From: Rob Hartill <hartill@hyperreal.com>
Message-Id: <199602170707.XAA21948@taz.hyperreal.com>
Subject: WWW Form Bug Report: "access log has entries with NULL method when
 read_request() gets a premature EOF" on SunOS 4.x (fwd)
To: new-httpd@hyperreal.com
Date: Fri, 16 Feb 1996 23:07:28 -0800 (PST)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1350      
Sender: owner-new-httpd@apache.org
Precedence: bulk
Reply-To: new-httpd@hyperreal.com


no ack

Forwarded message:
> From nobody@hyperreal.com  Fri Feb 16 14:54:05 1996
> Message-Id: <199602162254.OAA14109@taz.hyperreal.com>
> From: chrisd@name.net
> To: apache-bugs%apache.org@organic.com
> Date: Fri Feb 16 14:54:02 1996
> Subject: WWW Form Bug Report: "access log has entries with NULL method when read_request() gets a premature EOF" on SunOS 4.x
> 
> Submitter: chrisd@name.net
> Operating system: SunOS 4.x, version: 
> Extra Modules used: mod_log_common.c
> URL exhibiting problem: 
> 
> Symptoms:
> --
> Bug's a strong word.
> 
> I notice that entires in the access log appear
> with a "NULL" access method when the error log
> reports a "read timed out" error.  I have traced
> this to (I think) the behaviour of read_request()
> when getline()'s call to fgets() returns a NULL
> pointer.  This leaves the timeout still set and
> because r->method never gets set, it is filled
> in by log_transaction() as NULL.
> 
> I've checked around and this doesn't to be
> part of the Common Log Format spec, although it
> might be.  I only noticed the problem because
> it produces odd effects in our log analyzer.
> I'd like to know if this is a deliberate
> choice, and if so, I'll tweak the analyzer.
> 
> Lovely server, BTW.
> 
> Thanks,
> 
> Chris Darroch
> NAME Webmaster
> 416-943-1010x230
> 
> --
> 
> Backtrace:
> --
> 
> --
>