164: status = U_ZERO_ERROR;
! 165: intl_convert_utf8_to_utf16(&uhaystack, &uhaystack_len, (char *) haystack, haystack_len, &status );
! 166:
! 167: if ( U_FAILURE( status ) ) {
! 168: /* Set global error code. */
! 169: intl_error_set_code( NULL, status TSRMLS_CC );
! 170:
! 171: /* Set error messages. */
! 172: intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
! 173: if (uhaystack) {
! 174: efree( uhaystack );
! 175: }
! 176: return -1;
! 177: }
! 178:
! 179: if ( f_ignore_case ) {
! 180: grapheme_intl_case_fold(&uhaystack, &uhaystack, &uhaystack_len, &status );
! 181: }
! 182:
! 183: /* get a pointer to the haystack taking into account the offset */
! 184: bi = NULL;
! 185: status = U_ZERO_ERROR;
! 186: bi = grapheme_get_break_iterator(u_break_iterator_buffer, &status TSRMLS_CC );
! 187:
! 188: puhaystack = grapheme_get_haystack_offset(bi, uhaystack, uhaystack_len, offset);
! 189:
! 190: if ( NULL == puhaystack ) {
! 191: intl_error_set( NULL, U_ILLEGAL_ARGUMENT_ERROR, "grapheme_strpos: Offset not contained in string", 1 TSRMLS_CC );
! 192: if (uhaystack) {
! 193: efree( uhaystack );
! 194: }
! 195: ubrk_close (bi);
! 196: return -1;
! 197: }
! 198:
! 199: uneedle = NULL;
! 200: uneedle_len = 0;
! 201: status = U_ZERO_ERROR;
! 202: intl_convert_utf8_to_utf16(&uneedle, &uneedle_len, (char *) needle, needle_len, &status );
! 203:
! 204: if ( U_FAILURE( status ) ) {
! 205: /* Set global error code. */
! 206: intl_error_set_code( NULL, status TSRMLS_CC );
! 207:
! 208: /* Set error messages. */
! 209: intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
! 210: if (uhaystack) {
! 211: efree( uhaystack );
! 212: }
! 213: if (uneedle) {
! 214: efree( uneedle );
! 215: }
! 216: ubrk_close (bi);
! 217: return -1;
! 218: }
! 219:
! 220: if ( f_ignore_case ) {
! 221: grapheme_intl_case_fold(&uneedle, &uneedle, &uneedle_len, &status );
! 222: }
! 223:
! 224: ret_pos = -1; /* -1 represents 'not found' */
! 225:
! 226: /* back up until there's needle_len characters to compare */
! 227:
! 228: uhaystack_end = uhaystack + uhaystack_len;
! 229: pos = ubrk_last(bi);
! 230: puhaystack = uhaystack + pos;
! 231:
! 232: while ( uhaystack_end - puhaystack < uneedle_len ) {
! 233:
! 234: pos = ubrk_previous(bi);
! 235:
! 236: if ( UBRK_DONE == pos ) {
! 237: break;
! 238: }
! 239:
! 240: puhaystack = uhaystack + pos;
! 241: }
! 242:
! 243: /* is there enough haystack left to hold the needle? */
! 244: if ( ( uhaystack_end - puhaystack ) < uneedle_len ) {
! 245: /* not enough, not found */
! 246: goto exit;
! 247: }
! 248:
! 249: while ( UBRK_DONE != pos ) {
! 250:
! 251: if (!u_memcmp(uneedle, puhaystack, uneedle_len)) { /* needle_len - 1 in zend memnstr? */
! 252:
! 253: /* does the grapheme in the haystack end at the same place as the last grapheme in the needle? */
! 254:
! 255: if ( ubrk_isBoundary(bi, pos + uneedle_len) ) {
! 256:
! 257: /* found it, get grapheme count offset */
! 258: ret_pos = grapheme_count_graphemes(bi, uhaystack, pos);
! 259: break;
! 260: }
! 261:
! 262: /* set position back */
! 263: ubrk_isBoundary(bi, pos);
! 264: }
! 265:
! 266: pos = ubrk_previous(bi);
! 267: puhaystack = uhaystack + pos;
! 268: }
! 269:
! 270: exit:
! 271: if (uhaystack) {
! 272: efree( uhaystack );
! 273: }
! 274: if (uneedle) {
! 275: efree( uneedle );
! 276: }
! 277: ubrk_close (bi);
