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

referer redirect fixes:

Browse Source 
- allow redirect if current Host: matches
- redirect request uri to host root, not current host path
This commit is contained in:
Ilya Kreymer 2014-02-09 20:19:43 +00:00
parent 2a828fab32
commit 8b2bfa570c
1 changed files with 19 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
pywb/archivalrouter.py
View File

@ -135,8 +135,8 @@ class ReferRedirect:
>>> ReferRedirect(['http://example:9090/']).match_prefixs >>> ReferRedirect(['http://example:9090/']).match_prefixs
['http://example:9090/'] ['http://example:9090/']
>>> test_redir('http://localhost:8080/', '/other.html', 'http://localhost:8080/coll/20131010/http://example.com/path/page.html') >>> test_redir('http://localhost:8080/', '/diff_path/other.html', 'http://localhost:8080/coll/20131010/http://example.com/path/page.html')
'http://localhost:8080/coll/20131010/http://example.com/path/other.html' 'http://localhost:8080/coll/20131010/http://example.com/diff_path/other.html'
>>> test_redir('http://localhost:8080/', '/../other.html', 'http://localhost:8080/coll/20131010/http://example.com/path/page.html') >>> test_redir('http://localhost:8080/', '/../other.html', 'http://localhost:8080/coll/20131010/http://example.com/path/page.html')
'http://localhost:8080/coll/20131010/http://example.com/other.html' 'http://localhost:8080/coll/20131010/http://example.com/other.html'
@ -146,7 +146,7 @@ class ReferRedirect:
# Custom collection # Custom collection
>>> test_redir('http://localhost:8080/', '/other.html', 'http://localhost:8080/complex/123/20131010/http://example.com/path/page.html', coll='complex/123') >>> test_redir('http://localhost:8080/', '/other.html', 'http://localhost:8080/complex/123/20131010/http://example.com/path/page.html', coll='complex/123')
'http://localhost:8080/complex/123/20131010/http://example.com/path/other.html' 'http://localhost:8080/complex/123/20131010/http://example.com/other.html'
# With timestamp included # With timestamp included
>>> test_redir('http://localhost:8080/', '/20131010/other.html', 'http://localhost:8080/coll/20131010/http://example.com/index.html') >>> test_redir('http://localhost:8080/', '/20131010/other.html', 'http://localhost:8080/coll/20131010/http://example.com/index.html')
@ -160,6 +160,10 @@ class ReferRedirect:
>>> test_redir('http://example:8080/', '/other.html', 'http://localhost:8080/coll/20131010/http://example.com/path/page.html') >>> test_redir('http://example:8080/', '/other.html', 'http://localhost:8080/coll/20131010/http://example.com/path/page.html')
False False
# Right Host
>>> test_redir('http://localhost:8080/', '/other.html', 'http://example.com:8080/coll/20131010/http://example.com/path/page.html', http_host = 'example.com:8080')
'http://example.com:8080/coll/20131010/http://example.com/other.html'
# With custom SCRIPT_NAME # With custom SCRIPT_NAME
>>> test_redir('http://localhost:8080/', '/../other.html', 'http://localhost:8080/extra/coll/20131010/http://example.com/path/page.html', '/extra') >>> test_redir('http://localhost:8080/', '/../other.html', 'http://localhost:8080/extra/coll/20131010/http://example.com/path/page.html', '/extra')
'http://localhost:8080/extra/coll/20131010/http://example.com/other.html' 'http://localhost:8080/extra/coll/20131010/http://example.com/other.html'
@ -188,13 +192,14 @@ class ReferRedirect:
if referrer is None: if referrer is None:
return None return None
# ensure referrer starts with one of allowed hosts
if not any (referrer.startswith(i) for i in self.match_prefixs):
return None
# get referrer path name # get referrer path name
ref_split = urlparse.urlsplit(referrer) ref_split = urlparse.urlsplit(referrer)
# ensure referrer starts with one of allowed hosts
if not any (referrer.startswith(i) for i in self.match_prefixs):
if ref_split.netloc != env.get('HTTP_HOST'):
return None
path = ref_split.path path = ref_split.path
app_path = env['SCRIPT_NAME'] app_path = env['SCRIPT_NAME']
@ -222,9 +227,9 @@ class ReferRedirect:
rewriter = ref_request.urlrewriter rewriter = ref_request.urlrewriter
rel_request_uri = env['REL_REQUEST_URI'][1:] rel_request_uri = env['REL_REQUEST_URI']
timestamp_path = rewriter.wburl.timestamp + '/' timestamp_path = '/' + rewriter.wburl.timestamp + '/'
# check if timestamp is already part of the path # check if timestamp is already part of the path
if rel_request_uri.startswith(timestamp_path): if rel_request_uri.startswith(timestamp_path):
@ -242,8 +247,12 @@ if __name__ == "__main__" or utils.enable_doctests():
import handlers import handlers
def test_redir(match_host, request_uri, referrer, script_name = '', coll = 'coll'): def test_redir(match_host, request_uri, referrer, script_name = '', coll = 'coll', http_host = None):
env = {'REL_REQUEST_URI': request_uri, 'HTTP_REFERER': referrer, 'SCRIPT_NAME': script_name} env = {'REL_REQUEST_URI': request_uri, 'HTTP_REFERER': referrer, 'SCRIPT_NAME': script_name}
if http_host:
env['HTTP_HOST'] = http_host
routes = [Route(coll, handlers.BaseHandler())] routes = [Route(coll, handlers.BaseHandler())]
redir = ReferRedirect(match_host) redir = ReferRedirect(match_host)