Beaker 1.4

Beaker 1.4 functions as a helpful library for caching and sessions for use with web applications and stand-alone Python scripts and applications. It comes with WSGI middleware  for easy drop-in use with WSGI  based web applications, and caching decorators for ease of use with any Python based application.

Major Features:

1. Lazy-Loading Sessions: No performance hit for having sessions active in a request unless they're actually used
2. Performance: Utilizes a multiple-reader / single-writer locking system to prevent the Dog Pile effect when caching.
3. Mulitple Back-ends: File-based, DBM files, memcached, memory, and database (via SQLAlchemy) back-ends available for sessions and caching
4. Cookie-based Sessions: SHA-1 signatures with optional AES encryption for client-side cookie-based session storage
5. Flexible Caching: Data can be cached per function to different back-ends, with different expirations, and different keys
6. Extensible Back-ends: Add more backends using setuptools entrypoints to support new back-ends.

• Fix bug with hmac on Python 2.4. Patch from toshio, closes ticket #2133 from the TurboGears2 Trac.
• Fix bug with occasional ValueError from FileNamespaceManager.do_open. Fixes #10.
• Fixed bug with session files being saved despite being new and not saved.
• Fixed bug with CacheMiddleware overwriting configuration with default arguments despite prior setting.
• Fixed bug with SyntaxError not being caught properly in entry point discovery.
• Changed to using BlobProperty for Google Datastore.
• Added domain/path properties to the session. This allows one to dynamically set the cookie’s domain and/or path on the fly, which will then be set on the cookie for the session.
• Added support for cookie-based sessions in Jython via the JCE (Java Cryptography Extensions). Patch from Alex Grönholm.
• Update Beaker database extensions to work with SQLAlchemy 0.6 PostgreSQL, and Jython.