1<?php 2 3use dokuwiki\Utf8\Unicode; 4/** 5 * Class to safely store UTF-8 in a Filename 6 * 7 * Encodes a utf8 string using only the following characters 0-9a-z_.-% 8 * characters 0-9a-z in the original string are preserved, "plain". 9 * all other characters are represented in a substring that starts 10 * with '%' are "converted". 11 * The transition from converted substrings to plain characters is 12 * marked with a '.' 13 * 14 * @author Christopher Smith <chris@jalakai.co.uk> 15 * @date 2010-04-02 16 */ 17class SafeFN { 18 19 // 'safe' characters are a superset of $plain, $pre_indicator and $post_indicator 20 private static $plain = '-./[_0123456789abcdefghijklmnopqrstuvwxyz'; // these characters aren't converted 21 private static $pre_indicator = '%'; 22 private static $post_indicator = ']'; 23 24 /** 25 * Convert an UTF-8 string to a safe ASCII String 26 * 27 * conversion process 28 * - if codepoint is a plain or post_indicator character, 29 * - if previous character was "converted", append post_indicator to output, clear "converted" flag 30 * - append ascii byte for character to output 31 * (continue to next character) 32 * 33 * - if codepoint is a pre_indicator character, 34 * - append ascii byte for character to output, set "converted" flag 35 * (continue to next character) 36 * 37 * (all remaining characters) 38 * - reduce codepoint value for non-printable ASCII characters (0x00 - 0x1f). Space becomes our zero. 39 * - convert reduced value to base36 (0-9a-z) 40 * - append $pre_indicator characater followed by base36 string to output, set converted flag 41 * (continue to next character) 42 * 43 * @param string $filename a utf8 string, should only include printable characters - not 0x00-0x1f 44 * @return string an encoded representation of $filename using only 'safe' ASCII characters 45 * 46 * @author Christopher Smith <chris@jalakai.co.uk> 47 */ 48 public static function encode($filename) { 49 return self::unicodeToSafe(Unicode::fromUtf8($filename)); 50 } 51 52 /** 53 * decoding process 54 * - split the string into substrings at any occurrence of pre or post indicator characters 55 * - check the first character of the substring 56 * - if its not a pre_indicator character 57 * - if previous character was converted, skip over post_indicator character 58 * - copy codepoint values of remaining characters to the output array 59 * - clear any converted flag 60 * (continue to next substring) 61 * 62 * _ else (its a pre_indicator character) 63 * - if string length is 1, copy the post_indicator character to the output array 64 * (continue to next substring) 65 * 66 * - else (string length > 1) 67 * - skip the pre-indicator character and convert remaining string from base36 to base10 68 * - increase codepoint value for non-printable ASCII characters (add 0x20) 69 * - append codepoint to output array 70 * (continue to next substring) 71 * 72 * @param string $filename a 'safe' encoded ASCII string, 73 * @return string decoded utf8 representation of $filename 74 * 75 * @author Christopher Smith <chris@jalakai.co.uk> 76 */ 77 public static function decode($filename) { 78 return Unicode::toUtf8(self::safeToUnicode(strtolower($filename))); 79 } 80 81 public static function validatePrintableUtf8($printable_utf8) { 82 return !preg_match('#[\x01-\x1f]#',$printable_utf8); 83 } 84 85 public static function validateSafe($safe) { 86 return !preg_match('#[^'.self::$plain.self::$post_indicator.self::$pre_indicator.']#',$safe); 87 } 88 89 /** 90 * convert an array of unicode codepoints into 'safe_filename' format 91 * 92 * @param array int $unicode an array of unicode codepoints 93 * @return string the unicode represented in 'safe_filename' format 94 * 95 * @author Christopher Smith <chris@jalakai.co.uk> 96 */ 97 private static function unicodeToSafe($unicode) { 98 99 $safe = ''; 100 $converted = false; 101 102 foreach ($unicode as $codepoint) { 103 if ($codepoint < 127 && (strpos(self::$plain.self::$post_indicator,chr($codepoint))!==false)) { 104 if ($converted) { 105 $safe .= self::$post_indicator; 106 $converted = false; 107 } 108 $safe .= chr($codepoint); 109 } elseif ($codepoint == ord(self::$pre_indicator)) { 110 $safe .= self::$pre_indicator; 111 $converted = true; 112 } else { 113 $safe .= self::$pre_indicator.base_convert((string)($codepoint-32),10,36); 114 $converted = true; 115 } 116 } 117 if($converted) $safe .= self::$post_indicator; 118 return $safe; 119 } 120 121 /** 122 * convert a 'safe_filename' string into an array of unicode codepoints 123 * 124 * @param string $safe a filename in 'safe_filename' format 125 * @return array int an array of unicode codepoints 126 * 127 * @author Christopher Smith <chris@jalakai.co.uk> 128 */ 129 private static function safeToUnicode($safe) { 130 131 $unicode = []; 132 $split = preg_split('#(?=['.self::$post_indicator.self::$pre_indicator.'])#',$safe,-1,PREG_SPLIT_NO_EMPTY); 133 134 $converted = false; 135 foreach ($split as $sub) { 136 $len = strlen($sub); 137 if ($sub[0] != self::$pre_indicator) { 138 // plain (unconverted) characters, optionally starting with a post_indicator 139 // set initial value to skip any post_indicator 140 for ($i=($converted?1:0); $i < $len; $i++) { 141 $unicode[] = ord($sub[$i]); 142 } 143 $converted = false; 144 } elseif ($len==1) { 145 // a pre_indicator character in the real data 146 $unicode[] = ord($sub); 147 $converted = true; 148 } else { 149 // a single codepoint in base36, adjusted for initial 32 non-printable chars 150 $unicode[] = 32 + (int)base_convert(substr($sub,1),36,10); 151 $converted = true; 152 } 153 } 154 155 return $unicode; 156 } 157 158} 159