! 278:
! 279: return ret_pos;
! 280: }
! 281:
! 282: /* }}} */
! 283:
! 284: /* {{{ grapheme_strpos_utf16 - strrpos using utf16*/
! 285: int
! 286: grapheme_strpos_utf16(unsigned char *haystack, int32_t haystack_len, unsigned char*needle, int32_t needle_len, int32_t offset, int32_t *puchar_pos, int f_ignore_case TSRMLS_DC)
! 287: {
! 288: UChar *uhaystack, *puhaystack, *uneedle;
! 289: int32_t uhaystack_len, uneedle_len;
! 290: int ret_pos;
! 291: unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
! 292: UBreakIterator* bi;
! 293: UErrorCode status;
! 294:
! 295: *puchar_pos = -1;
! 296:
! 297: /* convert the strings to UTF-16. */
! 298:
! 299: uhaystack = NULL;
! 300: uhaystack_len = 0;
! 301: status = U_ZERO_ERROR;
! 302: intl_convert_utf8_to_utf16(&uhaystack, &uhaystack_len, (char *) haystack, haystack_len, &status );
! 303:
! 304: if ( U_FAILURE( status ) ) {
! 305: /* Set global error code. */
! 306: intl_error_set_code( NULL, status TSRMLS_CC );
! 307:
! 308: /* Set error messages. */
! 309: intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
! 310: if (uhaystack) {
! 311: efree( uhaystack );
! 312: }
! 313: return -1;
! 314: }
! 315:
! 316: /* get a pointer to the haystack taking into account the offset */
! 317: bi = NULL;
! 318: status = U_ZERO_ERROR;
! 319: bi = grapheme_get_break_iterator(u_break_iterator_buffer, &status TSRMLS_CC );
! 320:
! 321: puhaystack = grapheme_get_haystack_offset(bi, uhaystack, uhaystack_len, offset);
! 322: uhaystack_len = (uhaystack_len - ( puhaystack - uhaystack));
! 323:
! 324: if ( NULL == puhaystack ) {
! 325:
! 326: intl_error_set( NULL, U_ILLEGAL_ARGUMENT_ERROR, "grapheme_strpos: Offset not contained in string", 1 TSRMLS_CC );
! 327: if (uhaystack) {
! 328: efree( uhaystack );
! 329: }
! 330: ubrk_close (bi);
! 331:
! 332: return -1;
! 333: }
! 334:
! 335: if ( f_ignore_case ) {
! 336: grapheme_intl_case_fold(&uhaystack, &puhaystack, &uhaystack_len, &status );
! 337: }
! 338:
! 339: uneedle = NULL;
! 340: uneedle_len = 0;
! 341: status = U_ZERO_ERROR;
! 342: intl_convert_utf8_to_utf16(&uneedle, &uneedle_len, (char *) needle, needle_len, &status );
! 343:
! 344: if ( U_FAILURE( status ) ) {
! 345: /* Set global error code. */
! 346: intl_error_set_code( NULL, status TSRMLS_CC );
! 347:
! 348: /* Set error messages. */
! 349: intl_error_set_custom_msg( NULL, "Error converting input string to UTF-16", 0 TSRMLS_CC );
! 350: if (uhaystack) {
! 351: efree( uhaystack );
! 352: }
! 353: if (uneedle) {
! 354: efree( uneedle );
! 355: }
! 356: ubrk_close (bi);
! 357:
! 358: return -1;
! 359: }
! 360:
! 361: if ( f_ignore_case ) {
! 362: grapheme_intl_case_fold(&uneedle, &uneedle, &uneedle_len, &status );
! 363: }
! 364:
! 365: ret_pos = grapheme_memnstr_grapheme(bi, puhaystack, uneedle, uneedle_len, puhaystack + uhaystack_len );
! 366:
! 367: *puchar_pos = ubrk_current(bi);
! 368:
! 369: if (uhaystack) {
! 370: efree( uhaystack );
! 371: }
! 372: if (uneedle) {
! 373: efree( uneedle );
! 374: }
! 375: ubrk_close (bi);
! 376:
! 377: return ret_pos;
! 378: }
! 379:
! 380: /* }}} */
! 381:
! 382: /* {{{ grapheme_ascii_check: ASCII check */
! 383: int grapheme_ascii_check(const unsigned char *day, int32_t len)
! 384: {
! 385: int ret_len = len;
! 386: while ( len-- ) {
! 387: if ( *day++ > 0x7f )
! 388: return -1;
