Just a second ACK. From studying the debugging docs in the 2nd reply, it appears that I was doing the correct thing, and yet it does not work. From running with the -w parameter I found a trivial mistake to fix. From running with the -d parameter, I received the mixed blessing of reams of unknown data. Time for meditation.

A fun approach; maybe overkill here. It’s multi-arg system that avoids the shell, not capture.

Well sure :) but if you don't capture it, it just goes into the regular place, at which point, the question becomes, why not do a redirect instead?

Oh right, the OP didn't ask that exactly, that was just the question that brought me to this question (from a cgi execute another cgi) ... looked close enough :)

I am doing some sort of redirect, in that the HTML output of the original PERL file is properly captured into a file that I can then load in a browser.

I think I am skipping over some extremely low level of ignorance here. The answer is doubtless "intuitively obvious to the most casual observer", but I am evidently too tense?

I am doing some sort of redirect, in that the HTML output of the original PERL file is properly captured into a file that I can then load in a browser.

I think I am skipping over some extremely low level of ignorance here. The answer is doubtless "intuitively obvious to the most casual observer", but I am evidently too tense?

So have you considered installing a webserver to run your "cgi" ?

apacheet (Plack::App::CGIBin),