server-side rewrite: fix '#' rewriting

- only encode from request, not in WbUrl in general
- tests: add live rewrite test to ensure encoded '#' is used

pywb/apps/rewriterapp.py

@@ -135,6 +135,7 @@ class RewriterApp(object):
         return is_timegate
 
     def render_content(self, wb_url, kwargs, environ):
+        wb_url = wb_url.replace('#', '%23')
         wb_url = WbUrl(wb_url)
         is_timegate = self._check_accept_dt(wb_url, environ)
 

pywb/rewrite/test/test_wburl.py

@@ -54,7 +54,7 @@ u"""
 "('latest_replay', '', '', 'http://example.com?example=2', 'http://example.com?example=2')"
 
 >>> repr(WbUrl('http://example.com/xyz##abc'))
-"('latest_replay', '', '', 'http://example.com/xyz%23%23abc', 'http://example.com/xyz%23%23abc')"
+"('latest_replay', '', '', 'http://example.com/xyz##abc', 'http://example.com/xyz##abc')"
 
 # support urn: prefix
 >>> repr(WbUrl('urn:X-wpull:log'))

pywb/rewrite/wburl.py

@@ -148,7 +148,6 @@ class WbUrl(BaseWbUrl):
 
         if six.PY2 and isinstance(scheme_dom, six.binary_type):
             if scheme_dom == parts[0]:
-                url = url.replace('#', '%23')
                 return url
 
             scheme_dom = scheme_dom.decode('utf-8', 'ignore')
@@ -178,7 +177,6 @@ class WbUrl(BaseWbUrl):
 
             url += rest
 
-        url = url.replace('#', '%23')
         return url
 
     # ======================

tests/test_live_rewriter.py

@@ -24,6 +24,11 @@ class TestLiveRewriter(BaseConfigTest):
         assert '"test": "abc"' in resp.text
         assert resp.status_int == 200
 
+    def test_live_anchor_encode(self, fmod_sl):
+        resp = self.get('/live/{0}httpbin.org/anything/abc%23%23xyz', fmod_sl)
+        assert '"http://httpbin.org/anything/abc##xyz"' in resp.text
+        assert resp.status_int == 200
+
     def test_live_live_frame(self):
         resp = self.testapp.get('/live/http://example.com/')
         assert resp.status_int == 200