revisit of redirect optimization: (#753)

- if a revisit is of a redirect (3xx response) and revisit has http headers, return
the http headers with empty payload -- don't bother loading the original record
builds on changes in #751
- cleanup redirect revisit tests from #751

pywb/warcserver/resource/resolvingloader.py

@@ -75,6 +75,14 @@ class ResolvingLoader(object):
         # two index lookups
         # Case 1: if mimetype is still warc/revisit
         if cdx.get('mime') == 'warc/revisit' and headers_record:
+            if headers_record.http_headers:
+                status = headers_record.http_headers.get_statuscode()
+                # optimization: if redirect, don't load payload record, as it'll be ignored by browser
+                # always replay zero-length payload
+                if status and status.startswith('3'):
+                    headers_record.http_headers.replace_header('Content-Length', '0')
+                    return headers_record, headers_record
+
             payload_record = self._load_different_url_payload(cdx,
                                                               headers_record,
                                                               failed_files,

tests/test_integration.py

@@ -104,13 +104,11 @@ class TestWbIntegration(BaseConfigTest):
     def test_replay_content_head_non_zero_content_length_match(self):
         resp = self.testapp.get('/pywb/20140126200625id_/http://www.iana.org/_js/2013.1/jquery.js', status=200)
         length = resp.content_length
-        print('length', length)
 
         # Content-Length included if non-zero
         resp = self.testapp.head('/pywb/20140126200625id_/http://www.iana.org/_js/2013.1/jquery.js', status=200)
 
         #assert resp.headers['Content-Length'] == length
-        print('length', resp.content_length)
         assert resp.content_length == length
 
     def test_replay_content(self, fmod):

tests/test_redirect_revisits.py

@@ -1,3 +1,4 @@
+from .base_config_test import BaseConfigTest, CollsDirMixin, fmod
 
 from io import BytesIO
 import os
@@ -5,8 +6,6 @@ import os
 from warcio import WARCWriter, StatusAndHeaders
 from pywb.manager.manager import main as wb_manager
 
-from .base_config_test import BaseConfigTest, CollsDirMixin, fmod
-
 
 # ============================================================================
 class TestRevisits(CollsDirMixin, BaseConfigTest):
@@ -125,18 +124,22 @@ class TestRevisits(CollsDirMixin, BaseConfigTest):
         res = self.get('/revisits/20220101{0}/http://example.com/', fmod, status=301)
         assert res.headers["Custom"] == "4"
         assert res.headers["Location"].endswith("/20220101{0}/https://example.com/redirect-4".format(fmod))
-        assert res.text == 'some\ntext'
+        assert res.content_length == 0
+        assert res.text == ''
 
-    def test_different_url_revisit_and_response(self, fmod):
+    def test_different_url_response_and_revisit(self, fmod):
+        # response
         res = self.get('/revisits/20200101{0}/http://example.com/orig-2', fmod, status=301)
         assert res.headers["Custom"] == "2"
         assert res.headers["Location"].endswith("/20200101{0}/https://example.com/redirect-2".format(fmod))
         assert res.text == 'some\ntext'
 
+        # revisit
         res = self.get('/revisits/20220101{0}/http://example.com/orig-2', fmod, status=301)
         assert res.headers["Custom"] == "3"
         assert res.headers["Location"].endswith("/20220101{0}/https://example.com/redirect-3".format(fmod))
-        assert res.text == 'some\ntext'
+        assert res.content_length == 0
+        assert res.text == ''
 
     def test_orig(self, fmod):
         res = self.get('/revisits/20200101{0}/http://example.com/orig-1', fmod, status=301)