From 2e4d78d0791143b69ac77c4876727cfb73edcbbe Mon Sep 17 00:00:00 2001
From: Ilya Kreymer <ilya@archive.org>
Date: Sun, 19 Jan 2014 16:51:17 -0800
Subject: [PATCH] request_uri: only generate REQUEST_URI manually if not
 provided by wsgi framework only encode chars that are not allowed in path
 segment, per http://tools.ietf.org/html/rfc3986#section-3.3

---
 pywb/utils.py | 20 +++++++++++++++++---
 pywb/wbapp.py |  4 +++-
 2 files changed, 20 insertions(+), 4 deletions(-)

diff --git a/pywb/utils.py b/pywb/utils.py
index c74b777f..52c74de8 100644
--- a/pywb/utils.py
+++ b/pywb/utils.py
@@ -107,13 +107,27 @@ def iso_date_to_timestamp(string):
     return datetime_to_timestamp(iso_date_to_datetime(string))
 
 
-# adapted from wsgiref.request_uri, but doesn't include domain name and allows ':' in url
+# adapted -from wsgiref.request_uri, but doesn't include domain name and allows all characters
+# allowed in the path segment according to: http://tools.ietf.org/html/rfc3986#section-3.3
+# explained here: http://stackoverflow.com/questions/4669692/valid-characters-for-directory-part-of-a-url-for-short-links
 def request_uri(environ, include_query=1):
-    """Return the requested path, optionally including the query string"""
+    """
+    Return the requested path, optionally including the query string
+
+    # Simple test:
+    >>> request_uri({'PATH_INFO': '/web/example.com'})
+    '/web/example.com'
+
+    # Test all unecoded special chars and double-quote
+    # (double-quote must be encoded but not single quote)
+    >>> request_uri({'PATH_INFO': "/web/example.com/0~!+$&'()*+,;=:\\\""})
+    "/web/example.com/0~!+$&'()*+,;=:%22"
+    """
     from urllib import quote
-    url = quote(environ.get('SCRIPT_NAME', '')+environ.get('PATH_INFO',''),safe='/;=,:')
+    url = quote(environ.get('SCRIPT_NAME', '') + environ.get('PATH_INFO',''), safe='/~!$&\'()*+,;=:@')
     if include_query and environ.get('QUERY_STRING'):
         url += '?' + environ['QUERY_STRING']
+
     return url
 
 
diff --git a/pywb/wbapp.py b/pywb/wbapp.py
index b81467f0..3e425608 100644
--- a/pywb/wbapp.py
+++ b/pywb/wbapp.py
@@ -82,7 +82,9 @@ except:
 
 
 def application(env, start_response):
-    env['REQUEST_URI'] = request_uri(env)
+    if not env.get('REQUEST_URI'):
+        env['REQUEST_URI'] = request_uri(env)
+
     response = None
 
     try: