ILIAS  release_5-4 Revision v5.4.26-12-gabc799a52e6
UtfNormal.php
Go to the documentation of this file.
1<?php
2# Copyright (C) 2004 Brion Vibber <brion@pobox.com>
3# http://www.mediawiki.org/
4#
5# This program is free software; you can redistribute it and/or modify
6# it under the terms of the GNU General Public License as published by
7# the Free Software Foundation; either version 2 of the License, or
8# (at your option) any later version.
9#
10# This program is distributed in the hope that it will be useful,
11# but WITHOUT ANY WARRANTY; without even the implied warranty of
12# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13# GNU General Public License for more details.
14#
15# You should have received a copy of the GNU General Public License along
16# with this program; if not, write to the Free Software Foundation, Inc.,
17# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18# http://www.gnu.org/copyleft/gpl.html
19
21require_once dirname(__FILE__) . '/UtfNormalUtil.php';
22
27
28# Load compatibility decompositions on demand if they are needed.
31
32define('UNICODE_HANGUL_FIRST', 0xac00);
33define('UNICODE_HANGUL_LAST', 0xd7a3);
34
35define('UNICODE_HANGUL_LBASE', 0x1100);
36define('UNICODE_HANGUL_VBASE', 0x1161);
37define('UNICODE_HANGUL_TBASE', 0x11a7);
38
39define('UNICODE_HANGUL_LCOUNT', 19);
40define('UNICODE_HANGUL_VCOUNT', 21);
41define('UNICODE_HANGUL_TCOUNT', 28);
42define('UNICODE_HANGUL_NCOUNT', UNICODE_HANGUL_VCOUNT * UNICODE_HANGUL_TCOUNT);
43
44define('UNICODE_HANGUL_LEND', UNICODE_HANGUL_LBASE + UNICODE_HANGUL_LCOUNT - 1);
45define('UNICODE_HANGUL_VEND', UNICODE_HANGUL_VBASE + UNICODE_HANGUL_VCOUNT - 1);
46define('UNICODE_HANGUL_TEND', UNICODE_HANGUL_TBASE + UNICODE_HANGUL_TCOUNT - 1);
47
48define('UNICODE_SURROGATE_FIRST', 0xd800);
49define('UNICODE_SURROGATE_LAST', 0xdfff);
50define('UNICODE_MAX', 0x10ffff);
51define('UNICODE_REPLACEMENT', 0xfffd);
52
53
54define('UTF8_HANGUL_FIRST', "\xea\xb0\x80" /*codepointToUtf8( UNICODE_HANGUL_FIRST )*/);
55define('UTF8_HANGUL_LAST', "\xed\x9e\xa3" /*codepointToUtf8( UNICODE_HANGUL_LAST )*/);
56
57define('UTF8_HANGUL_LBASE', "\xe1\x84\x80" /*codepointToUtf8( UNICODE_HANGUL_LBASE )*/);
58define('UTF8_HANGUL_VBASE', "\xe1\x85\xa1" /*codepointToUtf8( UNICODE_HANGUL_VBASE )*/);
59define('UTF8_HANGUL_TBASE', "\xe1\x86\xa7" /*codepointToUtf8( UNICODE_HANGUL_TBASE )*/);
60
61define('UTF8_HANGUL_LEND', "\xe1\x84\x92" /*codepointToUtf8( UNICODE_HANGUL_LEND )*/);
62define('UTF8_HANGUL_VEND', "\xe1\x85\xb5" /*codepointToUtf8( UNICODE_HANGUL_VEND )*/);
63define('UTF8_HANGUL_TEND', "\xe1\x87\x82" /*codepointToUtf8( UNICODE_HANGUL_TEND )*/);
64
65define('UTF8_SURROGATE_FIRST', "\xed\xa0\x80" /*codepointToUtf8( UNICODE_SURROGATE_FIRST )*/);
66define('UTF8_SURROGATE_LAST', "\xed\xbf\xbf" /*codepointToUtf8( UNICODE_SURROGATE_LAST )*/);
67define('UTF8_MAX', "\xf4\x8f\xbf\xbf" /*codepointToUtf8( UNICODE_MAX )*/);
68define('UTF8_REPLACEMENT', "\xef\xbf\xbd" /*codepointToUtf8( UNICODE_REPLACEMENT )*/);
69#define( 'UTF8_REPLACEMENT', '!' );
70
71define('UTF8_OVERLONG_A', "\xc1\xbf");
72define('UTF8_OVERLONG_B', "\xe0\x9f\xbf");
73define('UTF8_OVERLONG_C', "\xf0\x8f\xbf\xbf");
74
75# These two ranges are illegal
