Support external links in MWInternalLinkAnnotation

As URLs from the clipboard are always absolute, we need to detect if these
are from the same wiki as the current document, and if so convert back to
relative for Parsoid.

Bug: 58136
Change-Id: Id251afe65193fc6356628f1deb5ed757f8a6d347

modules/ve-mw/dm/annotations/ve.dm.MWInternalLinkAnnotation.js

@@ -35,13 +35,33 @@ ve.dm.MWInternalLinkAnnotation.static.name = 'link/mwInternal';
 ve.dm.MWInternalLinkAnnotation.static.matchRdfaTypes = ['mw:WikiLink'];
 
 ve.dm.MWInternalLinkAnnotation.static.toDataElement = function ( domElements ) {
-	// Get title from href
+
+	function regexEscape( str ) {
+		return str.replace( /([.?*+^$[\]\\(){}|-])/g, '\\$1' );
+	}
+
+	var matches, normalizedTitle,
+		// Protocol relative base
+		relativeBase = new mw.Uri(  mw.config.get( 'wgArticlePath' ) ).toString().replace( /^https?:/, '' ),
+		relativeBaseRegex = new RegExp( regexEscape( relativeBase ).replace( regexEscape( '$1' ), '(.*)' ) ),
+		href = domElements[0].getAttribute( 'href' ),
+		// Protocol relative href
+		relativeHref = href.replace( /^https?:/, '' );
+
+	// Check if this matches the server's article path
+	matches = relativeHref.match ( relativeBaseRegex );
+	if ( matches ) {
+		// Take the relative path
+		href = matches[1];
+	}
+
 	// The href is simply the title, unless we're dealing with a page that has slashes in its name
 	// in which case it's preceded by one or more instances of "./" or "../", so strip those
 	/*jshint regexp:false */
-	var matches = domElements[0].getAttribute( 'href' ).match( /^((?:\.\.?\/)*)(.*)$/ ),
-		// Normalize capitalisation and underscores
-		normalizedTitle = ve.dm.MWInternalLinkAnnotation.static.normalizeTitle( matches[2] );
+	matches = href.match( /^((?:\.\.?\/)*)(.*)$/ );
+	// Normalize capitalisation and underscores
+	normalizedTitle = ve.dm.MWInternalLinkAnnotation.static.normalizeTitle( matches[2] );
+
 	return {
 		'type': 'link/mwInternal',
 		'attributes': {

modules/ve-mw/test/dm/ve.dm.mwExample.js

@@ -5,6 +5,8 @@
  * @license The MIT License (MIT); see LICENSE.txt
  */
 
+/*global mw */
+
 /**
  * @class
  * @singleton
@@ -155,6 +157,32 @@ ve.dm.mwExample.MWTransclusion.mixedStoreItems = {
 	'value': $( ve.dm.mwExample.MWTransclusion.mixed ).toArray()
 };
 
+ve.dm.mwExample.MWInternalLink = {
+	'absoluteHref': new mw.Uri( '/wiki/Foo/Bar' ).toString()
+};
+
+ve.dm.mwExample.MWInternalLink.absoluteOpen = '<a rel="mw:WikiLink" href="' + ve.dm.mwExample.MWInternalLink.absoluteHref + '">';
+ve.dm.mwExample.MWInternalLink.absoluteData = {
+	'type': 'link/mwInternal',
+	'attributes': {
+		'title': 'Foo/Bar',
+		'origTitle': 'Foo/Bar',
+		'normalizedTitle': 'Foo/Bar',
+		'hrefPrefix': ''
+	},
+	'htmlAttributes': [
+		{
+			'values': {
+				'href': ve.dm.mwExample.MWInternalLink.absoluteHref,
+				'rel': 'mw:WikiLink'
+			},
+			'computed': {
+				'href': ve.dm.mwExample.MWInternalLink.absoluteHref
+			}
+		}
+	]
+};
+
 ve.dm.mwExample.MWBlockImage = {
 	'html':
 		'<figure typeof="mw:Image/Thumb" class="mw-halign-right foobar">' +
@@ -1501,6 +1529,31 @@ ve.dm.mwExample.domToDataCases = {
 			{ 'type': '/internalList' }
 		]
 	},
+	'internal link with absolute path': {
+		'body': '<p>' + ve.dm.mwExample.MWInternalLink.absoluteOpen + 'Foo</a></p>',
+		'data': [
+			{ 'type': 'paragraph' },
+			[
+				'F',
+				[ ve.dm.mwExample.MWInternalLink.absoluteData ]
+			],
+			[
+				'o',
+				[ ve.dm.mwExample.MWInternalLink.absoluteData ]
+			],
+			[
+				'o',
+				[ ve.dm.mwExample.MWInternalLink.absoluteData ]
+			],
+			{ 'type': '/paragraph' },
+			{ 'type': 'internalList' },
+			{ 'type': '/internalList' }
+		],
+		'normalizedBody': '<p><a rel="mw:WikiLink" href="Foo/Bar">Foo</a></p>',
+		'mwConfig': {
+			'wgArticlePath': '/wiki/$1'
+		}
+	},
 	'numbered external link (empty mw:Extlink)': {
 		'body': '<p>Foo<a rel="mw:ExtLink" href="http://www.example.com"></a>Bar</p>',
 		'data': [