[maemo-developers] AT command AT+CSMP=1,167,0,8 not possible on N900

From: 白い熊 maemo-developers_maemo.org at sumou.com
Date: Mon Mar 1 13:22:36 EET 2010
Marius Vollmer <marius.vollmer at nokia.com> writes:
> Ahh, tough.  Maybe you can wrap pnatd somehow so that the sub-process
> that Emacs sees does only exit at the very end.

Yeah, the problem is that the only solution of this sort, I've found is
just init pnatd, wait half a second, send one AT command, wait half a
second etc. After the SMS dispatch, wait 10 seconds, then kill the pnatd
process.

The problem is, sometimes due to network congestion or what have you,
the wait should be even longer, when sending the SMS, than 10, sometimes
it's really fast. It's very long especially, when sending multiple parts
of a long SMS.

But then increasing these intervals becomes unbearable...

I'm just thinking, surely there must be some dynamic way to start the
pnatd process and monitor from Emacs it's ongoing output, i.e. wait for
the OK after sending one AT command and then commence...

The USSD widget guys are doing something of that sort in Python at
https://garage.maemo.org/plugins/scmsvn/viewcvs.php/trunk/ussd-common/src/usr/bin/ussdquery.py?root=ussd-widget&view=markup

But one should be able to do something of this sort in Emacs.

I considered running pnatd without a filter, just outputting its feed to
a buffer. Then at periodic intervals process whether the buffer has been
updated, tail the buffer's contents and check for an error or OK. But
this still requires some looping sleep-for periods etc.

Could the solution be via Transaction Queues? Never used those, but it
seems that queueing the AT commands could be the way to go, I'm just
thinking it might rest on the same problem, i.e. the process feedback
would only update and give hints to the transaction queue after the
process end. Trying to study now, how these Transaction Queues work...
-- 
C уважением / 宜しく御願い致します / Best regards / S pozdravem / Z poważaniem / Mit freundlichen Grüßen

白い熊
More information about the maemo-developers mailing list