W3C home > Mailing lists > Public > www-jigsaw@w3.org > July to August 1996

Re: CGI problems

From: Remy AMOUROUX <Remy.Amouroux@imag.fr>
Date: Fri, 9 Aug 1996 15:25:54 +0200
Message-Id: <199608091325.PAA20198@elronde.inrialpes.fr>
To: www-jigsaw@w3.org

Hello

>>>>> "AB" == Anselm Baird-Smith <abaird@w3.org> writes:

AB> I have been working on the CgiResource, preparing it for the next
AB> release. I know lots of people had trouble with it.

Maybe I come a bit late, but here is one exemple of trouble.

	Doing some experiences with jigsaw, I tried to use the
CgiResource to access a database through the CGI given by the
vendor. The archive of the mailing-list gave me the hints to do so and
the creation of a CgiResource for the well-known test-cgi shell script
has been done without any problem. I made the same with the database
access cgi program. But I met a problem which I was not able to
solve.

I illustrate the problem with test-cgi :

1) Using the following url, with an http server like apache, 
	http://mysite/cgi-bin/test-cgi/truc/muche
   the test-cgi is executed and you will have
	 PATH_INFO = /truc/muche

   The same url used with the jigsaw server gives :
	<h1>Document not found<h1>
	Document /cgi-bin/test-cgi/truc/muche not found on this server

So, the script is _never executed_ on a jigsaw server. Maybe, I'm
doing something wrong but I don't know what !

   The problem doesn't come from the CgiResource, I think the code of
the class is never called in this case.

   I tried to define a resource named /cgi-bin/test-cgi/truc/muche,
but it's not possible using the Editor interface. Surely because my
CgiResource test-cgi is a FileResource and not a DirectoryResource.

2) The previous problem made me look at the output of the test-cgi
script :
	when you request 
		http://mysite/cgi-bin/test-cgi
	to the jigsaw server, you have :
		PATH_INFO = /cgi-bin/test-cgi
	But doing the same with another server (apache for exemple),
you will have :
		PATH_INFO = /
	I had a look at the CgiResource code, the value of PATH_INFO
is set with a call to request.getURI(). 
	
	It's a small thing but we all have to deal with legacy
application and I think the problem will arise very often.

	the CGI/1.1 interface specification gives the following
definition :
	"PATH_INFO=The extra path information, as given by the
client. In other words, scripts can be accessed by their virtual
pathname, followed by extra information at the end of this path. The
extra information is sent as PATH_INFO. This information should be
decoded by the server if it comes from a URL before it is passed to
the CGI script" 

	So, is it choice in the jigsaw design to disallow this feature ?

AB> The most visible improvment I did to this resource, is to default
AB> command[0] (the script path), to the resource path, which means that
AB> you can now use the resource indexer, to declare all ".pl" files to be
AB> exported by a CgiResource. 

A good idea.

Rémy
Received on Friday, 9 August 1996 09:31:09 GMT

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