I finally found some time to look at the archiver code for Mailman. I made a small fix to it long ago and Barry never let me forget :-).
I fixed the archiver so that it generates useful HTML pages for mail messages that use non-ASCII character sets, either in the message body or in the headers (using encoded-words). It looks like it's working; I see lots of non-Roman characters where I used to see line noise. I also spent a lot of time refactoring the code; it had far too many long methods, including several that were almost line-for-line duplicates of the same method in a superclass.
