possible re-encoding bug ?

Developers,

I have asked this question on the users list , but have had no response. Since 
it appears to be a bug I guess maybe I should have tried this forum first, 
but I was not sure that it wasn't just my misunderstanding of kannel 
configuration options. Can anyone help shed some light on the behaviour I am 
observing ?.

I am experiencing a problem with character recoding when passing thru the
bearer box to the sms box . According to the specs the gsm encoding for the
dollar symbol ($) is 0x02 .    It should (?)  then be converted to the 
ISO-8859-1 encoding of 0x24  for internal use by kannel and eventually end up
being embedded as the $ char in any http url that I insert the message
content into.
The bearerbox debugging logs display this character in the incoming message
 as $ - I presume it has been already been converted from 0x02 before display
 in the log ? . The smsbox debug log shows the equivalent char in the
 outgoing generated url as %A4 which is what the urlencoding for ISO-8859-1
 the generic currency symbol that looks a bit like  '#'  ( ¤ ). This is what
 the gsm_to_latin1 mapping in the kannel src code  is set up to map the GSM
 byte value 0x24 to .
So it either appears that there has been an incorrect double remapping of the
message content ( the '$' char encoding translates from 0x02 to 0x24 and hence 
to 0xA4) 
OR
the incoming message encoding is NOT using the GSM 03.38 character set - even
though the bearerbox log shows it as being the default encoding
( data_coding: 0 = 0x00000000 )

Can any one comment on this behaviour ? Am I right in thinking this is a bug
 - or I am I missing something about the behaviour of the SMS system
 implenented for the mobile networks in Australia or perhaps about the way I
 should configure kannel ?

Regards
Peter Farmer