From: Hamish M. <ha...@cl...> - 2001-10-01 22:24:02
|
On Mon, Oct 01, 2001 at 02:22:45PM -0400, Chuck Byam wrote: > With the latest round of packet (server) bashing on aprssig it occurred to me > to revisit packet filtering on aprsd again. I made an attempt to "fix" > non-conforming packets before but that introduced problems of its own. For > example, if I truncated long payload fields packets were treated as "new" or > separate packets by other servers thereby injecting duplicate packets. > Should the server be so strict as to drop "any" packet that does not conform > to spec? How should "tactical calls" be handled? While this would fix most > of the griping, it does limit the use to only AX.25 packets with little room > for experimentation. Ideally the server wouldn't have to know every detail of the spec -- just enough to check that the packet is valid. It's too bad that the protocol makes that so hard. As to "tactical calls", I suggest that all packets that use them be dropped, as I consider their use completely unnecessary :-) But I don't think too many others feel that way. Hamish -- Hamish Moffatt VK3SB <ha...@de...> <ha...@cl...> |