Disabling recoding of SMS messages, upgrading from 1.3.1

Hi,

I have a setup at a client where we are using Kannel as a "transport"
layer for an application that handles split/concatenation of SMS
messages itself (and uses other transports apart from Kannel). In order
to do this I wrote a short patch for Kannel that allows sending of
messages directly in the GSM 03.38 7bit / UCS-2BE character set, and
disables recoding of messages to latin1 in the SMSC modules.

I would like to integrate this functionality into the upstream version
of Kannel. At present the use of charset_gsm_to_latin1 () and
charset_latin1_to_gsm () is hardcoded in the SMSCs -- would it make
sense to remove these conversions and convert the message only when it
is processed by smsbox, based on the setting of an option in the
configuration file? Is anyone else working on this/interested in such
functionality?

Regarding the production status of the current CVS, are there any major
issues if I were to upgrade a setup using 1.3.1 to CVS? (Aside from
those already documented, modules getting renamed and so on?). I have
several SMSC connections, using both UCP and SMPP.

-mato