searchbigword-diff.patch
Public
Actions

Authored By

	• bzimport
	Nov 21 2014, 9:23 PM

Size

1 KB

Referenced Files

None

Subscribers

None

searchbigword-diff.patch
View Options

	Index: languages/Language.php
	===================================================================
	--- languages/Language.php (revision 42655)
	+++ languages/Language.php (working copy)
	@@ -1519,20 +1519,38 @@

	wfProfileIn( __METHOD__ );
	if( function_exists( 'mb_strtolower' ) ) {
	- $out = preg_replace(
	- "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
	- "'U8' . bin2hex( \"$1\" )",
	+ $out = preg_replace_callback(
	+ "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
	+ array( 'self', 'encodeUnicodeCharForSearch' ),
	mb_strtolower( $string ) );
	} else {
	list( , $wikiLowerChars ) = self::getCaseMaps();
	- $out = preg_replace(
	- "/([\\xc0-\\xff][\\x80-\\xbf]*)/e",
	- "'U8' . bin2hex( strtr( \"\$1\", \$wikiLowerChars ) )",
	+ $out = preg_replace_callback(
	+ "/([\\xc0-\\xff][\\x80-\\xbf]*)/",
	+ array( 'self', 'encodeMixedCaseUnicodeForSearch' ),
	$string );
	}
	+
	+ ## Pad small words so they get picked up in the index.
	+ $out = preg_replace( "/\b\w{1,4}\b/", 'SMALL$0', $out );
	+
	wfProfileOut( __METHOD__ );
	return $out;
	}
	+
	+ private function encodeMixedCaseUnicodeForSearch( $char ) {
	+ static $wikiLowerChars = null;
	+
	+ if (!$wikiLowerChars) {
	+ list( , $wikiLowerChars ) = self::getCaseMaps();
	+ }
	+
	+ return 'U8' . bin2hex( strtr( $char, $wikiLowerChars ) );
	+ }
	+
	+ private function encodeUnicodeCharForSearch( $char ) {
	+ return 'U8' . bin2hex( $char );
	+ }

	function convertForSearchResult( $termsArray ) {
	# some languages, e.g. Chinese, need to do a conversion