How it works

This technique has been desribed by Adrian Holovaty in this blog post and previously by Honza Kral. The idea is to first render the template with certain blocks denoted as “phased,” such that they will not be rendered, and will remain valid template code that can be rendered with a second pass.

The second pass fetches the partially-rendered template from the cache and performs a second render on it, using RequestContext to provide user-specific context to the template. This enables very fast generation of pages that have user-specific content, by bypassing the need to use the CACHE_MIDDLEWARE_ANONYMOUS_ONLY setting.

This implementation uses a secret delimiter that makes it safe against the possibility of template code injection vulnerabilities, as it only passes any given text through the template parser once. The phased blocks can also contain cached context.

Basic Implementation

django-phased contains a templatetag, phased, which defines blocks that are to be parsed during the second phase. A middleware class, PhasedRenderMiddleware, processes the response to render the parts that were skipped during the first rendering.

A special subclass of UpdateCacheMiddleware that drops the “Vary: Cookie” header from response when it updates the cache is also included, which, if used in place of the standard UpdateCacheMiddleware will prevent the cache middleware from varying the cache key based on cookies, thus enabling caching of pages in authenticated sessions.