Interesting. I think this particular piece of the implementation of URI::Escape is just a little too fragile. I had problems until I doubled up the backslashes. That was with version 3.13 of URI::Escape. Then, I upgraded to the current 3.20 version and find that it broken again: only the backslash for \w needs to be doubled, the one for slash has to be single. You probably have some other version for which it's broken in yet a different way.
Here's a tested version -- written as a bash script. It is tested against Revision 3.20 of URI::Escape. #! /bin/bash perl -MHTML::Entities -MURI::Escape -ne \ 'chomp;print qq{<A href="http://host/}, scalar uri_escape( $_, "^\\w-.\/"), qq{">}, encode_entities($_), qq{</A>\n};' I had to put back a "^" that had been omitted in what you said you used. The name "host" is just a placeholder, by the way, you'll need to change that. And there's also the problem of DocumentRoot not usually being the same as / that you may need to do something about. ------------------------------------------------ Get the award winning ISP, AT&T WorldNet Service http://download.att.net/webtag ***************************************************************** To unsubscribe from this list, send mail to [EMAIL PROTECTED] with the text 'unsubscribe gnhlug' in the message body. *****************************************************************