76define('UTF8_FDD0', "\xef\xb7\x90" /*codepointToUtf8( 0xfdd0 )*/);
77define('UTF8_FDEF', "\xef\xb7\xaf" /*codepointToUtf8( 0xfdef )*/);
78define('UTF8_FFFE', "\xef\xbf\xbe" /*codepointToUtf8( 0xfffe )*/);
79define('UTF8_FFFF', "\xef\xbf\xbf" /*codepointToUtf8( 0xffff )*/);
80
81define('UTF8_HEAD', false);
82define('UTF8_TAIL', true);
83
84
88define('UNORM_NONE', 1);
89define('UNORM_NFD', 2);
90define('UNORM_NFKD', 3);
91define('UNORM_NFC', 4);
92define('UNORM_DEFAULT', UNORM_NFC);
93define('UNORM_NFKC', 5);
94define('UNORM_FCD', 6);
95
96define('NORMALIZE_ICU', function_exists('utf8_normalize'));
97
113{
125 public static function cleanUp($string)
126 {
127 if (NORMALIZE_ICU) {
128 # We exclude a few chars that ICU would not.
129 $string = preg_replace(
130 '/[\x00-\x08\x0b\x0c\x0e-\x1f]/',
132 $string
133 );
134 $string = str_replace(UTF8_FFFE, UTF8_REPLACEMENT, $string);
135 $string = str_replace(UTF8_FFFF, UTF8_REPLACEMENT, $string);
136
137 # UnicodeString constructor fails if the string ends with a
138 # head byte. Add a junk char at the end, we'll strip it off.
139 return rtrim(utf8_normalize($string . "\x01", UNORM_NFC), "\x01");
140 } elseif (UtfNormal::quickIsNFCVerify($string)) {
141 # Side effect -- $string has had UTF-8 errors cleaned up.
142 return $string;
143 } else {
144 return UtfNormal::NFC($string);
145 }
146 }
147
157 public static function toNFC($string)
158 {
159 if (NORMALIZE_ICU) {
160 return utf8_normalize($string, UNORM_NFC);
161 } elseif (UtfNormal::quickIsNFC($string)) {
162 return $string;
163 } else {
164 return UtfNormal::NFC($string);
165 }
166 }
167
176 public static function toNFD($string)
177 {
178 if (NORMALIZE_ICU) {
179 return utf8_normalize($string, UNORM_NFD);
180 } elseif (preg_match('/[\x80-\xff]/', $string)) {
181 return UtfNormal::NFD($string);
182 } else {
183 return $string;
184 }
185 }
186
196 public static function toNFKC($string)
197 {
198 if (NORMALIZE_ICU) {
199 return utf8_normalize($string, UNORM_NFKC);
200 } elseif (preg_match('/[\x80-\xff]/', $string)) {
201 return UtfNormal::NFKC($string);
202 } else {
203 return $string;
204 }
205 }
206
216 public static function toNFKD($string)
217 {
218 if (NORMALIZE_ICU) {
219 return utf8_normalize($string, UNORM_NFKD);
220 } elseif (preg_match('/[\x80-\xff]/', $string)) {
221 return UtfNormal::NFKD($string);
222 } else {
223 return $string;
224 }
225 }
226
232 public static function loadData()
233 {
234 global $utfCombiningClass;
235 if (!isset($utfCombiningClass)) {
236 require_once(dirname(__FILE__) . '/UtfNormalData.inc');
237 }
238 }
239
247 public static function quickIsNFC($string)
248 {
249 # ASCII is always valid NFC!
250 # If it's pure ASCII, let it through.
251 if (!preg_match('/[\x80-\xff]/', $string)) {
252 return true;
253 }
254
256 global $utfCheckNFC, $utfCombiningClass;
257 $len = strlen($string);
258 for ($i = 0; $i < $len; $i++) {
259 $c = $string[$i];
260 $n = ord($c);
261 if ($n < 0x80) {
262 continue;
263 } elseif ($n >= 0xf0) {
264 $c = substr($string, $i, 4);
265 $i += 3;
266 } elseif ($n >= 0xe0) {
267 $c = substr($string, $i, 3);
268 $i += 2;
269 } elseif ($n >= 0xc0) {
270 $c = substr($string, $i, 2);
271 $i++;
272 }
273 if (isset($utfCheckNFC[$c])) {
274 # If it's NO or MAYBE, bail and do the slow check.
