backup/pywb
1
0
Fork 0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-15 00:03:28 +01:00
Code Issues Projects Releases Activity

memento: add rel="memento" header to timegate as well, improve memento test, clearly differntiate between

Browse Source 
timegate redirect and intermediate resource redirect, related to #70
This commit is contained in:
Ilya Kreymer 2015-02-16 09:59:03 -08:00
parent c4d5dd4690
commit 9623f95439
3 changed files with 32 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
pywb/framework/memento.py
View File

@ -29,6 +29,7 @@ class MementoReqMixin(object):
raise BadRequestException('Invalid Accept-Datetime: ' +
accept_datetime)
# note: this changes from LATEST_REPLAY -> REPLAY
self.wb_url.set_replay_timestamp(timestamp)
@ -68,26 +69,29 @@ class MementoRespMixin(object):
elif wbrequest.options['is_proxy']:
is_memento = True
# otherwise only for replay
else:
# otherwise only if timestamp replay (and not a timegate)
elif not is_timegate:
is_memento = (wbrequest.wb_url.type == wbrequest.wb_url.REPLAY)
link = []
if is_memento:
if is_memento or is_timegate:
if cdx:
http_date = timestamp_to_http_date(cdx['timestamp'])
ts = cdx['timestamp']
# for top frame
elif wbrequest.wb_url.timestamp:
http_date = timestamp_to_http_date(wbrequest.wb_url.timestamp)
ts = wbrequest.wb_url.timestamp
else:
http_date = None
ts = None
if http_date:
self.status_headers.headers.append(('Memento-Datetime',
http_date))
if ts:
http_date = timestamp_to_http_date(ts)
canon_link = wbrequest.urlrewriter.get_new_url(mod='')
if is_memento:
self.status_headers.headers.append(('Memento-Datetime',
http_date))
canon_link = wbrequest.urlrewriter.get_new_url(mod='', timestamp=ts)
link.append(self.make_memento_link(canon_link,
'memento',
http_date))

11
pywb/webapp/replay_views.py
View File

@ -219,7 +219,8 @@ class ReplayView(object):
if wbrequest.custom_params.get('noredir'):
return None
redir_needed = (wbrequest.options.get('is_timegate', False))
is_memento_timegate = (wbrequest.options.get('is_timegate', False))
redir_needed = is_memento_timegate
if not redir_needed and self.redir_to_exact:
redir_needed = (cdx['timestamp'] != wbrequest.wb_url.timestamp)
@ -238,15 +239,19 @@ class ReplayView(object):
# FF shows a confirm dialog, so can't use 307 effectively
# was: statusline = '307 Same-Method Internal Redirect'
return None
elif is_memento_timegate:
statusline = '302 Found'
else:
# clear cdx line to indicate internal redirect
statusline = '302 Internal Redirect'
cdx = None
status_headers = StatusAndHeaders(statusline,
[('Location', new_url)])
# don't include cdx to indicate internal redirect
return self.response_class(status_headers,
wbrequest=wbrequest)
wbrequest=wbrequest,
cdx=cdx)
def _reject_self_redirect(self, wbrequest, cdx, status_headers):
"""

16
tests/test_memento.py
View File

@ -39,6 +39,8 @@ class TestWb:
"""
TimeGate with no Accept-Datetime header
"""
dt = 'Mon, 27 Jan 2014 17:12:39 GMT'
resp = self.testapp.get('/pywb/http://www.iana.org/_css/2013.1/screen.css')
assert resp.status_int == 302
@ -48,6 +50,7 @@ class TestWb:
links = self.get_links(resp)
assert '<http://www.iana.org/_css/2013.1/screen.css>; rel="original"' in links
assert self.make_timemap_link('http://www.iana.org/_css/2013.1/screen.css') in links
assert self.make_memento_link('http://www.iana.org/_css/2013.1/screen.css', '20140127171239', dt) in links
assert MEMENTO_DATETIME not in resp.headers
@ -56,9 +59,10 @@ class TestWb:
def test_timegate_accept_datetime(self):
"""
TimeGate with Accept-Datetime header
TimeGate with Accept-Datetime header, matching exactly
"""
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04'}
dt = 'Sun, 26 Jan 2014 20:08:04 GMT'
headers = {ACCEPT_DATETIME: dt}
resp = self.testapp.get('/pywb//http://www.iana.org/_css/2013.1/screen.css', headers=headers)
assert resp.status_int == 302
@ -68,7 +72,7 @@ class TestWb:
links = self.get_links(resp)
assert '<http://www.iana.org/_css/2013.1/screen.css>; rel="original"' in links
assert self.make_timemap_link('http://www.iana.org/_css/2013.1/screen.css') in links
assert self.make_memento_link('http://www.iana.org/_css/2013.1/screen.css', '20140126200804', dt) == links[0], links[0]
assert MEMENTO_DATETIME not in resp.headers
@ -79,7 +83,7 @@ class TestWb:
"""
Not a timegate, but an 'intermediate resource', redirect to closest timestamp
"""
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04'}
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04 GMT'}
# not a timegate, partial timestamp /2014/ present
resp = self.testapp.get('/pywb/2014/http://www.iana.org/_css/2013.1/screen.css', headers=headers)
@ -286,7 +290,7 @@ rel="memento"; datetime="Fri, 03 Jan 2014 03:03:41 GMT"'
# simulate proxy mode by setting REQUEST_URI
request_uri = 'http://www.iana.org/_css/2013.1/screen.css'
extra = dict(REQUEST_URI=request_uri, SCRIPT_NAME='')
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04'}
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04 GMT'}
resp = self.testapp.get('/x-ignore-this-x', extra_environ=extra, headers=headers)
@ -330,7 +334,7 @@ rel="memento"; datetime="Fri, 03 Jan 2014 03:03:41 GMT"'
"""
Non WbUrl memento path -- just ignore ACCEPT_DATETIME
"""
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04'}
headers = {ACCEPT_DATETIME: 'Sun, 26 Jan 2014 20:08:04 GMT'}
resp = self.testapp.get('/pywb/', headers=headers)
assert resp.status_int == 200