This repository has three branches with three different version of the
script.

I've run this twice and I've found that every PyBlosxom instance is a
little different and requires a different (sometimes very different)
version of this script. If you want to use it, you will *need*
to modify it to fit your needs. I've included three version of this
script in three different branches. Please use git to look at the three
to pick and choose what you want.

1. pyblosxom2wxr-snarfed.sh: This is the version available on
   http://snarfed.org/pyblosxom2wxr

   It assumes posts are .txt files which are HTML extension
   and assumes both posts and pages as distinct types.

2. pyblosxom2wxr-revealingerrors.sh: coverted http://revealingerrors.com
   
   This assumes only posts, in .txt files, but which are all in fact
   markdown files. It also takes into account 'tags' using the tags.py
   plugin and assumes that they are the second line of every file
   following a "#tags" line.

   REQUIRES: markdown_py
   
3. pyblosxom2wxr-copyrighteous.sh: converted http://mako.cc/copyrighteous

   This deals with two types of files: .txt files which are standard
   html pyblosxom files as well as .rst files which are restructured
   text. It also looks for comma separated tags in the "#tags " line of
   each file.

   REQUIRES: python-docutils

   HUGE WARNING: This version of the script changes the GUID or the
   permalinks. It does *not* keep the links that worked with your old
   version of PyBlosxom. I done because I usd date based URLS and I
   wanted to get rid of them. The script *also* writes a new file,
   called "url_mapping_list" in the local directory. I use that to set
   up a series of 301 permanant redirects with Apache so that people
   going to the old links still arrive at the page. 

Known Bugs / Workaround
-------------------------

One bug I have not worked around is that comments are created with the
'comment_author' field in the database including either the IP address
or the IP address and time timestamp pre-pended to the name. I think
this might be a bug in the Wordpress import code. Rather than debug it,
I wrote a simple perl script (fix_comment.pl) which will fix the
comments in the database. It's an ugly kluge but it worked for me.
Please read it carefully before running it.