Spicklemire, Jerry writes:
When I call "REQUEST.redirect(other_url)", the REQUEST isn't destroyed, in fact all the parameters are retained, right down to the original method, so that a "get" method passes parameters appended to the URL, and a "post" method includes all the original REQUEST parameters in the header. The HTTP/1.1 specification is unfortunately very vague with respect to what should be done with the request parameters. Thus, browsers have some freedom what they do about them.
I am unsure what behaviour I would favour: If I have a GET method, I would usually not like the browser to append the old request parameters. For, if I would like them available, I can easily add the QUERY_STRING myself. However, if I have a POST method, I cannot provide the request body in the redirect. It would be natural when the browser would send the original request body just to the new location. I think, the second behaviour is more in line with the notion of redirection: send the same request to a different URL. However, my browser (Netscape 4) implements the first behaviour.
Not what I need. All I want to do is to pipe the output of a page generation
process back to the requestor (a PDF file, in this case), so that it appears
to be a page being returned in response to a get, submitted with no parameters. Why are you unhappy about the parameters. Usually, they do not hurt.
Dieter