bug test before release of 1.2b8

[Aram Mirzadeh <aw...@luers.qosina.com>]

Before we release b8, I would like to change the following item for an 
internal test, and if it causes no problems on any of our OS's then 
release it in b8.  

This is mainly a linux bug: 


In listen() call for the linux networking code it seems that linux truncates
the parameter with an unsigned char and the value of 512 (which we are passing)
it being treated as a 0.

I would like to try to change this number to 511 which would produce a 256
I belive.  The line is in http_main: around 1900...  If you could just 
change it on your own server, compile and do a quick test to see if anything
looks outta wack.  If not I'll make up a patch for the tree. 

Thanks. 

<Aram>

-- 
/*
 * Aram Mirzadeh, MIS Manager, Qosina Corp. http://www.qosina.com/~awm/
 * Apache Development Team, awm@hyperreal.com http://www.hyperreal.com/
 * PGP Key - http://www.qosina.com/~awm/pgpkey.html
 * BE 49 9D F6 2A A7 22 FC  02 E9 1E 3D F7 0C 67 A0
 *
 * I haven't lost my mind -- it's backed up on tape somewhere.
 */

Re: bug test before release of 1.2b8

[Chuck Murcko <ch...@topsail.org>]

Aram Mirzadeh wrote:
> 
> Before we release b8, I would like to change the following item for an
> internal test, and if it causes no problems on any of our OS's then
> release it in b8.
> 
> This is mainly a linux bug:
> 
> In listen() call for the linux networking code it seems that linux truncates
> the parameter with an unsigned char and the value of 512 (which we are passing)
> it being treated as a 0.
> 
> I would like to try to change this number to 511 which would produce a 256
> I belive.  The line is in http_main: around 1900...  If you could just
> change it on your own server, compile and do a quick test to see if anything
> looks outta wack.  If not I'll make up a patch for the tree.
> 
Maybe it would be better to make this value 255 for linux, and leave it
alone otherwise.

I think the 511 will also make it 255 if it's high bit truncated.

-- 
chuck
Chuck Murcko
The Topsail Group, West Chester PA USA
chuck@topsail.org

Re: bug test before release of 1.2b8

[Dean Gaudet <dg...@arctic.org>]

I took a look at listen() in libc (5.4.17) and the code in the kernel
(2.0.28), it looks to use an int everywhere for passing that parm... are
you sure it's getting truncated?  'cause a value of 0 would cause the
kernel to choose a backlog of 1, which would be really really
noticeable...

Dean

On Tue, 4 Mar 1997, Aram Mirzadeh wrote:

> 
> Before we release b8, I would like to change the following item for an 
> internal test, and if it causes no problems on any of our OS's then 
> release it in b8.  
> 
> This is mainly a linux bug: 
> 
> 
> In listen() call for the linux networking code it seems that linux truncates
> the parameter with an unsigned char and the value of 512 (which we are passing)
> it being treated as a 0.
> 
> I would like to try to change this number to 511 which would produce a 256
> I belive.  The line is in http_main: around 1900...  If you could just 
> change it on your own server, compile and do a quick test to see if anything
> looks outta wack.  If not I'll make up a patch for the tree. 
> 
> Thanks. 
> 
> <Aram>
> 
> -- 
> /*
>  * Aram Mirzadeh, MIS Manager, Qosina Corp. http://www.qosina.com/~awm/
>  * Apache Development Team, awm@hyperreal.com http://www.hyperreal.com/
>  * PGP Key - http://www.qosina.com/~awm/pgpkey.html
>  * BE 49 9D F6 2A A7 22 FC  02 E9 1E 3D F7 0C 67 A0
>  *
>  * I haven't lost my mind -- it's backed up on tape somewhere.
>  */
>