|Ned Batchelder : Blog | Code | Text | Site|
» Home : Code
Created 6 November 2011
Caveman is a package that parses and validates HTML5 cache manifests. HTML5 applications are notoriously picky about their cache manifests, and it is difficult to check that all of the components are as they should be.
Caveman provides one command-line script, check_manifest, which pulls an HTML page, scrapes it for used resources, parses its cache manifest, and validates the resources against the manifest:
Problems with the manifest are written to the standard output. More detail about the process is output if you set the log level to a different severity with the --level=LEVEL switch, where LEVEL is DEBUG, INFO, WARNING, or ERROR.
Caveman has been designed to be usable from your own code, for example, as part of a larger validation process, or as part of unit tests in your web application.
Caveman's work is done by the ManifestChecker class. You instantiate it with two helpers: a get_url function that fetches data from URLs, and a logger object that gets logged messages:
Then call its check_manifest method with the URL of the HTML page:
No value is returned: the output has been logged to the logger object. See the docstrings in the source code for details.
As an example of programmatic use, here's one way to use Caveman in a Django test suite:
Caveman only pulls the HTML page you specify. Although it validates links to other HTML pages against the manifest, it does not pull those linked-to pages and verify their resources.
Certain rules from the HTML5 spec are not validated.
Caveman can be downloaded from PyPI: http://pypi.python.org/pypi/caveman.
The HTML5 cache manifest spec is at http://www.w3.org/TR/html5/offline.html.
Code repository and issue tracker are at bitbucket.org.