* (bug 1938) Fix normalization of character references in link text and

other replaced portions; run after link replacement but before tidy.
(Tidy will do some odd things if used, with breakage resulting from corner
cases.)

diff --git a/includes/Parser.php b/includes/Parser.php
index 2ff7786..daf10df 100644 (file)
--- a/includes/Parser.php
+++ b/includes/Parser.php
@@ -197,7 +197,6 @@ class Parser
                        '/<\\/center *>/i' => '</div>',
                );
                $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
-               $text = Sanitizer::normalizeCharReferences( $text );
                
                # only once and last
                $text = $this->doBlockLevels( $text, $linestart );
@@ -207,6 +206,7 @@ class Parser
                $this->mOutput->setTitleText($wgContLang->getParsedTitle());
                $text = $this->unstripNoWiki( $text, $this->mStripState );
                
+               $text = Sanitizer::normalizeCharReferences( $text );
                global $wgUseTidy;
                if ($wgUseTidy) {
                        $text = Parser::tidy($text);

diff --git a/maintenance/parserTests.txt b/maintenance/parserTests.txt
index 51203a5..e791a46 100644 (file)
--- a/maintenance/parserTests.txt
+++ b/maintenance/parserTests.txt
@@ -997,7 +997,7 @@ Interwiki link encoding conversion (bug 1636)
 *[[Wikipedia:ro:Olteniţa]]
 *[[Wikipedia:ro:Olteniţa]]
 !! result
-<ul><li><a href="http://en.wikipedia.org/wiki/ro:Olteni%C5%A3a" class='extiw' title="Wikipedia:ro:Olteniţa">Wikipedia:ro:Olteni&#0355;a</a>
+<ul><li><a href="http://en.wikipedia.org/wiki/ro:Olteni%C5%A3a" class='extiw' title="Wikipedia:ro:Olteniţa">Wikipedia:ro:Olteni&#355;a</a>
 </li><li><a href="http://en.wikipedia.org/wiki/ro:Olteni%C5%A3a" class='extiw' title="Wikipedia:ro:Olteniţa">Wikipedia:ro:Olteni&#355;a</a>
 </li></ul>
 
@@ -1136,7 +1136,7 @@ Magic links: PMID
 !! input
 PMID 1234
 !! result
-<p><a href='http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&amp;db=pubmed&amp;dopt=Abstract&amp;list_uids=1234' class='external' title="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=1234">PMID 1234</a>
+<p><a href='http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&amp;db=pubmed&amp;dopt=Abstract&amp;list_uids=1234' class='external' title="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&amp;db=pubmed&amp;dopt=Abstract&amp;list_uids=1234">PMID 1234</a>
 </p>
 !! end
 
@@ -2038,6 +2038,16 @@ Media link with nasty text
 </p>
 !! end
 
+
+!! test
+Character reference normalization in link text (bug 1938)
+!! input
+[[Main Page|this&that]]
+!! result
+<p><a href="/wiki/Main_Page" title="Main Page">this&amp;that</a>
+</p>
+!!end
+
 TODO:
 more images
 more tables