httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Slemko <>
Subject protocol/2602: GET /cgi-bin/cginame%3Fparam=/path/filename does not replace %3f before searching cginame?param... (fwd)
Date Thu, 16 Jul 1998 03:50:07 GMT
Isn't the current behaviour correct?

Isn't ? one of the special characters that have different meanings and
must not be encoded if being used for their special reserved meanings but
must be encoded if they aren't being used for their reserved value?

---------- Forwarded message ----------
Date: 13 Jul 1998 08:35:01 -0000
From: Andrei Petrescu <>
Subject: protocol/2602: GET /cgi-bin/cginame%3Fparam=/path/filename  does not replace %3f
before searching cginame?param...

>Number:         2602
>Category:       protocol
>Synopsis:       GET /cgi-bin/cginame%3Fparam=/path/filename  does not replace %3f before
searching cginame?param...
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    apache
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Mon Jul 13 01:40:01 PDT 1998
>Release:        1.2.1
It's not an os problem, it's in the URL parser of apache (the %hh stuff for
quoting special hex characters; HTTP standard says that URL's containing
%hh quoted characters should be equivalent to the ones containing the characters
If some browser requests a CGI using %3F instead of '?',Apache will parse
the URL incorrectly and mistakenly search for a file containing '?' 
(e.g. cginame?xxx=yyy/zzz/ttt)

It is true that most browsers never use %3f with a cgi, but I need this
in order to enable windoze clients to download generated files (otherwise
windows will goof up the filename putting the cginame and a ? instead).
So, I modified another cgi to yield url's like
so that netscape for windoze will use file.ext instead of cginame? as the
filename for d/l. But now apache says 404 not found, and a system trace 
says apache looks for cginame?filename=/path/file.ext and gets ENOENT.
Did you fix this problem in 1.6.* ?
If not, can you fix it ?
Try to replace ? by %3f in a cgi name and load it like that, it will
When parsing an URL, replace the %xx characters before anything else, especially before searching

the cgi file.
[In order for any reply to be added to the PR database, ]
[you need to include <apbugs@Apache.Org> in the Cc line ]
[and leave the subject line UNCHANGED.  This is not done]
[automatically because of the potential for mail loops. ]

View raw message