Would it be worthwhile to add that the user can run "open $htmlfile_path" to examine the download? In the case we think we know what's wrong (HTML was downloaded), it could also be handy to print out the originating mirror rather than have people go through debug log.