From e70448e53a138cef51327d5c532c700dc4a5570f Mon Sep 17 00:00:00 2001
From: Gabriel Wicke <gwicke@wikimedia.org>
Date: Thu, 24 May 2012 10:18:41 +0200
Subject: [PATCH] Use text/x-mediawiki content type, and handle tokenizer
 errors without --debug

Change-Id: I154cd344306aa05ada7ff30f631d487f39fa9739
---
 api/ParserService.js                      |  2 +-
 modules/parser/mediawiki.tokenizer.peg.js | 31 +++++++++++++++--------
 2 files changed, 21 insertions(+), 12 deletions(-)

diff --git a/api/ParserService.js b/api/ParserService.js
index 1f56b56ad4..5e542d19fb 100644
--- a/api/ParserService.js
+++ b/api/ParserService.js
@@ -149,7 +149,7 @@ app.get(/\/(.*)/, function(req, res){
  */
 app.post(/\/(.*)/, function(req, res){
 	env.pageName = req.params[0];
-	res.setHeader('Content-Type', 'text/plain; charset=UTF-8');
+	res.setHeader('Content-Type', 'text/x-mediawiki; charset=UTF-8');
 	var p = new html5.Parser();
 	p.parse( req.body.content );
 	new WikitextSerializer({env: env}).serializeDOM( 
diff --git a/modules/parser/mediawiki.tokenizer.peg.js b/modules/parser/mediawiki.tokenizer.peg.js
index a429af53ec..1741770c76 100644
--- a/modules/parser/mediawiki.tokenizer.peg.js
+++ b/modules/parser/mediawiki.tokenizer.peg.js
@@ -91,15 +91,28 @@ PegTokenizer.prototype.process = function( text, cacheKey ) {
 		text += "\n";
 	}
 
+	var chunkCB;
+	if ( this.canCache ) {
+		chunkCB = this.onCacheChunk.bind( this );
+	} else {
+		chunkCB = this.emit.bind( this, 'chunk' );
+	}
 	// XXX: Commented out exception handling during development to get
 	// reasonable traces.
-	//try {
-		var chunkCB;
-		if ( this.canCache ) {
-			chunkCB = this.onCacheChunk.bind( this );
-		} else {
-			chunkCB = this.emit.bind( this, 'chunk' );
+	if ( ! this.env.debug ) {
+		try {
+			this.tokenizer.tokenize(text, 'start', 
+					// callback
+					chunkCB,
+					// inline break test
+					this
+					);
+			this.onEnd();
+		} catch (e) {
+			console.warn( 'Tokenizer error in ' + cacheKey + ': ' + e );
+			console.trace();
 		}
+	} else {
 		this.tokenizer.tokenize(text, 'start', 
 				// callback
 				chunkCB,
@@ -107,11 +120,7 @@ PegTokenizer.prototype.process = function( text, cacheKey ) {
 				this
 				);
 		this.onEnd();
-	/*} catch (e) {
-		err = e;
-		console.warn( e );
-		console.trace();
-	}*/
+	}
 };
 
 PegTokenizer.prototype.onCacheChunk = function ( chunk ) {