From ac24c9a656bb713dc29b86f54ca74dc435d1a16b Mon Sep 17 00:00:00 2001
From: Pierre-Yves Chibon <pingou@pingoured.fr>
Date: Dec 29 2015 08:59:21 +0000
Subject: Rely on urllib and aiohttp.MultiDict to do the url arguments parsing


This gives us a much more robust url argument parsing and corresponds
to what's in the doc:
http://aiohttp.readthedocs.org/en/stable/server.html#handling-get-params

---

diff --git a/mdapi/__init__.py b/mdapi/__init__.py
index 89965d9..4b9b07f 100644
--- a/mdapi/__init__.py
+++ b/mdapi/__init__.py
@@ -23,6 +23,7 @@
 Top level of the mdapi aiohttp application.
 '''
 import os
+import urllib
 
 try:
     import simplejson as json
@@ -31,7 +32,7 @@ except ImportError:
 
 import asyncio
 import werkzeug
-from aiohttp import web
+from aiohttp import web, MultiDict
 
 import mdapi.lib as mdapilib
 import mdapi.file_lock as file_lock
@@ -95,8 +96,8 @@ def _get_pkg(branch, name):
 
 def _get_pretty(request):
     pretty = False
-    query_string = request.query_string.lower()
-    if query_string in ['pretty=1', 'pretty=true']:
+    get_params = MultiDict(urllib.parse.parse_qsl(request.query_string.lower()))
+    if str(get_params.get('pretty', None)) in ['1', 'true']:
         pretty = True
     # Assume pretty if html is requested and pretty is not disabled
     elif 'text/html' in request.headers.get('ACCEPT', ''):