RE: [PATCH] Chararcter @ in SMPP

> -----Original Message-----
> From: Dedy Sutanto [mailto:[email protected]]

> Patching charset.c is not good.
> But if it affected other driver (not only SMPP), 
> Anybody have a clue where the right way to go? (To each driver?)

Yes - use alt-charset to supply the correct character-set for each SMSC on a per SMSC basis and convert to and from that character set

> I think, we must go back to previous implementation of SMPP 
> driver like in 1.0.2 / 1.0.3, where smsc_smpp.c has its own 
> character mapping.

Don't go there. there are many SMPP implementation that are correct, and even if you don't break the entire Kannel to support some broken implementation, breaking the SMPP driver is not better.

--
Oded Arbel
m-Wise mobile solutions
[email protected]

+972-9-9581711 (116)
+972-67-340014

::..
"Everything that can be invented has been invented."
	-- Charles Duell, head of the U.S. Patent Office, 1899


> 
> Regards
> -dedy-
> 
> 
> > I have to disagree with you.
> > In GSM character set @ is coded as 0x00.
> > If your SMSC doesnt want to have @ encoded as 0x00 but wants to have
> ISO8859-1 as input, this is something which should be taken 
> care of in > the
> SMSC driver or on the SMSC itself. patching charset.c would 
> make all SMSC
> drivers fail. 0x00 encoding is correct according to the
> > standard GSM character set. If your SMSC wants some other 
> character set as
> input, well then you should make a patch in smsc_smpp however
> > its very new to me that SMPP specifies non GSM character 
> set as default.
> 
> >  And frankly if Logica and Comverse do it different, this 
> doesnt mean
> automatically they follow the SMPP specification.
> 
> > So I vote -2 also for this patch.
> 
> > IF this is really to be done, then we should have a config 
> parameter for
> SMPP driver to tell which charset to send messages
> > to the SMSC. Be aware if @ is broken, all the other 
> accented characters
> are probably too.
> 
> actually this is not the case. I have seen this problem with 
> 3 different
> smsc's, and not only smpp. and even if other characters are 
> allright, smsc's
> tend to suppose to get @ not as 0x00 but as something else .. 
> and the same
> in incoming messages ..
> 
> but generally - yes - patching charset.c in this respect is 
> not the right
> thing to do ..
> 
> 
> k
> 
> 
> 
>