From eb41544de5fa31c2ed70cadaaa9204e4ca23d8c5 Mon Sep 17 00:00:00 2001
From: Conrad Irwin <conrad@users.mediawiki.org>
Date: Mon, 21 Jun 2010 01:17:36 +0000
Subject: [PATCH] (bug 22784) Fix normalization of whitespace in autocomment
 links

put this specific normalisation into Sanitizer::normalizeSectionNameWhitespace
instead of repeating it three times.
---
 RELEASE-NOTES               |  1 +
 includes/Linker.php         | 10 +++-------
 includes/Sanitizer.php      | 12 ++++++++++++
 includes/parser/Parser.php  |  5 ++---
 maintenance/parserTests.txt | 11 +++++++++++
 5 files changed, 29 insertions(+), 10 deletions(-)

diff --git a/RELEASE-NOTES b/RELEASE-NOTES
index 1310543a54..b8de16db5c 100644
--- a/RELEASE-NOTES
+++ b/RELEASE-NOTES
@@ -208,6 +208,7 @@ it from source control: http://www.mediawiki.org/wiki/Download_from_SVN
   Special:Upload after following a red link
 * Correct the escaping of the autosummary URI fragments.
 * (bug 23642) Recognize mime types of MS OpenXML documents.
+* (bug 22784) Normalise underscores and spaces in autocomments.
 
 === API changes in 1.17 ===
 * (bug 22738) Allow filtering by action type on query=logevent.
diff --git a/includes/Linker.php b/includes/Linker.php
index f6e45d1408..c7f7df9df6 100644
--- a/includes/Linker.php
+++ b/includes/Linker.php
@@ -1018,17 +1018,13 @@ class Linker {
 		if ( $title ) {
 			$section = $auto;
 
-			# Generate a valid anchor name from the section title.
-			# Hackish, but should generally work - we strip wiki
-			# syntax, including the magic [[: that is used to
-			# "link rather than show" in case of images and
-			# interlanguage links.
+			# Remove links that a user may have manually put in the autosummary
+			# This could be improved by copying as much of Parser::stripSectionName as desired.
 			$section = str_replace( '[[:', '', $section );
 			$section = str_replace( '[[', '', $section );
 			$section = str_replace( ']]', '', $section );
 
-			# Most of Title:: expects fragments to be escaped
-			$section = Title::escapeFragmentForURL( $section );
+			$section = Sanitizer::normalizeSectionNameWhitespace( $section ); # bug 22784
 			if ( $local ) {
 				$sectionTitle = Title::newFromText( '#' . $section );
 			} else {
diff --git a/includes/Sanitizer.php b/includes/Sanitizer.php
index 00837fdadf..2a3f8473cf 100644
--- a/includes/Sanitizer.php
+++ b/includes/Sanitizer.php
@@ -1166,6 +1166,18 @@ class Sanitizer {
 			$text );
 	}
 
+	/**
+	 * Normalizes whitespace in a section name, such as might be returned
+	 * by Parser::stripSectionName(), for use in the id's that are used for
+	 * section links.
+	 *
+	 * @param $section String
+	 * @return String
+	 */
+	static function normalizeSectionNameWhitespace( $section ) {
+		return trim( preg_replace( '/[ _]+/', ' ', $section ) );
+	}
+
 	/**
 	 * Ensure that any entities and character references are legal
 	 * for XML and XHTML specifically. Any stray bits will be
diff --git a/includes/parser/Parser.php b/includes/parser/Parser.php
index 7b28c14519..2a681980ce 100644
--- a/includes/parser/Parser.php
+++ b/includes/parser/Parser.php
@@ -3859,8 +3859,7 @@ class Parser {
 
 			# For the anchor, strip out HTML-y stuff period
 			$safeHeadline = preg_replace( '/<.*?'.'>/', '', $safeHeadline );
-			$safeHeadline = preg_replace( '/[ _]+/', ' ', $safeHeadline );
-			$safeHeadline = trim( $safeHeadline );
+			$safeHeadline = Sanitizer::normalizeSectionNameWhitespace( $safeHeadline );
 
 			# Save headline for section edit hint before it's escaped
 			$headlineHint = $safeHeadline;
@@ -5172,7 +5171,7 @@ class Parser {
 	public function guessSectionNameFromWikiText( $text ) {
 		# Strip out wikitext links(they break the anchor)
 		$text = $this->stripSectionName( $text );
-		$text = trim( preg_replace( '/[ _]+/', ' ', $text ) );
+		$text = Sanitizer::normalizeSectionNameWhitespace( $text );
 		return '#' . Sanitizer::escapeId( $text, 'noninitial' );
 	}
 
diff --git a/maintenance/parserTests.txt b/maintenance/parserTests.txt
index e3612b179f..89a9587f94 100644
--- a/maintenance/parserTests.txt
+++ b/maintenance/parserTests.txt
@@ -7816,6 +7816,17 @@ title=[[Main Page]]
 <a href="/wiki/Main_Page#section" title="Main Page">#section</a>
 !! end
 
+!! test
+Space normalisation on autocomment (bug 22784)
+!! options
+comment
+title=[[Main Page]]
+!!input
+/* __hello__world__ */
+!! result
+<span class="autocomment"><a href="/wiki/Main_Page#hello_world" title="Main Page">→</a>__hello__world__</span>
+!! end
+
 !!article
 MediaWiki:bad image list
 !!text
-- 
2.20.1