diff options
Diffstat (limited to 'include/utf8/utils/specials.php')
| -rw-r--r-- | include/utf8/utils/specials.php | 131 |
1 files changed, 131 insertions, 0 deletions
diff --git a/include/utf8/utils/specials.php b/include/utf8/utils/specials.php new file mode 100644 index 0000000..69219dc --- /dev/null +++ b/include/utf8/utils/specials.php @@ -0,0 +1,131 @@ +<?php + +/** +* Utilities for processing "special" characters in UTF-8. "Special" largely means anything which would +* be regarded as a non-word character, like ASCII control characters and punctuation. This has a "Roman" +* bias - it would be unaware of modern Chinese "punctuation" characters for example. +* Note: requires utils/unicode.php to be loaded +* @version $Id: specials.php,v 1.2 2006/10/16 21:13:59 harryf Exp $ +* @package utf8 +* @subpackage utils +* @see utf8_is_valid +*/ + +/** +* Used internally. Builds a PCRE pattern from the $UTF8_SPECIAL_CHARS +* array defined in this file +* The $UTF8_SPECIAL_CHARS should contain all special characters (non-letter/non-digit) +* defined in the various local charsets - it's not a complete list of +* non-alphanum characters in UTF-8. It's not perfect but should match most +* cases of special chars. +* This function adds the control chars 0x00 to 0x19 to the array of +* special chars (they are not included in $UTF8_SPECIAL_CHARS) +* @package utf8 +* @subpackage utils +* @return string +* @see utf8_from_unicode +* @see utf8_is_word_chars +* @see utf8_strip_specials +*/ +function utf8_specials_pattern() +{ + static $pattern = null; + + if (!$pattern) + { + $UTF8_SPECIAL_CHARS = array( + 0x001a, 0x001b, 0x001c, 0x001d, 0x001e, 0x001f, 0x0020, 0x0021, 0x0022, 0x0023, + 0x0024, 0x0025, 0x0026, 0x0027, 0x0028, 0x0029, 0x002a, 0x002b, 0x002c, + 0x002f, 0x003b, 0x003c, 0x003d, 0x003e, 0x003f, 0x0040, 0x005b, + 0x005c, 0x005d, 0x005e, 0x0060, 0x007b, 0x007c, 0x007d, 0x007e, + 0x007f, 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, 0x0088, + 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, 0x0090, 0x0091, 0x0092, + 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, + 0x009d, 0x009e, 0x009f, 0x00a0, 0x00a1, 0x00a2, 0x00a3, 0x00a4, 0x00a5, 0x00a6, + 0x00a7, 0x00a8, 0x00a9, 0x00aa, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af, 0x00b0, + 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7, 0x00b8, 0x00b9, 0x00ba, + 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00bf, 0x00d7, 0x00f7, 0x02c7, 0x02d8, 0x02d9, + 0x02da, 0x02db, 0x02dc, 0x02dd, 0x0300, 0x0301, 0x0303, 0x0309, 0x0323, 0x0384, + 0x0385, 0x0387, 0x03b2, 0x03c6, 0x03d1, 0x03d2, 0x03d5, 0x03d6, 0x05b0, 0x05b1, + 0x05b2, 0x05b3, 0x05b4, 0x05b5, 0x05b6, 0x05b7, 0x05b8, 0x05b9, 0x05bb, 0x05bc, + 0x05bd, 0x05be, 0x05bf, 0x05c0, 0x05c1, 0x05c2, 0x05c3, 0x05f3, 0x05f4, 0x060c, + 0x061b, 0x061f, 0x0640, 0x064b, 0x064c, 0x064d, 0x064e, 0x064f, 0x0650, 0x0651, + 0x0652, 0x066a, 0x0e3f, 0x200c, 0x200d, 0x200e, 0x200f, 0x2013, 0x2014, 0x2015, + 0x2017, 0x2018, 0x2019, 0x201a, 0x201c, 0x201d, 0x201e, 0x2020, 0x2021, 0x2022, + 0x2026, 0x2030, 0x2032, 0x2033, 0x2039, 0x203a, 0x2044, 0x20a7, 0x20aa, 0x20ab, + 0x20ac, 0x2116, 0x2118, 0x2122, 0x2126, 0x2135, 0x2190, 0x2191, 0x2192, 0x2193, + 0x2194, 0x2195, 0x21b5, 0x21d0, 0x21d1, 0x21d2, 0x21d3, 0x21d4, 0x2200, 0x2202, + 0x2203, 0x2205, 0x2206, 0x2207, 0x2208, 0x2209, 0x220b, 0x220f, 0x2211, 0x2212, + 0x2215, 0x2217, 0x2219, 0x221a, 0x221d, 0x221e, 0x2220, 0x2227, 0x2228, 0x2229, + 0x222a, 0x222b, 0x2234, 0x223c, 0x2245, 0x2248, 0x2260, 0x2261, 0x2264, 0x2265, + 0x2282, 0x2283, 0x2284, 0x2286, 0x2287, 0x2295, 