Re: fastCGI for Jigsaw?

Steve Harris (
Wed, 27 Nov 1996 09:40:27 -0500

Message-Id: <>
Subject: Re: fastCGI for Jigsaw? 
Date: Wed, 27 Nov 1996 09:40:27 -0500
From: Steve Harris <>

> Steve Harris writes:
>  |
>  |I haven't looked at Jigsaw, though I have intended to for a while. As for the 
>  |fcgi java code, one caveat - it more or less reflects the state of fast-cgi as 
>  |of last you should probably talk to someone else here about using it 
>  |(I'm no longer in that group).
>  |
> Ok, who?
This list is being monitored, so hopefully someone will get back to you on 
this issue. From my point of view, the issue is what has been changed/added to 
fast-cgi since the java classes were written. Since you have the Java source, 
you might want to include some of these features, but this would have to be 
worked out with whoever is currently suporting fast-cgi. I will look into 
this, and get back to you next week.

>  |As far as the  classes go, you would probably need Message, Request, and 
>  |GlobalDefs, to go along with InputStream and OutputStream. You probably would 
>  |not need the Interface class. But then I'm not sure what you would be doing 
>  |with these clases.
> Well, Jigsaw is written in Java.  I was thinking that I would use the
> *.java code and incorporate them into a resource that Jigsaw uses to
> launch/dispatch the fcgi request to the fcgi application.  Sort of
> take the place of the cgi-fcgi gateway, or the fcgi.c code in the NCSA
> server.
Yes, it would seem that you would need the classes I mentioned, though not all 
of the methods in Message would be used. One thing that might be a bit 
confusing is that there is a single static Request object in the Interface 
class. This was done to maintain compatiblity with the C code at that time, 
but I'm pretty sure (confirmed by a quick grep) that I coded all the other 
classes not to refer to this static request directly, but to take it as a 
parameter, anticipating a point when fasg-cgi would be able to handle multiple 
requests concurrently (through threading or some other mechanism). 
> Thanks for getting back sooo fast.  You guys do a great job!
Thanks for your comment. I'd certainly like to see what you come up with.
> -- 
> Brian Millett                    
> Technology Applications Inc.     "Heaven can not exist,
> (314) 530-1981                          If the family is not eternal"
>                   F. Ballard Washburn