From b5ad1e694b3b233bb2dcd90b0aeafcb7dc18a5c6 Mon Sep 17 00:00:00 2001
From: Brion Vibber <brion@users.mediawiki.org>
Date: Mon, 7 Feb 2005 03:56:22 +0000
Subject: [PATCH] * Fix link in image alt text, using replaceLinkHolders with
 the global parser. This is a hack; such things should not be globally linked.
 * Use the common attribute normalization code in the alt text
 de-tagification.

---
 includes/Linker.php    | 10 +++++++---
 includes/Sanitizer.php | 25 +++++++++++++++++++++++++
 2 files changed, 32 insertions(+), 3 deletions(-)

diff --git a/includes/Linker.php b/includes/Linker.php
index 25fad56d8e..ee36c912da 100644
--- a/includes/Linker.php
+++ b/includes/Linker.php
@@ -519,9 +519,13 @@ class Linker {
 			if ( '' == $manual_thumb ) $url = $img->createThumb( $width );
 		}
 
-		$alt = preg_replace( '/<[^>]*>/', '', $alt );
-		$alt = preg_replace('/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/', '&amp;', $alt);
-		$alt = str_replace( array('<', '>', '"'), array('&lt;', '&gt;', '&quot;'), $alt );
+		# FIXME: This is a gross hack using a global.
+		# Replace link color holders in the caption text so the
+		# text portion can be placed int the alt/title attributes.
+		global $wgParser;
+		$wgParser->replaceLinkHolders( $alt );
+		
+		$alt = Sanitizer::stripAllTags( $alt );
 
 		$u = $nt->escapeLocalURL();
 		if ( $url == '' ) {
diff --git a/includes/Sanitizer.php b/includes/Sanitizer.php
index 2729efcb7c..fec5d720be 100644
--- a/includes/Sanitizer.php
+++ b/includes/Sanitizer.php
@@ -795,6 +795,31 @@ class Sanitizer {
 			);
 		return $whitelist;
 	}
+	
+	/**
+	 * Take a fragment of (potentially invalid) HTML and return
+	 * a version with any tags removed, encoded suitably for literal
+	 * inclusion in an attribute value.
+	 *
+	 * @param string $text HTML fragment
+	 * @return string
+	 */
+	function stripAllTags( $text ) {
+		# Actual <tags>
+		$text = preg_replace( '/<[^>]*>/', '', $text );
+		
+		# Normalize &entities and whitespace
+		$text = Sanitizer::normalizeAttributeValue( $text );
+		
+		# Will be placed into "double-quoted" attributes,
+		# make sure remaining bits are safe.
+		$text = str_replace(
+			array('<', '>', '"'),
+			array('&lt;', '&gt;', '&quot;'),
+			$text );
+		
+		return $text;
+	}
 
 }
 
-- 
2.20.1