0x2297, 0x22a5, 0x22c5, 0x2310, + 0x2320, 0x2321, 0x2329, 0x232a, 0x2469, 0x2500, 0x2502, 0x250c, 0x2510, 0x2514, + 0x2518, 0x251c, 0x2524, 0x252c, 0x2534, 0x253c, 0x2550, 0x2551, 0x2552, 0x2553, + 0x2554, 0x2555, 0x2556, 0x2557, 0x2558, 0x2559, 0x255a, 0x255b, 0x255c, 0x255d, + 0x255e, 0x255f, 0x2560, 0x2561, 0x2562, 0x2563, 0x2564, 0x2565, 0x2566, 0x2567, + 0x2568, 0x2569, 0x256a, 0x256b, 0x256c, 0x2580, 0x2584, 0x2588, 0x258c, 0x2590, + 0x2591, 0x2592, 0x2593, 0x25a0, 0x25b2, 0x25bc, 0x25c6, 0x25ca, 0x25cf, 0x25d7, + 0x2605, 0x260e, 0x261b, 0x261e, 0x2660, 0x2663, 0x2665, 0x2666, 0x2701, 0x2702, + 0x2703, 0x2704, 0x2706, 0x2707, 0x2708, 0x2709, 0x270c, 0x270d, 0x270e, 0x270f, + 0x2710, 0x2711, 0x2712, 0x2713, 0x2714, 0x2715, 0x2716, 0x2717, 0x2718, 0x2719, + 0x271a, 0x271b, 0x271c, 0x271d, 0x271e, 0x271f, 0x2720, 0x2721, 0x2722, 0x2723, + 0x2724, 0x2725, 0x2726, 0x2727, 0x2729, 0x272a, 0x272b, 0x272c, 0x272d, 0x272e, + 0x272f, 0x2730, 0x2731, 0x2732, 0x2733, 0x2734, 0x2735, 0x2736, 0x2737, 0x2738, + 0x2739, 0x273a, 0x273b, 0x273c, 0x273d, 0x273e, 0x273f, 0x2740, 0x2741, 0x2742, + 0x2743, 0x2744, 0x2745, 0x2746, 0x2747, 0x2748, 0x2749, 0x274a, 0x274b, 0x274d, + 0x274f, 0x2750, 0x2751, 0x2752, 0x2756, 0x2758, 0x2759, 0x275a, 0x275b, 0x275c, + 0x275d, 0x275e, 0x2761, 0x2762, 0x2763, 0x2764, 0x2765, 0x2766, 0x2767, 0x277f, + 0x2789, 0x2793, 0x2794, 0x2798, 0x2799, 0x279a, 0x279b, 0x279c, 0x279d, 0x279e, + 0x279f, 0x27a0, 0x27a1, 0x27a2, 0x27a3, 0x27a4, 0x27a5, 0x27a6, 0x27a7, 0x27a8, + 0x27a9, 0x27aa, 0x27ab, 0x27ac, 0x27ad, 0x27ae, 0x27af, 0x27b1, 0x27b2, 0x27b3, + 0x27b4, 0x27b5, 0x27b6, 0x27b7, 0x27b8, 0x27b9, 0x27ba, 0x27bb, 0x27bc, 0x27bd, + 0x27be, 0xf6d9, 0xf6da, 0xf6db, 0xf8d7, 0xf8d8, 0xf8d9, 0xf8da, 0xf8db, 0xf8dc, + 0xf8dd, 0xf8de, 0xf8df, 0xf8e0, 0xf8e1, 0xf8e2, 0xf8e3, 0xf8e4, 0xf8e5, 0xf8e6, + 0xf8e7, 0xf8e8, 0xf8e9, 0xf8ea, 0xf8eb, 0xf8ec, 0xf8ed, 0xf8ee, 0xf8ef, 0xf8f0, + 0xf8f1, 0xf8f2, 0xf8f3, 0xf8f4, 0xf8f5, 0xf8f6, 0xf8f7, 0xf8f8, 0xf8f9, 0xf8fa, + 0xf8fb, 0xf8fc, 0xf8fd, 0xf8fe, 0xfe7c, 0xfe7d); + + $pattern = preg_quote(utf8_from_unicode($UTF8_SPECIAL_CHARS), '/'); + $pattern = '/[\x00-\x19'.$pattern.']/u'; + } + + return $pattern; +} + +/** +* Checks a string for whether it contains only word characters. This +* is logically equivalent to the \w PCRE meta character. Note that +* this is not a 100% guarantee that the string only contains alpha / +* numeric characters but just that common non-alphanumeric are not +* in the string, including ASCII device control characters. +* @package utf8 +* @subpackage utils +* @param string to check +* @return boolean TRUE if the string only contains word characters +* @see utf8_specials_pattern +*/ +function utf8_is_word_chars($str) +{ + return !(bool) preg_match(utf8_specials_pattern(), $str); +} + +/** +* Removes special characters (nonalphanumeric) from a UTF-8 string +* +* This can be useful as a helper for sanitizing a string for use as +* something like a file name or a unique identifier. Be warned though +* it does not handle all possible non-alphanumeric characters and is +* not intended is some kind of security / injection filter. +* +* @package utf8 +* @subpackage utils +* @author Andreas Gohr <andi@splitbrain.org> +* @param string $string The UTF8 string to strip of special chars +* @param string (optional) $repl Replace special with this string +* @return string with common non-alphanumeric characters removed +* @see utf8_specials_pattern +*/ +function utf8_strip_specials($string, $repl='') +{ + return preg_replace(utf8_specials_pattern(), $repl, $string); +} |