59bea8ad59
To calculate the hash for a compress block, get_offline_hexdigest() calls smart_str on each template node, which resulted in the Python repr of the templates node which is not a text node.
This caused problems when using the {% static %} tag inside a compress block, since the {% static %} repr looks something like this:
>>> template.Template('{% load static %}{% static "foo" %}').nodelist
[<django.template.defaulttags.LoadNode at 0x10e747510>, <django.template.base.SimpleNode at 0x10e747490>]
The repr changes on every compilation (since the memory location changes), which caused the checksum to change every time, and thus made the offline mode unusable.
This patch renders the blocks completely, and calculates the checksums on the entire rendered block.
Django Compressor
Django Compressor combines and compresses linked and inline
Javascript or CSS in a Django templates into cacheable static files by
using the compress template tag.
HTML in between {% compress js/css %} and
{% endcompress %} is parsed and searched for CSS or JS.
These styles and scripts are subsequently processed with optional,
configurable compilers and filters.
The default filter for CSS rewrites paths to static files to be
absolute and adds a cache busting timestamp. For Javascript the default
filter compresses it using jsmin.
As the final result the template tag outputs a
<script> or <link> tag pointing to
the optimized file. These files are stored inside a folder and given an
unique name based on their content. Alternatively it can also return the
resulting content to the original template directly.
Since the file name is dependend on the content these files can be given a far future expiration date without worrying about stale browser caches.
The concatenation and compressing process can also be jump started
outside of the request/response cycle by using the Django management
command manage.py compress.
Configurability & Extendibility
Django Compressor is highly configurable and extendible. The HTML parsing is done using lxml or if it's not available Python's built-in HTMLParser by default. As an alternative Django Compressor provides a BeautifulSoup and a html5lib based parser, as well as an abstract base class that makes it easy to write a custom parser.
Django Compressor also comes with built-in support for CSS Tidy, YUI CSS and JS compressor, the Google's Closure Compiler, a Python port of Douglas Crockford's JSmin, a Python port of the YUI CSS Compressor cssmin and a filter to convert (some) images into data URIs.
If your setup requires a different compressor or other post-processing tool it will be fairly easy to implement a custom filter. Simply extend from one of the available base classes.
More documentation about the usage and settings of Django Compressor can be found on django_compressor.readthedocs.org.
The source code for Django Compressor can be found and contributed to on github.com/jezdez/django_compressor. There you can also file tickets.
The in-development
version of Django Compressor can be installed with
pip install django_compressor==dev or
easy_install django_compressor==dev.