httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From r..@ai.mit.edu (Robert S. Thau)
Subject Re: inneresting behavior
Date Tue, 18 Jun 1996 16:04:17 GMT
  This will issue a 302 redirect to the client.  Okay, this is all fine, and
  I believe in the spirit of the CGI spec, but I have to wonder - in the
  first case, how does the client know that the URL for the returned object
  is *different* than the URL in the request?  

Well, if we were a fully HTTP/1.1 compliant server, we would have sent
out a Content-location: header.  However, we aren't (and since there
aren't any 1.1 clients out there, it wouldn't necessarily matter much
if we were).

The bottom line, for the moment, is that if you want relative URLs to
work in the target of a redirect, you've got to return a 302 and have
the redirect done in the client, with the associated performance hit.

If relative URLs are not a problem, then the internal redirect does
provide a useful performance boost, but for the moment at least, you've
got to be aware of its limitations...

rst


Mime
View raw message