character set problem with sms_data_decode function

[Jeremy Lainé]

Hello,

There seems to be a problem with the way the sms_data_decode function  
(common/gsm-sms.c) detects the character set of an SMS. The first test  
in this function goes like this:

-----
/* Unicode */
if ((dcs.type & 0x08) == 0x08) {
-----

This is incorrect, because the bit of the Data Coding Scheme byte that  
is being tested does not always mean "UCS2". This bit can also be set  
if you have a 7bit Message Waiting Indication with the indication bit  
set to 1.

I have started working on a patch that corrects this problem:

https://svn.jerryweb.org/public/mkezx/trunk/packages/gnokii/patches/04_sms_decode.diff

However, I also noticed that the gn_sms_dcs contains both a byte value  
and a union (called 'u') which breaks down the different properties  
that can be contained in the DCS byte.

It seems to me that we should be doing something like:

- during the parsing of the SMS header, analyse the raw DCS byte and  
fill the 'u' union appropriately

- in sms_data_decode, use the appropriate gn_dcs_alphabet_type member  
of the 'u' union instead of the raw DCS byte to determine what  
alphabet to use

Your thoughts?

Jeremy

--
http://www.jerryweb.org/             : JerryWeb.org
http://sailcut.sourceforge.net/      : Sailcut CAD
http://opensource.polytechnique.org/ : Polytechnique.org Free Software