Received: (qmail 79208 invoked by uid 501); 14 Jan 2001 21:41:54 -0000
Message-Id: <20010114214154.79207.qmail@apache.org>
Date: 14 Jan 2001 21:41:54 -0000
From: James Marshall <james@jmarshall.com>
Reply-To: james@jmarshall.com
To: submit@bugz.apache.org
Subject: PATH_INFO containing %2F causes server to interpret PATH_INFO as part of file path.
X-Send-Pr-Version: 3.110

>Number:         7066
>Category:       protocol
>Synopsis:       PATH_INFO containing %2F causes server to interpret PATH_INFO as part of file path.
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    apache
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Sun Jan 14 13:50:01 PST 2001
>Closed-Date:    Fri Jun 13 15:02:44 PDT 2003
>Last-Modified:  Fri Jun 13 15:02:44 PDT 2003
>Originator:     james@jmarshall.com
>Release:        1.3.15-dev
>Organization:
>Environment:
It's actually on www.apache.org; printenv.cgi shows it.
>Description:
Compare

http://www.apache.org/printenv.cgi/path/info
http://www.apache.org/printenv.cgi/path%2Finfo

These should both call printenv.cgi, right?  The second case results in
"file not found", apparently because the server is looking for the file
/printenv.cgi/path/info .

Related to this, PATH_INFO is being un-URL-encoded before being passed to the
script, e.g.

http://www.apache.org/printenv.cgi/path%20info

results in PATH_INFO of "/path info".  Shouldn't it be passed along without
decoding it, i.e. "/path%20info"?  Otherwise, how could a script distinguish
between PATH_INFOs of "/path%2Finfo" and "/path/info"?  Or are such pairs
always semantically equivalent?

Sticky question and a can of worms, I know.  When are parts of the URL
encoded and when aren't they, etc.  The latest CGI draft seems to imply
that PATH_INFO treatment may vary by server (section 6.1.6), but this
results in non-portable scripts.  Granted, the CGI draft is a BCP.
>How-To-Repeat:
Try http://www.apache.org/printenv.cgi/path%2Finfo
>Fix:
What seems to be lacking is a clear definition of when things are encoded and 
when they aren't.  I know that's very tricky to come up with.  Yes, I've lost
sleep trying to figure this one out.
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open-closed
State-Changed-By: coar
State-Changed-When: Fri Jun 13 15:02:44 PDT 2003
State-Changed-Why:
This has been addressed in Apache 2.0.46 through the AllowEncodedSlashes directive.
Category-Changed-From-To: mod_cgi-protocol
Category-Changed-By: coar
Category-Changed-When: Fri Jun 13 15:02:44 PDT 2003

>Unformatted:
 [In order for any reply to be added to the PR database, you need]
 [to include <apbugs@Apache.Org> in the Cc line and make sure the]
 [subject line starts with the report component and number, with ]
 [or without any 'Re:' prefixes (such as "general/1098:" or      ]
 ["Re: general/1098:").  If the subject doesn't match this       ]
 [pattern, your message will be misfiled and ignored.  The       ]
 ["apbugs" address is not added to the Cc line of messages from  ]
 [the database automatically because of the potential for mail   ]
 [loops.  If you do not include this Cc, your reply may be ig-   ]
 [nored unless you are responding to an explicit request from a  ]
 [developer.  Reply only with text; DO NOT SEND ATTACHMENTS!     ]