index.html being added to CGI path info

Hi,

I was using fnord and I noticed that fnord's logic to append index.html 
to a directory seems to do it even when the directory is part of the 
path info of a CGI program. For instance, if I browse to 
http://foo.com/viewcvs.cgi/Repos1/, fnord passes /Respos1/index.html to 
the CGI script, which then complains that it can't find a repository there.

Is this considered correct behavior?