backup/pywb
1
0
Fork 0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-24 06:59:52 +01:00
Code Issues Projects Releases Activity

replay/memento: always include 'Content-Location' for in no-redir mode replay (not just for memento timegate), #122

Browse Source
This commit is contained in:
Ilya Kreymer 2015-07-19 00:11:25 -07:00
parent 66f5ad62b3
commit c2f99d6cfd
3 changed files with 22 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
pywb/framework/memento.py
View File

@ -106,10 +106,10 @@ class MementoRespMixin(object):
timestamp=ts, timestamp=ts,
url=url) url=url)
# Must set content location # set in replay_views -- Must set content location
if is_memento and is_timegate: #if is_memento and is_timegate:
self.status_headers.headers.append(('Content-Location', # self.status_headers.headers.append(('Content-Location',
canon_link)) # canon_link))
# don't set memento link for very long urls... # don't set memento link for very long urls...
if len(canon_link) < 512: if len(canon_link) < 512:

10
pywb/webapp/replay_views.py
View File

@ -191,6 +191,16 @@ class ReplayView(object):
response_iter, response_iter,
self.buffer_max_size) self.buffer_max_size)
# Set Content-Location if not exact capture
if not self.redir_to_exact:
mod = wbrequest.options.get('replay_mod', wbrequest.wb_url.mod)
canon_url = (wbrequest.urlrewriter.
get_new_url(timestamp=cdx['timestamp'],
url=cdx['url'],
mod=mod))
status_headers.headers.append(('Content-Location', canon_url))
response = self.response_class(status_headers, response = self.response_class(status_headers,
response_iter, response_iter,
wbrequest=wbrequest, wbrequest=wbrequest,

16
tests/test_integration.py
View File

@ -303,23 +303,23 @@ class TestWb:
def test_redirect_non_exact_latest_replay_ts(self): def test_redirect_non_exact_latest_replay_ts(self):
resp = self.testapp.get('/pywb-non-exact/http://example.com/') resp = self.testapp.get('/pywb-non-exact/http://example.com/')
assert resp.status_int == 302 assert resp.status_int == 200
assert resp.headers['Location'].endswith('/http://example.com') assert resp.headers['Content-Location'].endswith('/http://example.com')
# extract ts, which should be current time # extract ts, which should be current time
ts = resp.headers['Location'].rsplit('/http://')[0].rsplit('/', 1)[-1] ts = resp.headers['Content-Location'].rsplit('/http://')[0].rsplit('/', 1)[-1]
assert len(ts) == 14, ts assert ts == '20140127171251'
resp = resp.follow() #resp = resp.follow()
self._assert_basic_html(resp) #self._assert_basic_html(resp)
# ensure the current ts is present in the links # ensure the current ts is present in the links
assert '"{0}"'.format(ts) in resp.body assert '"{0}"'.format(ts) in resp.body
assert '/pywb-non-exact/{0}/http://www.iana.org/domains/example'.format(ts) in resp.body assert '/pywb-non-exact/http://www.iana.org/domains/example' in resp.body
# ensure ts is current ts # ensure ts is current ts
assert timestamp_now() >= ts, ts #assert timestamp_now() >= ts, ts
def test_redirect_relative_3(self): def test_redirect_relative_3(self):
# webtest uses Host: localhost:80 by default # webtest uses Host: localhost:80 by default