275 return false;
276 }
277 if (isset($utfCombiningClass[$c])) {
278 # Combining character? We might have to do sorting, at least.
279 return false;
280 }
281 }
282 return true;
283 }
284
291 public static function quickIsNFCVerify(&$string)
292 {
293 # Screen out some characters that eg won't be allowed in XML
294 $string = preg_replace('/[\x00-\x08\x0b\x0c\x0e-\x1f]/', UTF8_REPLACEMENT, $string);
295
296 # ASCII is always valid NFC!
297 # If we're only ever given plain ASCII, we can avoid the overhead
298 # of initializing the decomposition tables by skipping out early.
299 if (!preg_match('/[\x80-\xff]/', $string)) {
300 return true;
301 }
302
303 static $checkit = null, $tailBytes = null, $utfCheckOrCombining = null;
304 if (!isset($checkit)) {
305 # Load/build some scary lookup tables...
307 global $utfCheckNFC, $utfCombiningClass;
308
309 $utfCheckOrCombining = array_merge($utfCheckNFC, $utfCombiningClass);
310
311 # Head bytes for sequences which we should do further validity checks
312 $checkit = array_flip(array_map(
313 'chr',
314 array( 0xc0, 0xc1, 0xe0, 0xed, 0xef,
315 0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7,
316 0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff )
317 ));
318
319 # Each UTF-8 head byte is followed by a certain
320 # number of tail bytes.
321 $tailBytes = array();
322 for ($n = 0; $n < 256; $n++) {
323 if ($n < 0xc0) {
324 $remaining = 0;
325 } elseif ($n < 0xe0) {
326 $remaining = 1;
327 } elseif ($n < 0xf0) {
328 $remaining = 2;
329 } elseif ($n < 0xf8) {
330 $remaining = 3;
331 } elseif ($n < 0xfc) {
332 $remaining = 4;
333 } elseif ($n < 0xfe) {
334 $remaining = 5;
335 } else {
336 $remaining = 0;
337 }
338 $tailBytes[chr($n)] = $remaining;
339 }
340 }
341
342 # Chop the text into pure-ASCII and non-ASCII areas;
343 # large ASCII parts can be handled much more quickly.
344 # Don't chop up Unicode areas for punctuation, though,
345 # that wastes energy.
346 $matches = array();
347 preg_match_all(
348 '/([\x00-\x7f]+|[\x80-\xff][\x00-\x40\x5b-\x5f\x7b-\xff]*)/',
349 $string,
350 $matches
351 );
352
353 $looksNormal = true;
354 $base = 0;
355 $replace = array();
356 foreach ($matches[1] as $str) {
357 $chunk = strlen($str);
358
359 if ($str[0] < "\x80") {
360 # ASCII chunk: guaranteed to be valid UTF-8
361 # and in normal form C, so skip over it.
362 $base += $chunk;
363 continue;
364 }
365
366 # We'll have to examine the chunk byte by byte to ensure
367 # that it consists of valid UTF-8 sequences, and to see
368 # if any of them might not be normalized.
369 #
370 # Since PHP is not the fastest language on earth, some of
371 # this code is a little ugly with inner loop optimizations.
372
373 $head = '';
374 $len = $chunk + 1; # Counting down is faster. I'm *so* sorry.
375
376 for ($i = -1; --$len;) {
377 if ($remaining = $tailBytes[$c = $str[++$i]]) {
378 # UTF-8 head byte!
379 $sequence = $head = $c;
380 do {
381 # Look for the defined number of tail bytes...
382 if (--$len && ($c = $str[++$i]) >= "\x80" && $c < "\xc0") {
383 # Legal tail bytes are nice.
384 $sequence .= $c;
385 } else {
386 if (0 == $len) {
387 # Premature end of string!
388 # Drop a replacement character into output to
389 # represent the invalid UTF-8 sequence.
390 $replace[] = array( UTF8_REPLACEMENT,
391 $base + $i + 1 - strlen($sequence),
392 strlen($sequence) );
393 break 2;
394 } else {
395 # Illegal tail byte; abandon the sequence.