! 389: }
! 390:
! 391: return ret_len;
! 392: }
! 393:
! 394: /* }}} */
! 395:
! 396: /* {{{ grapheme_split_string: find and optionally return grapheme boundaries */
! 397: int grapheme_split_string(const UChar *text, int32_t text_length, int boundary_array[], int boundary_array_len TSRMLS_DC )
! 398: {
! 399: unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
! 400: UErrorCode status = U_ZERO_ERROR;
! 401: int ret_len, pos;
! 402: UBreakIterator* bi;
! 403:
! 404: bi = grapheme_get_break_iterator((void*)u_break_iterator_buffer, &status TSRMLS_CC );
! 405:
! 406: if( U_FAILURE(status) ) {
! 407: return -1;
! 408: }
! 409:
! 410: ubrk_setText(bi, text, text_length, &status);
! 411:
! 412: pos = 0;
! 413:
! 414: for ( ret_len = 0; pos != UBRK_DONE; ) {
! 415:
! 416: pos = ubrk_next(bi);
! 417:
! 418: if ( pos != UBRK_DONE ) {
! 419:
! 420: if ( NULL != boundary_array && ret_len < boundary_array_len ) {
! 421: boundary_array[ret_len] = pos;
! 422: }
! 423:
! 424: ret_len++;
! 425: }
! 426: }
! 427:
! 428: ubrk_close(bi);
! 429:
! 430: return ret_len;
! 431: }
! 432: /* }}} */
! 433:
! 434: /* {{{ grapheme_count_graphemes */
! 435: int32_t
! 436: grapheme_count_graphemes(UBreakIterator *bi, UChar *string, int32_t string_len)
! 437: {
! 438: int ret_len = 0;
! 439: int pos = 0;
! 440: UErrorCode status = U_ZERO_ERROR;
! 441:
! 442: ubrk_setText(bi, string, string_len, &status);
! 443:
! 444: do {
! 445:
! 446: pos = ubrk_next(bi);
! 447:
! 448: if ( UBRK_DONE != pos ) {
! 449: ret_len++;
! 450: }
! 451:
! 452: } while ( UBRK_DONE != pos );
! 453:
! 454: return ret_len;
! 455: }
! 456: /* }}} */
! 457:
! 458: /* {{{ grapheme_memnstr_grapheme: find needle in haystack using grapheme boundaries */
! 459: int32_t
! 460: grapheme_memnstr_grapheme(UBreakIterator *bi, UChar *haystack, UChar *needle, int32_t needle_len, UChar *end)
! 461: {
! 462: UChar *p = haystack;
! 463: UChar ne = needle[needle_len-1];
! 464: UErrorCode status;
! 465: int32_t grapheme_offset;
! 466:
! 467: end -= needle_len;
! 468:
! 469: while (p <= end) {
! 470:
! 471: if ((p = u_memchr(p, *needle, (end-p+1))) && ne == p[needle_len-1]) {
! 472:
! 473: if (!u_memcmp(needle, p, needle_len - 1)) { /* needle_len - 1 works because if needle_len is 1, we've already tested the char */
! 474:
! 475: /* does the grapheme end here? */
! 476:
! 477: status = U_ZERO_ERROR;
! 478: ubrk_setText (bi, haystack, (end - haystack) + needle_len, &status);
! 479:
! 480: if ( ubrk_isBoundary (bi, (p - haystack) + needle_len) ) {
! 481:
! 482: /* found it, get grapheme count offset */
! 483: grapheme_offset = grapheme_count_graphemes(bi, haystack, (p - haystack));
! 484:
! 485: return grapheme_offset;
! 486: }
! 487: }
! 488: }
! 489:
! 490: if (p == NULL) {
! 491: return -1;
! 492: }
! 493:
! 494: p++;
! 495: }
! 496:
! 497: return -1;
! 498: }
! 499:
! 500: /* }}} */
! 501:
! 502: /* {{{ grapheme_memrstr_grapheme: reverse find needle in haystack using grapheme boundaries */
! 503: inline void *grapheme_memrchr_grapheme(const void *s, int c, int32_t n)
! 504: {
! 505: register unsigned char *e;
! 506:
! 507: if (n <= 0) {
! 508: return NULL;
! 509: }
! 510:
! 511: for (e = (unsigned char *)s + n - 1; e >= (unsigned char *)s; e--) {
! 512: if (*e == (unsigned char)c) {
! 513: return (void *)e;
! 514: }
! 515: }
! 516:
! 517: return NULL;
! 518: }
! 519: /* }}} */
! 520:
! 521: /* {{{ grapheme_get_haystack_offset - bump the haystack pointer based on the grapheme count offset */
! 522: UChar *
! 523: grapheme_get_haystack_offset(UBreakIterator* bi, UChar *uhaystack, int32_t uhaystack_len, int32_t offset)
! 524: {
! 525: UErrorCode status;
! 526: int32_t pos;
! 527: int32_t (*iter_op)(UBreakIterator* bi);
! 528: int iter_incr;
! 529:
! 530: if ( NULL != bi ) {
! 531: status = U_ZERO_ERROR;
! 532: ubrk_setText (bi, uhaystack, uhaystack_len, &status);
! 533: }
! 534:
! 535: if ( 0 == offset ) {
! 536: return uhaystack;
! 537: }
! 538:
! 539: if ( offset < 0 ) {
! 540: iter_op = ubrk_previous;
! 541: ubrk_last(bi); /* one past the end */
! 542: iter_incr = 1;
! 543: }
! 544: else {
! 545: iter_op = ubrk_next;
! 546: iter_incr = -1;
! 547: }
! 548:
! 549: pos = 0;
! 550:
! 551: while ( pos != UBRK_DONE && offset != 0 ) {
! 552:
! 553: pos = iter_op(bi);
! 554:
! 555: if ( UBRK_DONE != pos ) {
! 556: offset += iter_incr;
! 557: }
! 558: }
! 559:
! 560: if ( offset != 0 ) {
! 561: return NULL;
! 562: }
! 563:
! 564: return uhaystack + pos;
! 565: }
! 566: /* }}} */
! 567:
! 568: /* {{{ grapheme_strrpos_ascii: borrowed from the php ext/standard/string.c */
! 569: int32_t
! 570: grapheme_strrpos_ascii(unsigned char *haystack, int32_t haystack_len, unsigned char *needle, int32_t needle_len, int32_t offset)
! 571: {
! 572: unsigned char *p, *e;
! 573:
! 574: if (offset >= 0) {
! 575: p = haystack + offset;
! 576: e = haystack + haystack_len - needle_len;
! 577: } else {
! 578: p = haystack;
! 579: if (needle_len > -offset) {
! 580: e = haystack + haystack_len - needle_len;
! 581: } else {
! 582: e = haystack + haystack_len + offset;
! 583: }
! 584: }
! 585:
! 586: if (needle_len == 1) {
! 587: /* Single character search can shortcut memcmps */
! 588: while (e >= p) {
! 589: if (*e == *needle) {
! 590: return (e - p + (offset > 0 ? offset : 0));
! 591: }
! 592: e--;
! 593: }
! 594: return -1;
! 595: }
! 596:
! 597: while (e >= p) {
! 598: if (memcmp(e, needle, needle_len) == 0) {
! 599: return (e - p + (offset > 0 ? offset : 0));
! 600: }
! 601: e--;
! 602: }
! 603:
! 604: return -1;
! 605: }
! 606:
! 607: /* }}} */
! 608:
! 609: /* {{{ grapheme_get_break_iterator: get a clone of the global character break iterator */
! 610: UBreakIterator*
! 611: grapheme_get_break_iterator(void *stack_buffer, UErrorCode *status TSRMLS_DC )
! 612: {
! 613: int32_t buffer_size;
! 614:
! 615: UBreakIterator *global_break_iterator = INTL_G( grapheme_iterator );
! 616:
! 617: if ( NULL == global_break_iterator ) {
! 618:
! 619: global_break_iterator = ubrk_open(UBRK_CHARACTER,
! 620: NULL, /* icu default locale - locale has no effect on this iterator */
! 621: NULL, /* text not set in global iterator */
! 622: 0, /* text length = 0 */
! 623: status);
! 624:
! 625: INTL_G(grapheme_iterator) = global_break_iterator;
! 626: }
! 627:
! 628: buffer_size = U_BRK_SAFECLONE_BUFFERSIZE;
! 629:
! 630: return ubrk_safeClone(global_break_iterator, stack_buffer, &buffer_size, status);
! 631: }
! 632: /* }}} */
! 633:
! 634: /*
! 635: * Local variables:
! 636: * tab-width: 4
! 637: * c-basic-offset: 4
! 638: * End:
! 639: * vim600: fdm=marker
! 640: * vim: noet sw=4 ts=4
! 641: */
! 642:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>