httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Laurie <>
Subject Re: Configurable nph filenames
Date Thu, 11 Jul 1996 17:22:09 GMT
Alexei Kosut wrote:
> On Thu, 11 Jul 1996, Rob Hartill wrote:
> >  
> > > Hmm, I think we already decided once not to have this, didn't we?
> > 
> > do you remember why?  It seems reasonable to me.
> The main reason is that nph scripts are a Bad Idea, it's generally agreed.
> Because they talk directly to the client, you can't do any header parsing
> (which, btw, can make the response an illegal one - I wonder how many nph-
> scripts deal correctly with HEAD, for example. HTTP/1.1 makes this problem
> about twenty times worse, because of the persistent connection stuff), you
> can't do keep-alive, you can't do chunking, and gosh himself probably
> couldn't help us if we tried to make nph scripts work with HTTP-NG (should
> it ever actually exist).
> Usually when we talk about this, we decide it'd be a good idea if CGI
> scripts could send a header that let them deactivate buffering of the
> response, thus allow most things people want to do with nph- scripts
> (server push and whatnot), but still making it a server-parsable response.
> But no one's written it yet.

FWIW I'll probably do this soon. I need this facility for some slow CGIs I've
got, and I've sworn not to do NPH. Unless someone else wants to do it (he says

Does anyone have any suggestions for the name of the header?



> -- Alexei Kosut <>            The Apache HTTP Server 

Ben Laurie                  Phone: +44 (181) 994 6435
Freelance Consultant and    Fax:   +44 (181) 994 6472
Technical Director          Email:
A.L. Digital Ltd,           URL:
London, England.

View raw message