🚀 go-pugleaf

Previously Graham E Kinns <Graham@grempc.demon.co.uk> said:
>robm@demon.net (Robert Mortimer) wrote:
>>At home on a 486 running novell dos 7 and a hayes
>>accura the dialer doesn't appear to receive the OK from the
>>modem until after the time out. Regardless of what the
>>time out is set to. If I dial in manually it's fine it just
>>seems to object to the script.
>Flow control setup OK?  Escaped any '\'s to '\\'?  Any difference 
>dropping the DTE rate?  Does it work if you ignore the OK?  Any 
>difference with just 'AT' rather than a setup string?  Have you got 
>any traceing turned on? (ISTR something similar while developing some 
>logging add-ons to the dialler, turned out to be the dialler stack 
>overrunning)

Flow control should be fine, it works reliably under KA9Q
The problem is the same .exe works fine on other machines.
It was ha[ppier if it didn't wait for the OK.

I think further experimentation is needed thanks for the pointers.

>>The other problem is ISDN, we in support where wondering if
>>anyone has got KA9Q working with ISDN and PAP authentication
>>and if they have could they please tell us how. We can get it
>>dial but not login.
>I've not seen anyone report it working to d.s.isdn and don't (yet ;) 

Likewise.

>have access myself, but having worked through the source,
>  ppp sl0 lcp remote authenticate pap
>  ppp sl0 pap user <nodename> <password>
>should be all you need to add to autoexec.net to get the PAP working
>(with suitable chopping of dialer ...).

Yeah that's what we thought and tried.
Sadly it dials and then deems to not bother with the above.
I'll check the authenticate bit again though.

>I had some email correspondance which suggested that KA9Q was crashing 
>somewhere around the authentication exchange - could be another 
>internal stack problem or a timing thing as, IIRC, it starts a new 
>session for the authentication - check nos.log for free pointer errors.
>Unfortunately we didn't make any progress from there ...

Hmm again something to look at.

>I haven't got round to looking at the JNOS source to see how different 
>it is but for 2.16d you'll have to increase one of the process stacks 
>to get the traceing running reliably if you want to debug it.  Drop me 
>a mail if you want the details - if you can find any clues as to where 
>it's falling over, I don't mind looking at any PPP logs or hunting 
>through the source ...

I suspect looking through logs and such may be what I'm being paid
for, I'll let you know how I get on and I may take you up on the
offer yet :-)

Thanks.
--
RobertM
Demon Internet Ltd
         "Trust the computer, the computer is your fiend"