[HECnet] continued adventure with buggy DECnet on VAX/VMS 3.5

[Paul Koning]

> On Dec 17, 2021, at 1:14 PM, Paul Koning <paulkoning at comcast.net> wrote:
> 
> 
> 
>> On Dec 17, 2021, at 11:08 AM, Supratim Sanyal <supratim at riseup.net> wrote:
>> 
>> I turned the Ethernet circuit off and set up a DMC line to DECnet/Python. What seems to be interesting is the Area does not seem to be propagating as part of the DECnet address and confusing DECnet/Python. [Non-Issue, just saying.]
>> 
>> The following four things are reported in a loop.
>> 
>> Event type 4.11, Initialization failure, line fault#012  From node 31.32 (PIPY), occurred 17-Dec-2021 16:03:05.499#012  Circuit = DDCMP-31-42#012    Reason = Circuit synchronization lost
>> 
>> Event type 4.10, Circuit up#012  From node 31.32 (PIPY), occurred 17-Dec-2021 16:03:15.473#012  Circuit = DDCMP-31-42#012 Adjacent node = 31.42
>> 
>> DDCMP-31-42 packet from wrong node 42
> 
> Interesting bug.  That means DECnet/VMS is sending a Phase III style node address even though it is reporting its version as Phase IV.  The spec is very clear, this isn't allowed and PyDECnet rejects it for that reason.
> 
> It would not be all that hard to put in a bug workaround and supply the local area number in this case, just as would be done for a Phase III neighbor.

I was wondering: does it do this only if you configure that node as an end node?  Would be interesting to see if you can make it go away by setting the node type to (level 1) router.

	paul