Re: accept(), solaris 2.6, devkit1.5.1
Stanley Gambarin (stanleyg@cs.bu.edu)
Sat, 13 Sep 1997 02:56:33 -0400 (EDT)
Date: Sat, 13 Sep 1997 02:56:33 -0400 (EDT)
From: Stanley Gambarin <stanleyg@cs.bu.edu>
To: suit <zoot@jobtrak.com>
Subject: Re: accept(), solaris 2.6, devkit1.5.1
In-Reply-To: <199709121053.DAA16330@jobtrak.com>
Message-Id: <Pine.GSO.3.95q.970913025442.25770A-100000@csa>
The assert failure you are refering to was fixed in the
2.0b1 mod_fastcgi module. If my memories serves me
correct, the fastcgi header was not casted to char *,
resulting in the error. Get the newest version of the
mod_fastcgi, grep for char * and apply the change to your
version of the module.
Stanley.
On Fri, 12 Sep 1997, suit wrote:
>
> for those wanting to use the devkit without accept() lock patches under solaris
> it does work under the following conditions:
>
> solaris 2.6 - [ looks like they fixed? their accept call ]
> apache 1.2.4 - [ pre 1.2.3 was hurting under solaris 2.6 ]
> fastcgi-devkit 1.5.1 - [ no patches applied ]
> apache-fastcgi module 1.4.3 - [ with duplicate typedef removed ]
>
> our peak loads went from 60 to 6 :)
>
> the only remaining wierdness is that the apache-module will ocasionally
> fail on a buffer length assert() - but we haven't been able to reliably
> recreate the situation yet ... *shrug*
>
> thanks for the help along the way,
> -rob
>
> --
>
> <zoot@jobtrak.com>
> "Code slinger / Wax mixer
> Noone can see me ... I am the future."
>
>
>