Given that he's reading one line at a time, this solution says:

If the current line does not contain the string
  href="http://[anything].html
then replace every instance (on this line) of
  href=[anything].html
with
  href=[anything].asp
[download]

If there are multiple links on one line, and any one of them is external, this solution will skip all links on that line. If a link spans more than one line, it will not be handled properly. After looking at this same problem posted here what, 4 times now, I'm thinking maybe the best approach is to use HTML::TokeParser, analyze each link (using the convenient get_tag method), and rewrite the HTML with CGI or something so it's nice and pretty.

--isotope
http://www.skylab.org/~isotope/