My mistake...
It appears that the SMSC we're connecting to is sending/receiving data to us as LATIN
characters (except the GSM extended characters). They're also declaring the
data_coding to be 0, which I believe is indicative of the 7bit GSM alphabet. Kannel
performs a GSM->Latin conversion (and a Latin->GSM conversion when sending TO the
SMSC) and thus the double conversion.
Is there a way to tell kannel not to do that conversion? I'm using 1.3.1 development
version.
Many thanks,
Bryce
-----Original Message-----
From: Bryce Pasechnik
Sent: Wed 10/1/2003 9:35 AM
To: [EMAIL PROTECTED]
Cc:
Subject: Kannel Character set issue GSM->Latin appears to run TWICE!
hey guys, I must be doing something silly in my kannel use/config. However,
this is what happens:
When I send from a handset to my operator assigned shortcode the character
"@", it comes to kannel and I see the message in the kannel.log debug output as being
of hex value '00'. When it reaches kannel's access.log, it is now showing up as an
inverted exclimation mark. It is also passed onto our web service as an inverted
exclimation mark. At first I was confused by this mapping, but a quick look at how a
GSM -> latin conversion works, one realizes that Kannel appears to be converting the
data TWICE.
00 in GSM to latin1 ==> decimal 64 (the @ symbol).
64 in GSM to latin1 ==> decimal 161 (the inverted !)
A similar problem seems to be occuring when I use kannel to send to my
handset. In order for an @ symbol to appear on the handset, I need to send it decimal
161, as it appears to go from latin1 -> gsm twice.
What can I do to alleviate this?
Thanks,
Bryce
