extract latinize() for search::utf2latin

1 files changed, 10 insertions, 3 deletions

diff --git a/it_html.class b/it_html.class
index 47444a6..e628a23 100644
--- a/it_html.class
+++ b/it_html.class
@@ -415,6 +415,13 @@ function sanitize($html)
 	return $result;
 }
 
+/**
+ * Replace or remove all illegal characters from a latin-1 string
+ */
+function latinize($string)
+{
+	return preg_replace('/[\x00-\x08\x0b-\x0c\x0e-\x1f\x80-\x9f]/', ' ', strtr($string, array("\x80" => "EUR", "\x82" => "'", "\x84" => "\"", "\x85" => "...", "\x8a" => "S", "\x8c" => "OE", "\x8e" => "Z", "\x91" => "'", "\x92" => "'", "\x93" => "\"", "\x94" => "\"", "\x96" => "-", "\x97" => "-", "\x9a" => "s", "\x9e" => "z")));
+}
 
 /**
  * Shortcut: return htmlspecialchars($string) and encode forbidden characters 80-9f if latin1 is output
@@ -423,10 +430,10 @@ function sanitize($html)
  */
 function Q($string)
 {
-	if ($GLOBALS['it_html']->p['charset'] == "iso-8859-1")
-		$string = preg_replace('/[\x00-\x08\x0b-\x0c\x0e-\x1f\x80-\x9f]/', ' ', strtr($string, array("\x80" => "EUR", "\x82" => "'", "\x84" => "\"", "\x85" => "...", "\x8a" => "S", "\x8c" => "OE", "\x8e" => "Z", "\x91" => "'", "\x92" => "'", "\x93" => "\"", "\x94" => "\"", "\x96" => "-", "\x97" => "-", "\x9a" => "s", "\x9e" => "z")));
+	if (preg_match('/[<>&"\x00-\x08\x0b-\x0c\x0e-\x1f\x80-\x9f]/', $string))
+		$string = htmlspecialchars($GLOBALS['it_html']->p['charset'] == "iso-8859-1" ? it_html::latinize($string) : $string);
 
-	return htmlspecialchars($string);
+	return $string;
 }