Comments on: Python code to decode SMS PDU

By: Colm O'Shea

Colm O'Shea — Fri, 18 Dec 2009 09:17:28 +0000

The following table will map GSM character codes to their latin1 equivalent:
gsm_to_latin1 = [64, 163, 36, 165, 232, 233, 249, 236, 242, 199, 10, 216, 248, 13, 197, 229, 16, 95, 32, 32, 32, 32, 32, 32, 32, 32, 32, 27, 198, 230, 223, 201, 32, 33, 34, 35, 164, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 161, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 196, 214, 209, 220, 167, 191, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 228, 246, 241, 252, 224, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 94, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 123, 125, 32, 32, 32, 32, 32, 92, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 91, 126, 93, 32, 124, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32]

Then change the following line of decode_user_data:
out += chr(byte)
to:
out += chr(gsm_to_latin1[byte])

By: Mike

Mike — Tue, 01 Sep 2009 03:27:20 +0000

Hi Pretty Trivial, Can you post your code as well ? It will save me quite some time. Thanks a lot!

Pretty Trivial : I wrote my own one without looking at yours in the end, my decode was pretty different too! You can’t decode properly without using message length. It will work most of the time but the 7-bit GSM encoding of default alphabet uses 0×00 for the @ symbol, plus you can use your own encodings on it too if you like which may also use 0×00. Try making a string that is 8 bytes long and making the last character \x00. Every time \x00 falls on the end of a string in an octet that is a multiple of 7 then it is lost. The only way to know it is there is to use the message length.

By: Pretty Trivial

Pretty Trivial — Tue, 18 Aug 2009 04:28:32 +0000

I wrote my own one without looking at yours in the end, my decode was pretty different too! You can’t decode properly without using message length. It will work most of the time but the 7-bit GSM encoding of default alphabet uses 0×00 for the @ symbol, plus you can use your own encodings on it too if you like which may also use 0×00. Try making a string that is 8 bytes long and making the last character \x00. Every time \x00 falls on the end of a string in an octet that is a multiple of 7 then it is lost. The only way to know it is there is to use the message length.

By: Pretty Trivial

Pretty Trivial — Tue, 18 Aug 2009 00:14:13 +0000

Come on, that is pretty trivial. All you need is the info on the first hit from gooling “sms pdu” and it tells you how to do it. I was about to make this then I googled “python pdu sms” and you beat me to it. Thanks for saving 20 minutes!

By: Carl Audet

Carl Audet — Thu, 11 Jun 2009 21:35:42 +0000

Thank you very much, this was really helpful, especially given how rusty was my python-fu (now I remmber why I like it so much). You should verify your def decode_user_data function though, it leaves the last characters out for strings that have a multiple of 8 as a length.
I’ve added that line before the return statement:

if strip == 7: out += chr(ls)

(ok, not that elegant, but that fixes it)

You probably are also aware that the module as it is doesn’t handle characters that don’t have a 1:1 mapping from unicode / GSM encoding ( fixed it with a simple dict lookup).

Whoever designed that PDU user data protocol deserves to be repeatedly slapped in the face…

Thanks again for sharing.

Cheers.