At 10:37 PM 2/28/00 -0500, Joe Subich, W8IK wrote:
>The "problem" seems to be unique to nodes running ARcluster
>software. Try the same thing with a node using one of the
>other varieties (CLX, Pavilion, etc.) and the symptom does
>not appear.
Quite possible - I used two different sites, both running the ARcluster
stuff. I don't know what would happen on a node connected by other than
telnet (like a local packet cluster via VHF) or what would happen on nodes
running different versions of the software. But it seems to me that wu3v is
one of (if not the) biggest clusters available on the internet, capable of
supporting hundreds of connections during contests so I would expect
WriteLog to properly support it, if the problem turns out to be a
difference in implementation.
>I would wager that ARCluster gets itself in a CR/LF loop if
>one asks for more data than it has in the queue. Do a
>SH/DX/10 and you're OK ... do a SH/DX/100 and you're sure to
>crash.
Nope, that's not the case - you can easily verify by using TELNET on your
windows machine to connect a node and simulating the same thing you'd get
from the WriteLog packet window. There were easily 100 spots available on
both systems I tested, but the WriteLog packet window puked after
displaying only some of the spots. Never did the telnet connection have any
trouble displaying the data.
The point of my posting was not to point out an obscure bug when you use
sh/dx/100 but to describe the easiest way I could find of making the bug
appear. It's my understanding that several users have had the problem show
up without ever issuing a sh/dx command.
-Marty KI0LO
martyt@pobox.com
--
WWW: http://www.contesting.com/writelog
Submissions: writelog@contesting.com
Administrative requests: writelog-REQUEST@contesting.com
Problems: owner-writelog@contesting.com
|