396 $replace[] = array( UTF8_REPLACEMENT,
397 $base + $i - strlen($sequence),
398 strlen($sequence) );
399 # Back up and reprocess this byte; it may itself
400 # be a legal ASCII or UTF-8 sequence head.
401 --$i;
402 ++$len;
403 continue 2;
404 }
405 }
406 } while (--$remaining);
407
408 if (isset($checkit[$head])) {
409 # Do some more detailed validity checks, for
410 # invalid characters and illegal sequences.
411 if ($head == "\xed") {
412 # 0xed is relatively frequent in Korean, which
413 # abuts the surrogate area, so we're doing
414 # this check separately to speed things up.
415
416 if ($sequence >= UTF8_SURROGATE_FIRST) {
417 # Surrogates are legal only in UTF-16 code.
418 # They are totally forbidden here in UTF-8
419 # utopia.
420 $replace[] = array( UTF8_REPLACEMENT,
421 $base + $i + 1 - strlen($sequence),
422 strlen($sequence) );
423 $head = '';
424 continue;
425 }
426 } else {
427 # Slower, but rarer checks...
428 $n = ord($head);
429 if (
430 # "Overlong sequences" are those that are syntactically
431 # correct but use more UTF-8 bytes than are necessary to
432 # encode a character. Naïve string comparisons can be
433 # tricked into failing to see a match for an ASCII
434 # character, for instance, which can be a security hole
435 # if blacklist checks are being used.
436 ($n < 0xc2 && $sequence <= UTF8_OVERLONG_A)
437 || ($n == 0xe0 && $sequence <= UTF8_OVERLONG_B)
438 || ($n == 0xf0 && $sequence <= UTF8_OVERLONG_C)
439
440 # U+FFFE and U+FFFF are explicitly forbidden in Unicode.
441 || ($n == 0xef &&
442 ($sequence == UTF8_FFFE)
443 || ($sequence == UTF8_FFFF))
444
445 # Unicode has been limited to 21 bits; longer
446 # sequences are not allowed.
447 || ($n >= 0xf0 && $sequence > UTF8_MAX)) {
448 $replace[] = array( UTF8_REPLACEMENT,
449 $base + $i + 1 - strlen($sequence),
450 strlen($sequence) );
451 $head = '';
452 continue;
453 }
454 }
455 }
456
457 if (isset($utfCheckOrCombining[$sequence])) {
458 # If it's NO or MAYBE, we'll have to rip
459 # the string apart and put it back together.
460 # That's going to be mighty slow.
461 $looksNormal = false;
462 }
463
464 # The sequence is legal!
465 $head = '';
466 } elseif ($c < "\x80") {
467 # ASCII byte.
468 $head = '';
469 } elseif ($c < "\xc0") {
470 # Illegal tail bytes
471 if ($head == '') {
472 # Out of the blue!
473 $replace[] = array( UTF8_REPLACEMENT, $base + $i, 1 );
474 } else {
475 # Don't add if we're continuing a broken sequence;
476 # we already put a replacement character when we looked
477 # at the broken sequence.
478 $replace[] = array( '', $base + $i, 1 );
479 }
480 } else {
481 # Miscellaneous freaks.
482 $replace[] = array( UTF8_REPLACEMENT, $base + $i, 1 );
483 $head = '';
484 }
485 }
486 $base += $chunk;
487 }
488 if (count($replace)) {
489 # There were illegal UTF-8 sequences we need to fix up.
490 $out = '';
491 $last = 0;
492 foreach ($replace as $rep) {
493 list($replacement, $start, $length) = $rep;
494 if ($last < $start) {
495 $out .= substr($string, $last, $start - $last);
496 }
497 $out .= $replacement;
498 $last = $start + $length;
499 }
500 if ($last < strlen($string)) {
501 $out .= substr($string, $last);
502 }
503 $string = $out;
504 }
505 return $looksNormal;
506 }
507
508 # These take a string and run the normalization on them, without
509 # checking for validity or any optimization etc. Input must be
510 # VALID UTF-8!
517 public static function NFC($string)
518 {
519 return UtfNormal::fastCompose(UtfNormal::NFD($string));
520 }
521
528 public static function NFD($string)
529 {
531 global $utfCanonicalDecomp;
534 );
535 }
536
543 public static function NFKC($string)
544 {
546 }
547
554 public static function NFKD($string)
555 {
557 if (!isset($utfCompatibilityDecomp)) {
558 require_once('UtfNormalDataK.inc');
559 }
562 );
563 }
564
565
576 public static function fastDecompose($string, $map)
577 {
579 $len = strlen($string);
580 $out = '';
581 for ($i = 0; $i < $len; $i++) {
582 $c = $string[$i];
583 $n = ord($c);
584 if ($n < 0x80) {
585 # ASCII chars never decompose
586 # THEY ARE IMMORTAL
587 $out .= $c;
588 continue;
589 } elseif ($n >= 0xf0) {
590 $c = substr($string, $i, 4);
591 $i += 3;
592 } elseif ($n >= 0xe0) {
593 $c = substr($string, $i, 3);
594 $i += 2;
595 } elseif ($n >= 0xc0) {
596 $c = substr($string, $i, 2);
597 $i++;
598 }
599 if (isset($map[$c])) {
600 $out .= $map[$c];
601 continue;
602 } else {
604 # Decompose a hangul syllable into jamo;
605 # hardcoded for three-byte UTF-8 sequence.
606 # A lookup table would be slightly faster,
607 # but adds a lot of memory & disk needs.
608 #
609 $index = ((ord($c[0]) & 0x0f) << 12
610 | (ord($c[1]) & 0x3f) << 6
611 | (ord($c[2]) & 0x3f))
613 $l = intval($index / UNICODE_HANGUL_NCOUNT);
616 $out .= "\xe1\x84" . chr(0x80 + $l) . "\xe1\x85" . chr(0xa1 + $v);
617 if ($t >= 25) {
618 $out .= "\xe1\x87" . chr(0x80 + $t - 25);
619 } elseif ($t) {
620 $out .= "\xe1\x86" . chr(0xa7 + $t);
621 }
622 continue;
623 }
624 }
625 $out .= $c;
626 }
627 return $out;
628 }
629
638 public static function fastCombiningSort($string)
639 {
641 global $utfCombiningClass;
642 $len = strlen($string);
643 $out = '';
644 $combiners = array();
645 $lastClass = -1;
646 for ($i = 0; $i < $len; $i++) {
647 $c = $string[$i];
648 $n = ord($c);
649 if ($n >= 0x80) {
650 if ($n >= 0xf0) {
651 $c = substr($string, $i, 4);
652 $i += 3;
653 } elseif ($n >= 0xe0) {
654 $c = substr($string, $i, 3);
655 $i += 2;
656 } elseif ($n >= 0xc0) {
657 $c = substr($string, $i, 2);
658 $i++;
659 }
660 if (isset($utfCombiningClass[$c])) {
661 $lastClass = $utfCombiningClass[$c];
662 if (isset($combiners[$lastClass])) {
663 $combiners[$lastClass] .= $c;
664 } else {
665 $combiners[$lastClass] = $c;
666 }
667 continue;
668 }
669 }
670 if ($lastClass) {
671 ksort($combiners);
672 $out .= implode('', $combiners);
673 $combiners = array();
674 }
675 $out .= $c;
676 $lastClass = 0;
677 }
678 if ($lastClass) {
679 ksort($combiners);
680 $out .= implode('', $combiners);
681 }
682 return $out;
683 }
684
693 public static function fastCompose($string)
694 {
697 $len = strlen($string);
698 $out = '';
699 $lastClass = -1;
700 $lastHangul = 0;
701 $startChar = '';
702 $combining = '';
703 $x1 = ord(substr(UTF8_HANGUL_VBASE, 0, 1));
704 $x2 = ord(substr(UTF8_HANGUL_TEND, 0, 1));
705 for ($i = 0; $i < $len; $i++) {
706 $c = $string[$i];
707 $n = ord($c);
708 if ($n < 0x80) {
709 # No combining characters here...
