Infinite loop when reading "file:" HTML

I'm experiencing a problem with libwww when it reads files locally
rather than fetching them via a socket.  This problem affects even
the webbot program, so it is not unique to the code that I am
writing.

I searched the mailing list and discovered that Kinuko Yasuda
reported the same problem back in mid January, 2000, and he
even provided a fix:

  http://lists.w3.org/Archives/Public/www-lib/2000JanMar/0025.html

This fix works fine for me, so I'm curious why it hasn't been
incorporated into the latest CVS source.  Is there a better fix
that someone is working on?

If you want to reproduce the problem, just run webbot on some local
files.  I did something like this:

webbot -depth 200 -prefix file:/home/wayne/path file:/home/wayne/path/foo

(Keep in mind that "foo" links to several other pages.)

..wayne..