Re: Re: FCGI application testing

What about automating the web client piece for testing rather than looking
at the FCGI piece?  You could use some of the Perl modules out there or
a command-line client like lynx or wget to send in the requests and save
whatever return values are necessary between subsequent requests.

Steve

On Thu, 30 Sep 2004, Michael Hartley wrote:

> Hi,
>
> Nice idea's but my fcgi application server is long running and needs to
> maintain state between session calls, so calling from the command line is
> unlikely to solve my problems.
>
> I'm up for the idea of a two way socket communication with the FCGI server
> but my stress test application would have to mimic the equivalent exchange
> protocol of the FCGI server and the FCGI webserver modules.
>
> I suspect a few technically brilliant pearls of wisdom from RobS or
> somebody with equivalent experience would be very gratefully received here.
>
> Somebody must have suffered the same needs to build a rigorous test rig
> against an FCGI application before?
>
>
> Cheers,
>
> Michael
>
>
>
>
> On Thu, 30 Sep 2004 08:11:00 -0400 (EDT), Steve Ferguson
> <[email protected]> wrote:
>
> > You could write it so that it can still function as a "regular" CGI,
> > which
> > isn't much work, at which point you could just feed it inputs on STDIN
> > and get the results on STDOUT/STDERR by running it on the command lnie.
> >
> > Or you could set up your app as an external FCGI server, then just open
> > sockets to it directly for testing.  Granted, this means you have to add
> > your own code to deal with listening on the socket and handling requests,
> > so this may be using a sledge hammer to crack a walnut.
> >
> > Steve
> >
> > On Thu, 30 Sep 2004, Michael Hartley wrote:
> >
> >>
> >> Does anyone have any suggestions for testing FCGI applications without
> >> making HTTP calls via a web server?
> >>
> >> I would like to generate test plans that call the FCGI application
> >> directly passing a fully formed FCGI stream to the application server
> >> and
> >> reading back the output with the destination probably being a file.
> >>
> >> Failing the possibility of an FCGI application test mechanism, does can
> >> anyone recommendations for executing reproducible test scripts against
> >> an
> >> FCGI application?
> >>
> >> Any comments, thoughts, suggestions or radical idea's would be gladly
> >> received!
> >>
> >> Kind regards,
> >>
> >> Michael Hartley
> >>
> >>
> >
>
>
>
>

-- 
Steve Ferguson                        | "Men never do evil so completely
Email: [email protected]        | and cheerfully as when they do
http://dogbert.dnsalias.org:8080/~stf | it from a religious conviction."
                                      |   --Blaise Pascal

___________________________________
fastcgi-developers mailing list
http://fastcgi.com/fastcgi-developers/