710 $out .= $startChar;
711 $out .= $combining;
712 $startChar = $c;
713 $combining = '';
714 $lastClass = 0;
715 continue;
716 } elseif ($n >= 0xf0) {
717 $c = substr($string, $i, 4);
718 $i += 3;
719 } elseif ($n >= 0xe0) {
720 $c = substr($string, $i, 3);
721 $i += 2;
722 } elseif ($n >= 0xc0) {
723 $c = substr($string, $i, 2);
724 $i++;
725 }
726 $pair = $startChar . $c;
727 if ($n > 0x80) {
728 if (isset($utfCombiningClass[$c])) {
729 # A combining char; see what we can do with it
730 $class = $utfCombiningClass[$c];
731 if (!empty($startChar) &&
732 $lastClass < $class &&
733 $class > 0 &&
734 isset($utfCanonicalComp[$pair])) {
735 $startChar = $utfCanonicalComp[$pair];
736 $class = 0;
737 } else {
738 $combining .= $c;
739 }
740 $lastClass = $class;
741 $lastHangul = 0;
742 continue;
743 }
744 }
745 # New start char
746 if ($lastClass == 0) {
747 if (isset($utfCanonicalComp[$pair])) {
748 $startChar = $utfCanonicalComp[$pair];
749 $lastHangul = 0;
750 continue;
751 }
752 if ($n >= $x1 && $n <= $x2) {
753 # WARNING: Hangul code is painfully slow.
754 # I apologize for this ugly, ugly code; however
755 # performance is even more teh suck if we call
756 # out to nice clean functions. Lookup tables are
757 # marginally faster, but require a lot of space.
758 #
759 if ($c >= UTF8_HANGUL_VBASE &&
760 $c <= UTF8_HANGUL_VEND &&
761 $startChar >= UTF8_HANGUL_LBASE &&
762 $startChar <= UTF8_HANGUL_LEND) {
763 #
764 #$lIndex = utf8ToCodepoint( $startChar ) - UNICODE_HANGUL_LBASE;
765 #$vIndex = utf8ToCodepoint( $c ) - UNICODE_HANGUL_VBASE;
766 $lIndex = ord($startChar[2]) - 0x80;
767 $vIndex = ord($c[2]) - 0xa1;
768
769 $hangulPoint = UNICODE_HANGUL_FIRST +
771 (UNICODE_HANGUL_VCOUNT * $lIndex + $vIndex);
772
773 # Hardcode the limited-range UTF-8 conversion:
774 $startChar = chr($hangulPoint >> 12 & 0x0f | 0xe0) .
775 chr($hangulPoint >> 6 & 0x3f | 0x80) .
776 chr($hangulPoint & 0x3f | 0x80);
777 $lastHangul = 0;
778 continue;
779 } elseif ($c >= UTF8_HANGUL_TBASE &&
780 $c <= UTF8_HANGUL_TEND &&
781 $startChar >= UTF8_HANGUL_FIRST &&
782 $startChar <= UTF8_HANGUL_LAST &&
783 !$lastHangul) {
784 # $tIndex = utf8ToCodepoint( $c ) - UNICODE_HANGUL_TBASE;
785 $tIndex = ord($c[2]) - 0xa7;
786 if ($tIndex < 0) {
787 $tIndex = ord($c[2]) - 0x80 + (0x11c0 - 0x11a7);
788 }
789
790 # Increment the code point by $tIndex, without
791 # the function overhead of decoding and recoding UTF-8
792 #
793 $tail = ord($startChar[2]) + $tIndex;
794 if ($tail > 0xbf) {
795 $tail -= 0x40;
796 $mid = ord($startChar[1]) + 1;
797 if ($mid > 0xbf) {
798 $startChar[0] = chr(ord($startChar[0]) + 1);
799 $mid -= 0x40;
800 }
801 $startChar[1] = chr($mid);
802 }
803 $startChar[2] = chr($tail);
804
805 # If there's another jamo char after this, *don't* try to merge it.
806 $lastHangul = 1;
807 continue;
808 }
809 }
810 }
811 $out .= $startChar;
812 $out .= $combining;
813 $startChar = $c;
814 $combining = '';
815 $lastClass = 0;
816 $lastHangul = 0;
817 }
818 $out .= $startChar . $combining;
819 return $out;
820 }
821
829 public static function placebo($string)
830 {
831 $len = strlen($string);
832 $out = '';
833 for ($i = 0; $i < $len; $i++) {
834 $out .= $string[$i];
835 }
836 return $out;
837 }
838}
$n
Definition: RandomTest.php:85
const UTF8_HANGUL_TEND
Definition: UtfNormal.php:63
const UTF8_OVERLONG_C
Definition: UtfNormal.php:73
const UNICODE_HANGUL_VCOUNT
Definition: UtfNormal.php:40
global $utfCombiningClass
Definition: UtfNormal.php:23
const UTF8_MAX
Definition: UtfNormal.php:67
const UTF8_HANGUL_LEND
Definition: UtfNormal.php:61
const UNORM_NFC
Definition: UtfNormal.php:91
const UTF8_SURROGATE_FIRST
Definition: UtfNormal.php:65
const UNICODE_HANGUL_LCOUNT
Definition: UtfNormal.php:39
const UTF8_HANGUL_TBASE
Definition: UtfNormal.php:59
const UNORM_NFKC
Definition: UtfNormal.php:93
global $utfCompatibilityDecomp
Definition: UtfNormal.php:29
const UNICODE_HANGUL_LBASE
Definition: UtfNormal.php:35
const UTF8_HANGUL_VEND
Definition: UtfNormal.php:62
const UTF8_OVERLONG_A
Definition: UtfNormal.php:71
const UTF8_REPLACEMENT
Definition: UtfNormal.php:68
const UNICODE_HANGUL_TBASE
Definition: UtfNormal.php:37
const UTF8_OVERLONG_B
Definition: UtfNormal.php:72
const UNICODE_HANGUL_TCOUNT
Definition: UtfNormal.php:41
global $utfCanonicalDecomp
Definition: UtfNormal.php:23
const UNORM_NFKD
Definition: UtfNormal.php:90
const UTF8_FFFF
Definition: UtfNormal.php:79
const UNICODE_HANGUL_FIRST
Definition: UtfNormal.php:32
const UTF8_HANGUL_VBASE
Definition: UtfNormal.php:58
const UTF8_HANGUL_LAST
Definition: UtfNormal.php:55
const UTF8_HANGUL_LBASE
Definition: UtfNormal.php:57
const UTF8_FFFE
Definition: UtfNormal.php:78
global $utfCanonicalComp
Definition: UtfNormal.php:23
const UNORM_NFD
Definition: UtfNormal.php:89
const NORMALIZE_ICU
Definition: UtfNormal.php:96
const UNICODE_HANGUL_NCOUNT
Definition: UtfNormal.php:42
const UNICODE_HANGUL_VBASE
Definition: UtfNormal.php:36
const UTF8_HANGUL_FIRST
Definition: UtfNormal.php:54
global $l
Definition: afr.php:30
An exception for terminatinating execution or to throw for unit testing.
static quickIsNFC($string)
Returns true if the string is definitely in NFC.
Definition: UtfNormal.php:247
static fastCombiningSort($string)
Sorts combining characters into canonical order.
Definition: UtfNormal.php:638
static toNFD($string)
Convert a UTF-8 string to normal form D, canonical decomposition.
Definition: UtfNormal.php:176
static toNFC($string)
Convert a UTF-8 string to normal form C, canonical composition.
Definition: UtfNormal.php:157
static loadData()
Load the basic composition data if necessary.
Definition: UtfNormal.php:232
static fastCompose($string)
Produces canonically composed sequences, i.e.
Definition: UtfNormal.php:693
static fastDecompose($string, $map)
Perform decomposition of a UTF-8 string into either D or KD form (depending on which decomposition ma...
Definition: UtfNormal.php:576
static cleanUp($string)
The ultimate convenience function! Clean up invalid UTF-8 sequences, and convert to normal form C,...
Definition: UtfNormal.php:125
static NFKC($string)
Definition: UtfNormal.php:543
static toNFKC($string)
Convert a UTF-8 string to normal form KC, compatibility composition.
Definition: UtfNormal.php:196
static NFC($string)
Definition: UtfNormal.php:517
static placebo($string)
This is just used for the benchmark, comparing how long it takes to interate through a string without...
Definition: UtfNormal.php:829
static toNFKD($string)
Convert a UTF-8 string to normal form KD, compatibility decomposition.
Definition: UtfNormal.php:216
static quickIsNFCVerify(&$string)
Returns true if the string is definitely in NFC.
Definition: UtfNormal.php:291
static NFD($string)
Definition: UtfNormal.php:528
static NFKD($string)
Definition: UtfNormal.php:554
$i
Definition: disco.tpl.php:19
down()
Definition: down.php:2
$base
Definition: index.php:4
$index
Definition: metadata.php:60
instance(Loop $newLoop=null)
Retrieves or sets the global Loop object.
Definition: functions.php:173
$tail
Definition: tail.php:20
$start
Definition: bench.php:8
if($state['core:TerminatedAssocId'] !==null) $remaining