Re: CERN httpd and libwww 2.14 released

robm@ncsa.uiuc.edu (Rob McCool)
Message-id: <9311162047.AA20703@void.ncsa.uiuc.edu>
From: robm@ncsa.uiuc.edu (Rob McCool)
Date: Tue, 16 Nov 1993 14:47:36 -0600
In-Reply-To: George Phillips <phillips@cs.ubc.ca>
       "Re: CERN httpd and libwww 2.14 released" (Nov 16, 11:11am)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: George Phillips <phillips@cs.ubc.ca>, <luotonen@ptsun00.cern.ch>
Subject: Re: CERN httpd and libwww 2.14 released
Cc: <www-talk@nxoc01.cern.ch>
/*
 * Re: CERN httpd and libwww 2.14 released  by George Phillips (phillips@cs.ubc.ca)
 *    written on Nov 16, 11:11am.
 *
 * Ari Luotonen says:
 * > |		- currently is also compatible with NCSA server with
 * > |		  the exception of parameter unescaping
 * 
 * I wouldn't call it compatible if they differ on something so basic
 * as whether the arguments are unescaped or not.  There's a bigger
 * problem though -- NCSA's script interface is still a moving target.
 * It's not going to move far and it will be making some important
 * changes, but if CERN's script interface is frozen, it'll have
 * a hard time changing to be compatible.
 * 
 * I think this underscores why we really need something like a
 * SCRIPT_PROTO environment variable so that portable (between servers)
 * gateway scripts can be written.

I have a created an env. variable called "SERVER_SOFTWARE" which contains
NCSA/1.0a6 right now. Is this what you had in mind?

 * While I'm whining, it occurs to me that it'd be useful for a script
 * to have some way of knowing what method it is doing.  NCSA HTTPD
 * seems to cover this now by having different GET and POST scripts.
 * This is workable, but tossing in something to let the script
 * explicitly know (environment variable?) would be nice and would
 * make the external gateway scripts easier to run under the x-exec:
 * URL scheme.

I have created an env. variable called $SCRIPT_METHOD which contains either
"GET" or "POST".

 * I hate to dote on this, but the ability to share stuff between servers
 * is important.  Even if the server's don't exactly agree, it would
 * be nice if either portable scripts are possible or if Plexus can
 * support every script interface under the sun.
 */

It is important. I'd like to see a uniform script interface as well. 

--Rob