From 7feb0893eb3ec9a691c080d6fdf02839980c8357 Mon Sep 17 00:00:00 2001
From: Ilya Kreymer <ikreymer@gmail.com>
Date: Tue, 30 Sep 2014 10:57:25 -0700
Subject: [PATCH] rewrite: add 'application/json' to a seperate 'json' regex
 rewriter type (rewrite links only, no http), can be customized via rules
 wombat: add rewrite_style for rewriting style attrs query: don't include any
 filter in latest, custom filter can be used without any other filters tests:
 fix typos in tests

---
 pywb/rewrite/header_rewriter.py  |  2 ++
 pywb/rewrite/regex_rewriters.py  |  2 +-
 pywb/rewrite/rewriterules.py     |  1 +
 pywb/static/wombat.js            | 27 ++++++++++++++++++++-------
 pywb/webapp/query_handler.py     | 11 +++++++++--
 tests/test_config.yaml           |  1 +
 tests/test_proxy_http_cookie.py  |  1 -
 tests/test_proxy_https_cookie.py |  1 -
 8 files changed, 34 insertions(+), 12 deletions(-)

diff --git a/pywb/rewrite/header_rewriter.py b/pywb/rewrite/header_rewriter.py
index 2d505e88..950817d4 100644
--- a/pywb/rewrite/header_rewriter.py
+++ b/pywb/rewrite/header_rewriter.py
@@ -28,6 +28,8 @@ class HeaderRewriter:
                  'application/javascript',
                  'application/x-javascript'],
 
+        'json': ['application/json'],
+
         'xml':  ['/xml', '+xml', '.xml', '.rss'],
     }
 
diff --git a/pywb/rewrite/regex_rewriters.py b/pywb/rewrite/regex_rewriters.py
index df7a128e..97151190 100644
--- a/pywb/rewrite/regex_rewriters.py
+++ b/pywb/rewrite/regex_rewriters.py
@@ -123,7 +123,7 @@ class JSLinkAndLocationRewriter(JSLinkOnlyRewriter):
 
     def __init__(self, rewriter, rules=[], prefix='WB_wombat_'):
         rules = rules + [
-             (r'(?<!/)\blocation\b', RegexRewriter.add_prefix(prefix), 0),
+             (r'(?<!/)\blocation\b(?!\":)', RegexRewriter.add_prefix(prefix), 0),
              (r'(?<=document\.)domain', RegexRewriter.add_prefix(prefix), 0),
              (r'(?<=document\.)referrer', RegexRewriter.add_prefix(prefix), 0),
              (r'(?<=document\.)cookie', RegexRewriter.add_prefix(prefix), 0),
diff --git a/pywb/rewrite/rewriterules.py b/pywb/rewrite/rewriterules.py
index 226942c5..880540f7 100644
--- a/pywb/rewrite/rewriterules.py
+++ b/pywb/rewrite/rewriterules.py
@@ -22,6 +22,7 @@ class RewriteRules(BaseRule):
         self.rewriters['css'] = config.get('css_class', CSSRewriter)
         self.rewriters['xml'] = config.get('xml_class', XMLRewriter)
         self.rewriters['html'] = config.get('html_class', HTMLRewriter)
+        self.rewriters['json'] = config.get('json_class', JSLinkOnlyRewriter)
 
         self.parse_comments = config.get('parse_comments', False)
 
diff --git a/pywb/static/wombat.js b/pywb/static/wombat.js
index 962a32cd..d1fd3db0 100644
--- a/pywb/static/wombat.js
+++ b/pywb/static/wombat.js
@@ -280,7 +280,6 @@ WB_wombat_init = (function() {
         var href = extract_orig(this._orig_href);
         parser.href = href;
         
-        //console.log(this._orig_href + " -> " + tmp_href);
         this._autooverride = false;
         
         var _set_hash = function(hash) {
@@ -495,7 +494,7 @@ WB_wombat_init = (function() {
     }
 
