Fix for 404 occuring when signing out during the creation of, for example,
a new Document. This patch makes sure that the request.path is not altered
when requesting the create page (alteration of which caused the 404). This
also results in the seeds now being remembered when being forced to log in.
The editPost method now correctly switches to the edit path after a
successful edit.
Patch by: Lennard de Rijk
import re
from django.utils.text import compress_string
from django.utils.cache import patch_vary_headers
re_accepts_gzip = re.compile(r'\bgzip\b')
class GZipMiddleware(object):
"""
This middleware compresses content if the browser allows gzip compression.
It sets the Vary header accordingly, so that caches will base their storage
on the Accept-Encoding header.
"""
def process_response(self, request, response):
# It's not worth compressing non-OK or really short responses.
if response.status_code != 200 or len(response.content) < 200:
return response
patch_vary_headers(response, ('Accept-Encoding',))
# Avoid gzipping if we've already got a content-encoding.
if response.has_header('Content-Encoding'):
return response
# Older versions of IE have issues with gzipped pages containing either
# Javascript and PDF.
if "msie" in request.META.get('HTTP_USER_AGENT', '').lower():
ctype = response.get('Content-Type', '').lower()
if "javascript" in ctype or ctype == "application/pdf":
return response
ae = request.META.get('HTTP_ACCEPT_ENCODING', '')
if not re_accepts_gzip.search(ae):
return response
response.content = compress_string(response.content)
response['Content-Encoding'] = 'gzip'
response['Content-Length'] = str(len(response.content))
return response