🚀 go-pugleaf

Carlos Rimola wrote:
>At 02:05 AM 5/7/98 +0200, Stephen R. van den Berg wrote:
>>Stuart Pierce wrote:
>>>Configure>Mica>nodata-timeout {for all mica lines} 0

>>Careful here.  Up until release 3.3p7, setting the nodata-timeout
>>to 0 did *NOT* work as documented.  

>>Has this been fixed in the latest granite code?  Carlos?  Anyone?

>I am not sure that it has.  I just did a quick review of the most recent
>statement of code for Portware 2.2.30 and the closest thing I found was a
>fix that "adjusted connection timeouts".  Of course, there is one way to
>find out..

Well, ok, I invested some time and tried it a bit (on the production
unit)...  Seems like setting nodata-timeout in 3.4 using the K56Flex
granite code does not have any adverse effects anymore.
I'm running with nodata-timeout set to 0 on all lines, and it still
works.

I did try briefly to see in the nodata-timeout does kick in at some
specified time, but it seemed not to work at all.
What I did:
1. Dial in using a terminal program.
2. I'm now in the menu-processor of the NB, logged in under some username.
3. At this point, I changed the nodata-timeout of the mica modem I was
   logged in on to be set to 5.
4. I waited 5 seconds.
5. Nothing happened.

Could it be that:
1. The nodata-timeout change does not take effect until the next fresh
   connection?
2. They actually took up my suggestion and multiplied the timeout values
   with 10?  (So, 5 means 50 second timeout).

>Note that the portware (the firmware running on the modem card) is/was
>under complete control of Cisco and only their developers had access to the
>code needed to fix this type of problem.

Sadly, yes.  BTW, Carlos?, could you still give us a short braindump
on the latest 3.4 Mica>Diag commands?  Since this is not documented anywhere
in the normal man pages.

>>Also, one other thing, the country code setting for the Mica modems...
>...
>>Has this been fixed in the latest granite code too?

>Same answer.  I do see some indications in the statement of code that would
>lead me to believe that the country code is now more meaningful.

>If you do try it, let us know the results..

Haven't tried this yet.  The trouble is that the whole mica card needs
to be reset to change this.
-- 
Sincerely,                                                          srb@cuci.nl
           Stephen R. van den Berg (AKA BuGless).

This sentance contains threee errors.