     //============================================
-    function rewrite_attr(elem, name) {
+    function rewrite_attr(elem, name, func) {
         if (!elem || !elem.getAttribute) {
             return;
         }
@@ -510,17 +509,31 @@ WB_wombat_init = (function() {
             return;
         }
 
-        //var orig_value = value;        
-        value = rewrite_url(value);
+        value = func(value);
 
         elem.setAttribute(name, value);
     }
-    
+
+    //============================================
+    function rewrite_style(value)
+    {
+        //console.log("style rewrite: " + value);
+
+        STYLE_REGEX = /(url\s*\(\s*[\\"']*)([^)'"]+)([\\"']*\s*\))/g;
+
+        function style_replacer(match, n1, n2, n3, offset, string) {
+            return n1 + rewrite_url(n2) + n3;
+        }
+
+        return value.replace(STYLE_REGEX, style_replacer);
+    }
+
     //============================================
     function rewrite_elem(elem)
     {
-        rewrite_attr(elem, "src");
-        rewrite_attr(elem, "href");
+        rewrite_attr(elem, "src", rewrite_url);
+        rewrite_attr(elem, "href", rewrite_url);
+        rewrite_attr(elem, "style", rewrite_style);
         
         if (elem && elem.getAttribute && elem.getAttribute("crossorigin")) {
             elem.removeAttribute("crossorigin");
diff --git a/pywb/webapp/query_handler.py b/pywb/webapp/query_handler.py
index 92526179..ee03b8b9 100644
--- a/pywb/webapp/query_handler.py
+++ b/pywb/webapp/query_handler.py
@@ -76,10 +76,15 @@ class QueryHandler(object):
         return self.make_cdx_response(wbrequest, cdx_iter, params['output'])
 
     def load_cdx(self, wbrequest, params):
+        print(params)
         if wbrequest:
             # add any custom filter from the request
             if wbrequest.query_filter:
-                params['filter'].extend(wbrequest.query_filter)
+                filters = params.get('filter')
+                if filters:
+                    filters.extend(wbrequest.query_filter)
+                else:
+                    params['filter'] = wbrequest.query_filter
 
             if wbrequest.custom_params:
                 params.update(wbrequest.custom_params)
@@ -144,7 +149,9 @@ class QueryHandler(object):
 
             wburl.LATEST_REPLAY:
                 {'sort': 'reverse',
-                 'filter': ['statuscode:[23]..|-'],
+       # Not appropriate as default
+       # Should be an option to configure status code filtering in general
+       #         'filter': ['statuscode:[23]..|-'],
                  'limit': '1',
                  'resolveRevisits': True,
                 }
diff --git a/tests/test_config.yaml b/tests/test_config.yaml
index 2d0fc5b6..ea0f21a9 100644
--- a/tests/test_config.yaml
+++ b/tests/test_config.yaml
@@ -92,6 +92,7 @@ absoulte_paths: true
 # <route>: <package or file path>
 static_routes:
           static/test/route: pywb/static/
+          static/default: pywb/static/
 
 # Enable simple http proxy mode
 enable_http_proxy: true
diff --git a/tests/test_proxy_http_cookie.py b/tests/test_proxy_http_cookie.py
index c7330984..a79ef308 100644
--- a/tests/test_proxy_http_cookie.py
+++ b/tests/test_proxy_http_cookie.py
@@ -94,7 +94,6 @@ class TestProxyHttpCookie:
         global sesh_key
         sesh_key = sesh1
 
-        global sesh_key
         sesh2 = self.session.cookies.get('__pywb_proxy_sesh', domain='.iana.org')
         assert sesh_key == sesh2
 
diff --git a/tests/test_proxy_https_cookie.py b/tests/test_proxy_https_cookie.py
index 91f24610..bfd1d398 100644
--- a/tests/test_proxy_https_cookie.py
+++ b/tests/test_proxy_https_cookie.py
@@ -108,7 +108,6 @@ class TestProxyHttpsCookie:
         global sesh_key
         sesh_key = sesh1
 
-        global sesh_key
         sesh2 = self.session.cookies.get('__pywb_proxy_sesh', domain='.iana.org')
         assert sesh_key == sesh2