From 23da83490a17dd72d75166c5307e760a7b8d856c Mon Sep 17 00:00:00 2001
From: Derk-Jan Hartman <hartman@users.mediawiki.org>
Date: Tue, 15 Jun 2010 13:47:38 +0000
Subject: [PATCH] Fix bug 16356. Use proper normalization for dumpInterwiki.inc
 It was previously breaking on multibyte characters and space characters in
 interwiki entries.

Patch by User:Umherirrender, Review by ^demon
---
 CREDITS                       | 1 +
 RELEASE-NOTES                 | 1 +
 maintenance/dumpInterwiki.inc | 4 +++-
 3 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/CREDITS b/CREDITS
index 7b24692378..10a77d417b 100644
--- a/CREDITS
+++ b/CREDITS
@@ -127,6 +127,7 @@ following names for their contribution to the product.
 * Str4nd
 * svip
 * Tisane
+* Umherirrender
 * Zachary Hauri
 
 == Translators ==
diff --git a/RELEASE-NOTES b/RELEASE-NOTES
index 35520832ff..8224914609 100644
--- a/RELEASE-NOTES
+++ b/RELEASE-NOTES
@@ -196,6 +196,7 @@ it from source control: http://www.mediawiki.org/wiki/Download_from_SVN
   accidently activated when hitting enter.
 * (bug 23845) Special:ListFiles now uses correct file names without underscores
 * Ask for permanent login in Special:Preferences only if $wgCookieExpiration > 0
+* (bug 16356) Repair dumpInterwiki.inc to use proper normalization.
 
 === API changes in 1.17 ===
 * (bug 22738) Allow filtering by action type on query=logevent.
diff --git a/maintenance/dumpInterwiki.inc b/maintenance/dumpInterwiki.inc
index 1f8bb55e06..97e5be99dd 100644
--- a/maintenance/dumpInterwiki.inc
+++ b/maintenance/dumpInterwiki.inc
@@ -29,7 +29,7 @@ class Site {
 }
 
 function getRebuildInterwikiDump() {
-	global $langlist, $languageAliases, $prefixRewrites;
+	global $langlist, $languageAliases, $prefixRewrites, $wgContLang;
 
 	# Multi-language sites
 	# db suffix => db suffix, iw prefix, hostname
@@ -102,6 +102,8 @@ function getRebuildInterwikiDump() {
 	# Global iterwiki map
 	foreach ( $lines as $line ) {
 		if ( preg_match( '/^\|\s*(.*?)\s*\|\|\s*(.*?)\s*$/', $line, $matches ) ) {
+			$prefix = $wgContLang->lc( $matches[1] );
+			$prefix = str_replace( ' ', '_', $prefix );
 			$prefix = strtolower( $matches[1] );
 			$url = $matches[2];
 			if ( preg_match( '/(wikipedia|wiktionary|wikisource|wikiquote|wikibooks|wikimedia)\.org/', $url ) ) {
-- 
2.20.1