You are viewing a plain text version of this content. The canonical link for it is here.
Posted to apache-bugdb@apache.org by Rautenbach Keith <Ke...@isltd.insignia.com> on 1997/02/15 04:00:01 UTC

protocol/174: Another HTTP/1.0 in the header problem: Fix for PR#100 not complete.

>Number:         174
>Category:       protocol
>Synopsis:       Another HTTP/1.0 in the header problem: Fix for PR#100 not complete.
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Fri Feb 14 19:00:01 1997
>Originator:     KeithRautenbach@isltd.insignia.com
>Organization:
apache
>Release:        1.2b5
>Environment:
n/a
>Description:
I have applied the fix for PR#100 to 1.2b4 sources (by extracting from the 1.2b6
source tree.

This nearly fixes a problem with stupid browsers (e.g. Sun's HotJava and
JavaWorkShop browsers) that fail to parse a header response if it does
not contain "HTTP/1.0".

Ordinary pages are OK, but if when the server tries to respond with
AUTH_REQUIRED then mod_browser has not been called by the time the header is
prepared so when the error response calls basic_http_header() the table entry
"force-response-1.0" has not yet been set!
>How-To-Repeat:
dbx in basic_http_header() and add 

BrowserMatch JavaWorkShop.1.0 force-response-1.0=True

to your access.conf
>Fix:
Um, not really...
a) Recommend my clients upgrade to new version of their browsers?
   Sun has fixed this in HostJava preBeta2.

b) I suppose I duplicate my document tree under /browser-bug/... and then
   <Directory /usr/local/etc/httpd/docs/browser-bug>
	SetEnv force-response-1.0 True
   </Directory>
%0
>Audit-Trail:
>Unformatted: