Here’s a simple starter for hacking on the Feed Validator, either to make fixes, or simply to use it for non-public feeds.
(This is for Linux – tips for Windows or other environments are welcome!)
Get the code
The Feed Validator doesn’t make official releases,
but it’s regularly updated and a current snapshot is almost always stable.
Choose a local directory to store the validator in – say,
Download a snapshot
Snapshots are available from https://github.com/rubys/feedvalidator/archives/master.
Or, use Git
Git is slightly more complicated to get working, but then you can keep current.
git clone git://github.com/rubys/feedvalidator.git
Later, to get the latest code, call
git pull from inside the
feedvalidator directory. See the
Git documentation for more information.
Run it from the command line
There's a command-line interface:
python src/demo.py <feed url>
(Or, hack that script to work how you want.)
Set it up for your web server
First install a web server, Python and additional character encoding support. Under Debian:
apt-get install apache2-mpm-worker python2.3 python2.3-iconvcodec
For simplicity, it’s easiest to run the validator out of a regular user account. That means you don't need to be root.
To do this, simply unpack or check out the Feed Validator into your
public_html directory. Now open a browser to
and we’re, hopefully, done.
Do you have special requirements?
If you have special requirements for things like where directories are to be placed, you may want to tweak the config.py file. For example, if you want to separate the code, the cgi, the docs, and css files, you copy/link to only the files you need:
FVDIR=<location of the feedvalidator directory> cd ~/public_html ln -s "$FVDIR/css" fv-css ln -s "$FVDIR/docs" fv-docs mkdir -p cgi-bin/fv cp "$FVDIR/check.cgi" "$FVDIR/config.py" cgi-bin/fv/
config.py to point to the other directories:
WEBDIR = '/'.join(os.environ['SCRIPT_FILENAME'].split('/')[0:-1]) DOCSURL='docs' CSSURL='css'
Now open a browser to
I don’t think it’s any disrespect to say that that the test cases are the Validator’s most valuable asset – currently over fourteen hundred separate cases of valid, and invalid, feeds, along with the precise errors. When you find a bug, or any other misfeature, a simple test case is the best way to get your point across.
To check the existing cases, run
This will check everything below the ‘testcases’ directory (you
can also specify a single test by filename).
Not clear? Suggestions or improvements?
Let us know on the feedvalidator-users discussion list!