Jamey helped me debug a PyBlosxom plugin problem this morning. The Folksonomy plugin was throwing errors when it was asked to generate an RSS feed for a specific tag (e.g. http://sarah.thesharps.us/tags/pyblosxom/index.rss).

The lighttpd error log showed this:

Traceback (most recent call last):
  File "/usr/lib/cgi-bin/pyblosxom.cgi", line 96, in ?
    p.run()
  File "/var/lib/python-support/python2.4/Pyblosxom/pyblosxom.py",
    line 203, in run
    blosxom_handler(self._request)
  File "/var/lib/python-support/python2.4/Pyblosxom/pyblosxom.py",
    line 984, in blosxom_handler
    defaultfunc=blosxom_file_list_handler)
  File "/var/lib/python-support/python2.4/Pyblosxom/tools.py",
    line 759, in run_callback
    output = func(input)
  File "/home/sarah/blog/plugins/folksonomy.py",
    line 482, in cb_filelist
    return getEntriesForTag( tag, args )
  File "/home/sarah/blog/plugins/folksonomy.py",
    line 492, in getEntriesForTag
    for entry_location in entrymap[tag]:
KeyError: 'open source/index'

The code that generates the string to search the tag entry list wasn't stripping off the "/index" part of the URL. The patch Jamey wrote is here. I should send in the patch to the author too...