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

urlrewrite: generalize support for overridable handle_custom_response() callback for handling modifiers (default support top-frame)

Browse Source 
pass headers to add_custom_params, include error message on error if available
headers: use add_header() to support multiple headers with same name
is_ajax(): check for X-Pywb-Requested-With header to make as ajax and not pass to upstream
This commit is contained in:
Ilya Kreymer 2016-04-07 10:37:40 -07:00
parent fa5d5e6bcc
commit f4cc143dc7
2 changed files with 31 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
urlrewrite/rewriteinputreq.py
View File

@ -42,6 +42,9 @@ class RewriteInputRequest(DirectWSGIInputRequest):
if cookie_val: if cookie_val:
value = cookie_val value = cookie_val
elif name == 'HTTP_X_PYWB_REQUESTED_WITH':
continue
elif name == 'HTTP_X_FORWARDED_PROTO': elif name == 'HTTP_X_FORWARDED_PROTO':
name = 'X-Forwarded-Proto' name = 'X-Forwarded-Proto'
value = self.splits.scheme value = self.splits.scheme

42
urlrewrite/rewriterapp.py
View File

@ -48,16 +48,9 @@ class RewriterApp(object):
rel_prefix = self.get_rel_prefix() rel_prefix = self.get_rel_prefix()
full_prefix = host_prefix + rel_prefix full_prefix = host_prefix + rel_prefix
if self.framed_replay and wb_url.mod == self.frame_mod: resp = self.handle_custom_response(wb_url, full_prefix, host_prefix, kwargs)
extra_params = self.get_top_frame_params(wb_url, kwargs) if resp is not None:
return self.frame_insert_view.get_top_frame(wb_url, return resp
full_prefix,
host_prefix,
request.environ,
self.frame_mod,
self.replay_mod,
coll='',
extra_params=extra_params)
urlrewriter = UrlRewriter(wb_url, urlrewriter = UrlRewriter(wb_url,
prefix=full_prefix, prefix=full_prefix,
@ -90,12 +83,19 @@ class RewriterApp(object):
stream=True) stream=True)
if r.status_code >= 400: if r.status_code >= 400:
error = None
try: try:
error = r.raw.read()
r.raw.close() r.raw.close()
except: except:
pass pass
data = dict(url=url, args=kwargs) if error:
error = error.decode('utf-8')
else:
error = ''
data = dict(url=url, args=kwargs, error=error)
raise HTTPError(r.status_code, exception=data) raise HTTPError(r.status_code, exception=data)
record = self.loader.parse_record_stream(r.raw) record = self.loader.parse_record_stream(r.raw)
@ -105,7 +105,7 @@ class RewriterApp(object):
cdx['timestamp'] = http_date_to_timestamp(r.headers.get('Memento-Datetime')) cdx['timestamp'] = http_date_to_timestamp(r.headers.get('Memento-Datetime'))
cdx['url'] = url cdx['url'] = url
self._add_custom_params(cdx, kwargs) self._add_custom_params(cdx, r.headers, kwargs)
if self.is_ajax(): if self.is_ajax():
head_insert_func = None head_insert_func = None
@ -129,7 +129,7 @@ class RewriterApp(object):
response.status = int(status_headers.get_statuscode()) response.status = int(status_headers.get_statuscode())
for n, v in status_headers.headers: for n, v in status_headers.headers:
response.headers[n] = v response.add_header(n, v)
return gen return gen
@ -158,6 +158,7 @@ class RewriterApp(object):
def is_ajax(self): def is_ajax(self):
value = request.environ.get('HTTP_X_REQUESTED_WITH') value = request.environ.get('HTTP_X_REQUESTED_WITH')
value = value or request.environ.get('HTTP_X_PYWB_REQUESTED_WITH')
if value and value.lower() == 'xmlhttprequest': if value and value.lower() == 'xmlhttprequest':
return True return True
@ -166,9 +167,22 @@ class RewriterApp(object):
def get_upstream_url(self, url, wb_url, closest, kwargs): def get_upstream_url(self, url, wb_url, closest, kwargs):
raise NotImplemented() raise NotImplemented()
def _add_custom_params(self, cdx, kwargs): def _add_custom_params(self, cdx, headers, kwargs):
pass pass
def get_top_frame_params(self, wb_url, kwargs): def get_top_frame_params(self, wb_url, kwargs):
return None return None
def handle_custom_response(self, wb_url, full_prefix, host_prefix, kwargs):
if self.framed_replay and wb_url.mod == self.frame_mod:
extra_params = self.get_top_frame_params(wb_url, kwargs)
return self.frame_insert_view.get_top_frame(wb_url,
full_prefix,
host_prefix,
request.environ,
self.frame_mod,
self.replay_mod,
coll='',
extra_params=extra_params)
return None