Request for comment: a new software to manage linux networking features

[Colin Alston]

On 20/02/2008 19:52 namnd wrote:
> If you don't keep a lot of network states yourself, your software can't 
> do many things useful, I hope that after some effort,  you will agree 
> with me on this point.

I'm confused as to how you assert we should agree without an 
explanation of why we should. Especially since it's an event based 
design, it is inherently stateless. Perhaps the confusion is that 
*events* should be read from the kernel not state being polled. 
Obviously if one were polling then the event trigger must have some 
state cached in order not to repeat firing of events on every poll.

>> It's also aimed at replacing ifupdown, not overriding it.
>>   
> Many packages call ifup/ifdown or rely on /etc/network/ifup.d/*, if you 
> don't provide all those features, other packages will break, people 
> won't accept it.

Those scripts would still be there as proposed in Martins explanation 
of the project.

If people are invoking those directly though, I put forward that they 
are already broken.

--
Colin Alston <colin at thusa.co.za>     ______
Linux & Internet Services           /\_\_\_\
Thusa Business Support (Pty) Ltd   /\/\_\_\_\
http://www.thusa.co.za/           /\/\/\_\_\_\
Tel:  (+27) 031 277 1257          \/\/\/_/_/_/
Fax:  (+27) 031 277 1269           \/\/_/_/_/
                                     \/_/_/_/
"Never attribute to malice that which can be adequately explained by 
stupidity." ~ Robert J. Hanlon