Commits

Author Commit Message Labels Comments Date
Randy Syring
adding tests for previous fix Even though I added the changes first, I backed them out after commit to make sure this test fails. Then put them back in and the test passes. #75 Session.regenerate_id() needed to help avoid session hijacking
Randy Syring
fixes for regenerate_id() #75 Session.regenerate_id() needed to help avoid session hijacking
Randy Syring
version needs to be 1.6 instead of 0.6
Randy Syring
session: re-factored cookie related settings, added httponly support httponly support is Python 2.6+ only, warning will be given for usage on < 2.6 also re-factored cookie related settings to try and avoid duplicated code. refs #62 Some way to set httponly cookies
Randy Syring
need with_statement future for memcached (+whitespace changes)
Randy Syring
added to docs on the importance of calling session.save() before start_response() refs #45 placement of session.save() in the middleware stack matters
Randy Syring
added Session.regenerate_id() refs #75 - Session.regenerate_id() needed to help avoid session hijacking
Randy Syring
whitespace changes
Randy Syring
whitespace changes
Randy Syring
setup.py: fix os not imported
Randy Syring
ignore komodo project file
Randy Syring
setup.py whitespace changes
zzzeek
* file backend no longer squashes unpickling errors. This was inconsistent behavior versus all the other backends. * invalidate_corrupt flag on Session now emits a warning. (#52)
zzzeek
* propose that we go to 1.6 on this one since we're on a roll * cache decorators @cache.cache(), @cache_region() won't include first argument named 'self' or 'cls' as part of the cache key. This allows reasonably safe usage for methods as well as functions. (#55) * factor down cache decorators + invalidate logic into two private functions called by both CacheManager and the standalone functions. * remove unused cache_region() body that somehow was …
zzzeek
* Fixed bug whereby CacheManager.invalidate wouldn't work for a function decorated by cache.cache(). (#61)
zzzeek
* Added an optimization to the FileNamespaceContainer when used with Session, such that the pickled contents of the file are not read a second time when session.save() is called. (#64)
zzzeek
* Basic container/file-based Session support working in Py3K. (#72) * Further Python 3 fixes * tests which require webetest are now all qualified since we don't have webtest for py3k yet * don't really need webob for session tests
zzzeek
dont need webob dependency here
zzzeek
- reorganize the memcached imports again so that a specific module name can be specified - add 'memcache_module' param for memcached backend, allows specification of specific library - Sphinx 1.0 seems to use pathto("py-modindex") and not pathto("modindex") - update copyright in sphinx docs - add docstrings and such to NamespaceManager and others, though this is an uphill battle...this is a really old API
zzzeek
- coverage by default is annoying - add a memcached ping to the start of test_memcached so no failures if memcached isn't running
w31rd0
Added tests for cookies
w31rd0
Added some basic tests for Session object
zzzeek
- remove all trailing whitespace
zzzeek
merge tip
zzzeek
add py3k to the programming langs
Ben Bangert
merge
Ben Bangert
Update recorded version.
Yang Zhang
Use UTC timezone to calculate expiry
zzzeek
* memcached backend uses pylibmc.ThreadMappedPool to ensure thread-local usage of pylibmc when that library is installed.
zzzeek
* Defer running of pkg_resources to look for external cache modules until requested. #66 Unfortunately, a simple test against a "package" is not conceived here. Would be best if an entrypoint were artificially shoved into pkg_resources. and they say my code is hard to follow...
  1. Prev
  2. Next