cookie_rewriter: ensure cookie paths are always relative

ikreymer · ikreymer · commit 466968e97444 · 2015-07-31T11:37:50.000-07:00
cookie_rewriter tests: add cookie rewriter tests for secure, httponly
html_rewriter tests: add &lt;base&gt; rel and abs rewrite tests
no cover for waitress as its not used by default
diff --git a/pywb/apps/cli.py b/pywb/apps/cli.py
@@ -49,12 +49,12 @@ def load(self):  #pragma: no cover
         pass
 
     def run(self):
-        if self.r.server == 'waitress':
+        if self.r.server == 'waitress':  #pragma: no cover
             self.run_waitress()
         else:
             self.run_wsgiref()
 
-    def run_waitress(self):
+    def run_waitress(self):  #pragma: no cover
         from waitress import serve
         print(self.desc)
         serve(self.application, port=self.r.port, threads=self.r.threads)
diff --git a/pywb/rewrite/cookie_rewriter.py b/pywb/rewrite/cookie_rewriter.py
@@ -18,7 +18,14 @@ def rewrite(self, cookie_str, header='Set-Cookie'):
 
         for name, morsel in cookie.iteritems():
             morsel = self.rewrite_cookie(name, morsel)
+
             if morsel:
+                path = morsel.get('path')
+                if path:
+                    inx = path.find(self.url_rewriter.rel_prefix)
+                    if inx > 0:
+                        morsel['path'] = path[inx:]
+
                 results.append((header, morsel.OutputString()))
 
         return results
@@ -78,23 +85,14 @@ class HostScopeCookieRewriter(WbUrlBaseCookieRewriter):
     """
 
     def rewrite_cookie(self, name, morsel):
-        new_path = None
-
         # if domain set, expand cookie to host prefix
         if morsel.get('domain'):
             del morsel['domain']
-            new_path = self.url_rewriter.rewrite('/')
+            morsel['path'] = self.url_rewriter.rewrite('/')
 
         # set cookie to rewritten path
         elif morsel.get('path'):
-            new_path = self.url_rewriter.rewrite(morsel['path'])
-
-        if new_path:
-            inx = new_path.find(self.url_rewriter.rel_prefix)
-            if inx > 0:
-                new_path = new_path[inx:]
-
-            morsel['path'] = new_path
+            morsel['path'] = self.url_rewriter.rewrite(morsel['path'])
 
         self._remove_age_opts(morsel)
         return morsel
diff --git a/pywb/rewrite/test/test_cookie_rewriter.py b/pywb/rewrite/test/test_cookie_rewriter.py
@@ -34,12 +34,15 @@
 
 
 # HostCookieRewriter -- set path to host
->>> rewrite_cookie('some=value; Path=/diff/path/;', urlrewriter, 'host')
+>>> rewrite_cookie('some=value; Path=/diff/path/', urlrewriter, 'host')
 [('Set-Cookie', 'some=value; Path=/pywb/20131226101010/http://example.com/diff/path/')]
 
 >>> rewrite_cookie('some=value; Domain=.example.com; Path=/diff/path/; Max-Age=1500', urlrewriter, 'host')
 [('Set-Cookie', 'some=value; Path=/pywb/20131226101010/http://example.com/')]
 
+>>> rewrite_cookie('some=value; Domain=.example.com; Secure; Path=/diff/path/; HttpOnly; Max-Age=1500', urlrewriter, 'host')
+[('Set-Cookie', 'some=value; httponly; Path=/pywb/20131226101010/http://example.com/')]
+
 
 # RootCookieRewriter -- always sets Path=/, removes Domain
 >>> rewrite_cookie('some=value; Path=/diff/path/;', urlrewriter, 'root')
@@ -62,7 +65,7 @@
 from pywb.rewrite.cookie_rewriter import MinimalScopeCookieRewriter, get_cookie_rewriter
 from pywb.rewrite.url_rewriter import UrlRewriter
 
-urlrewriter = UrlRewriter('20131226101010/http://example.com/some/path/index.html', '/pywb/')
+urlrewriter = UrlRewriter('20131226101010/http://example.com/some/path/index.html', 'http://localhost:8080/pywb/', rel_prefix='/pywb/')
 
 urlrewriter2 = UrlRewriter('em_/http://example.com/', '/preview/')
 
diff --git a/pywb/rewrite/test/test_html_rewriter.py b/pywb/rewrite/test/test_html_rewriter.py
@@ -24,6 +24,14 @@
 >>> parse('<html><head><base href="http://example.com/diff/path/file.html"/>')
 <html><head><base href="/web/20131226101010/http://example.com/diff/path/file.html"/>
 
+# Full Path
+>>> parse('<html><head><base href="http://example.com/diff/path/file.html"/>', urlrewriter=full_path_urlrewriter)
+<html><head><base href="http://localhost:80/web/20131226101010/http://example.com/diff/path/file.html"/>
+
+# Rel Base
+>>> parse('<html><head><base href="/other/file.html"/>', urlrewriter=full_path_urlrewriter)
+<html><head><base href="/web/20131226101010/http://example.com/other/file.html"/>
+
 >>> parse('<base href="static/"/><img src="image.gif"/>')
 <base href="/web/20131226101010/http://example.com/some/path/static/"/><img src="/web/20131226101010im_/http://example.com/some/path/static/image.gif"/>
 
@@ -186,6 +194,10 @@
                           '/web/',
                           rewrite_opts=dict(punycode_links=False))
 
+full_path_urlrewriter = UrlRewriter('20131226101010/http://example.com/some/path/index.html',
+                                   'http://localhost:80/web/',
+                                   rewrite_opts=dict(punycode_links=False))
+
 urlrewriter_pencode = UrlRewriter('20131226101010/http://example.com/some/path/index.html',
                                    '/web/',
                                    rewrite_opts=dict(punycode_links=True))
