[HECnet] Is 00000A a valid DECnet node name?

[Thomas DeBellis]

I've been bumping into a number of bugs in the Tops-20 DECnet 
implementation, some worse (like crashing with an ILMNRF and trashing 
parts of files) than others (simple job hangs and undocumented 
behavior.  Fortunately, I have been able to fix most of them; there's 
only one that I had to program around and another that I have yet to 
investigate.

Meanwhile, I have a great deal of testing to do of the SETND2 code.  
However, there are a number problems doing that, besides those alluded 
to, above.  For a number of cases, I have to start from a fresh boot 
with the node table completely empty.  There is no command to 'empty' or 
reset it.  So that's a pain.  Another thing you can't do is extract or 
inspect the full node population.

It occurred to me that the data structures and associated node 
management routines that I wrote for the decisioning logic in SETND2 
turn out to be fine for simulating the NODE% JSYS itself.  So what I do 
is fork an inferior version of SETND2 and trap the NODE% with a TFORK%.  
It's starting to come together to the point where I need to model the 
actual rules used for DECnet node syntax.

I know that a DECnet node can be a maximum of six characters long with 
only the numerals 0 (zero) to 9 (nine) and the letters A to Z.  However, 
I noticed some code in COMND% that checks to see that a node name has at 
least one alphabetic character in it.  I had never thought about that 
and was wondering what the actual standard says (or where that standard is).

Does is matter where the letter is?  In other words, are five numeral 
zero's followed by the letter 'A' valid?  Is '00000A' OK?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20210618/715e836e/attachment.htm>