* This implementation is slower than hungryDelimiterReplace but uses far less
* memory. The delimiters are literal strings, not regular expressions.
*
- * @param string $flags Regular expression flags
+ * @param $startDelim String: start delimiter
+ * @param $endDelim String: end delimiter
+ * @param $callback Callback: function to call on each match
+ * @param $subject String
+ * @param $flags String: regular expression flags
*/
# If the start delimiter ends with an initial substring of the end delimiter,
# e.g. in the case of C-style comments, the behaviour differs from the model
}
if ( $tokenType == 'start' ) {
- $inputPos = $tokenOffset + $tokenLength;
# Only move the start position if we haven't already found a start
# This means that START START END matches outer pair
if ( !$foundStart ) {
# Found start
+ $inputPos = $tokenOffset + $tokenLength;
# Write out the non-matching section
$output .= substr( $subject, $outputPos, $tokenOffset - $outputPos );
$outputPos = $tokenOffset;
$contentPos = $inputPos;
$foundStart = true;
+ } else {
+ # Move the input position past the *first character* of START,
+ # to protect against missing END when it overlaps with START
+ $inputPos = $tokenOffset + 1;
}
} elseif ( $tokenType == 'end' ) {
if ( $foundStart ) {
return $output;
}
- /*
+ /**
* Perform an operation equivalent to
*
* preg_replace( "!$startDelim(.*)$endDelim!$flags", $replace, $subject )
*
- * @param string $startDelim Start delimiter regular expression
- * @param string $endDelim End delimiter regular expression
- * @param string $replace Replacement string. May contain $1, which will be
- * replaced by the text between the delimiters
- * @param string $subject String to search
- * @return string The string with the matches replaced
+ * @param $startDelim String: start delimiter regular expression
+ * @param $endDelim String: end delimiter regular expression
+ * @param $replace String: replacement string. May contain $1, which will be
+ * replaced by the text between the delimiters
+ * @param $subject String to search
+ * @param $flags String: regular expression flags
+ * @return String: The string with the matches replaced
*/
static function delimiterReplace( $startDelim, $endDelim, $replace, $subject, $flags = '' ) {
$replacer = new RegexlikeReplacer( $replace );
/**
* More or less "markup-safe" explode()
* Ignores any instances of the separator inside <...>
- * @param string $separator
- * @param string $text
+ * @param $separator String
+ * @param $text String
* @return array
*/
static function explodeMarkup( $separator, $text ) {
* Escape a string to make it suitable for inclusion in a preg_replace()
* replacement parameter.
*
- * @param string $string
- * @return string
+ * @param $string String
+ * @return String
*/
static function escapeRegexReplacement( $string ) {
$string = str_replace( '\\', '\\\\', $string );
return new ArrayIterator( explode( $separator, $subject ) );
}
}
-
- /**
- * Clean characters that are invalid in the given character set
- * from a given string.
- *
- * @param $string \type{$string} String to clean
- * @param $charset \type{$string} Character set (if unspecified, assume $wgOutputEncoding)
- * @return \type{$string} Cleaned string
- */
- public static function cleanForCharset( $string, $charset='' ) {
- global $wgOutputEncoding;
- switch ( $charset ? $charset : $wgOutputEncoding ) {
- # UTF-8 should be all we need to worry about. :)
- case 'UTF-8':
- return self::cleanUtf8( $string );
- default:
- return $string;
- }
- }
-
- /**
- * Clean invalid UTF-8 characters and sequences from a given string,
- * replacing them with U+FFFD.
- * Should be RFC 3629 compliant.
- *
- * @param $string \type{$string} String to clean
- * @return \type{$string} Cleaned string
- */
- private static function cleanUtf8( $str ) {
- # HERE BE DRAGONS!
- # ABANDON ALL HOPE, ALL YE WHO ENTER THE BITWISE HELLFIRE.
-
- $illegal = array( 0xD800, 0xDB7F, 0xDB80, 0xDBFF,
- 0xDC00, 0xDF80, 0xDFFF, 0xFFFE, 0xFFFF );
- $len = strlen( $str );
- $left = $bytes = 0;
- for ( $i = 0; $i < $len; $i++ ) {
- $ch = ord( $str[$i] );
- if ( !$left ) {
- if ( !($ch & 0x80 ) )
- continue;
- $left = (( $ch & 0xFE ) == 0xFC ? 5 :
- (( $ch & 0xFC ) == 0xF8 ? 4 :
- (( $ch & 0xF8 ) == 0xF0 ? 3 :
- (( $ch & 0xF0 ) == 0xE0 ? 2 :
- (( $ch & 0xE0 ) == 0xC0 ? 1 :
- 0 )))));
- if ( $left ) {
- $bytes = $left + 1;
- $sum = $ch & ( 0xFF >> $bytes + 1 );
- continue;
- } else if ( $ch & 0x80 ) {
- $bytes = 1;
- }
- } else if ( ( $ch & 0xC0 ) == 0x80 ) {
- $sum <<= 6;
- $sum += $ch & 0x3F;
- if ( --$left ) continue;
- if ( ( $bytes == 2 && $sum < 0x80 ) ||
- ( $bytes == 3 && $sum < 0x800 ) ||
- ( $bytes == 4 && $sum < 0x10000 ) ||
- ( $bytes > 4 || $sum > 0x10FFFF ) ||
- in_array( $sum, $illegal ) ) {
- } else continue;
-
- } else {
- $bytes -= $left;
- $i--;
- }
-
- $str = ( substr( $str, 0, $i - $bytes + 1 ) .
- "\xEF\xBF\xBD" .
- substr( $str, $i + 1 ) );
- $i += 3 - $bytes;
- $len += 3 - $bytes;
- $left = 0;
- }
-
- return $str;
- }
}
/**