W3C home > Mailing lists > Public > www-jigsaw@w3.org > May to June 1996

Re: OLD CGI CALLS

From: Adam R. B. Jack <ajack@neonsoft.com>
Date: Sun, 23 Jun 1996 12:46:24 -0600 (MDT)
To: Anselm Baird-Smith <abaird@w3.org>
Cc: www-jigsaw@w3.org
Message-Id: <Pine.SOL.3.93.960623114502.12143A-100000@neondev1>
On Sun, 23 Jun 1996, Anselm Baird-Smith wrote:

> Does this sounds like something you possibly done ?
> If so, you need to 'reindex' - by using the parent directory resource
> reindexing mechanism - the already indexed CGI scripts..
> 
Indeed. I guess it is jigsaw newbie mistake #1. Thanks for taking time on a
Sunday to put me straight.

> Again, the CgiResource has some oversight - like the one you mention -
> If any of you sends me a diff for the env var settings, etc I will be
> glad to incorporate it. In particular I know some of the HTTP_* (if I
> recall) are missing.

The following five lines constituts the diff output for this change.
228c228
< 	addEnv ("SERVER_NAME", server.getHost(), env) ;
---
> 	addEnv ("SERVER_NAME", "jigsaw", env) ;

If I run accross any more I will email them.

------------------------------------------------------------------------------

I realize that Jigsaw is an attempt to explore new mechanisms, and I do 
apologize for continuing w/ CGI questions, however I wish to use Jigsaw as 
my only server and want to understand as much about it, as possible.

I would like to explain a low-priority situation in the hope that you can
again clarify the situation for me. Maybe there exists a Jigsaw solution
to this problem. I am in no way attempting to say this is a limitation of
Jigsaw -- so if no solution exists please freely say so.

The CGI software I am using is an HTML preprocessor for file with a certain 
extn. It parses the HTML for special 'extension' tags and replaces them with 
dynamically generated HTML output as appropriate. Sometimes it find no tags 
to replace and returns a local redirect to the file. (Note: Then why have
this extension on the file? Dont ask me! Lame I know -- hence the low
priority.) Anyway, such redirects would, on a normal server, be:

	..../cgi-bin/BLAH?/blah_dir/blah.XXX redirected to
	..../blah_dir/blah.XXX

Obviously -- I find that I can no longer depend upon extension XXX to determine
if a resource should be a CgiResource. (It forces the browser into a potentially
infinite redirect loop, which Lynx, and I assume the HTTP protocol, is too 
smart for.)

So, and maybe this is an example of thinking 'the old server way', I consider
setting a resource BLAH in cgi-bin and referring to the URLs :

	..../cgi-bin/BLAH?/blah_dir/blah.XXX

Even though I remove the XXX extension rule, and attempt to reindex all that
I might consider affected, I can not access URLs like this. Adding a Debug-
Filter I see a difference betweening attempting to access BLAH and BLAH?...
the latter not even saying 'should dump response', which I take to mean
that lookup failed.  (For what it is worth, I also tried adding a resource 
BLAH?/blah_dir/blah.XXX with even less success.)

Whenever time is available I would greatly appreciate it if you could set me
straight once more.

Many thanks,

Adam
--
NEON Inc. 7400 East Orchard Road, Suite 230, Englewood, CO 80111 +1-303-6943933 
Received on Sunday, 23 June 1996 14:50:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 9 April 2012 12:13:25 GMT