On Thu, Aug 29, 2002 at 11:36:01AM -0700, Guy Harris wrote:
| On Wed, Aug 28, 2002 at 12:46:41PM +0200, Hannes Gredler wrote:
| > what i am saying does not contradict do what you are saying :-);
| > i like the fallback idea ....
|
| It's *not* a fallback! It's intended to be *the* way to register
| well-known MAC addresses.
|
| > non-programmers can do it that way ... no problem with that, however if
| > a programmer wants to register his MAC addresses in the code we should permit
| > that freedom of choice
|
| But why would they want to do so, given that it's easier to do so by
| tweaking a file?
i scanned the cvs repository and did not find a ethers file that comes
with the standard distribution;
- so the file is not there yet;
- who takes care to set it up / bundle it with the distribution ?
secondly it would be possibly for a dissector to compute names;
- a good example of this is the cisco HSRP / VRRP protocol:
the MAC addresses [i'll stick with the HSRP example] follows the
convention 00:00:0c:07:ac:<group> so the dissector could
register the name All-HSRP-group-xx-routers for all
256 MAC addresses much more simpler than writing 256 lines
in a file;
all i am asking is freedom of choice .... and i really can't understand
why this is bad ?
[ ... ]
/hannes
