Re: Character set recoding.. Can we eliminate it?

----- Original Message -----
From: "Worik Macky Turei Stanton" <[email protected]>
To: "kannel-devel" <[email protected]>
Sent: Tuesday, December 18, 2001 11:18 PM
Subject: Character set recoding.. Can we eliminate it?


> Friends.
>
> I am only working with SMPP for now, and have no knowledge of other
> SMSC types so I may be barking up the wrong tree, but...
>
> I have been bitten by the calls to
> charset_gsm_to_latin1(msg->sms.msgdata); and
> charset_latin1_to_gsm(pdu->u.submit_sm.short_message); in smsc_smpp.c
> pdu_to_msg and msg_to_pdu.
>
> I solved my problem by commenting the calls out.  I know what
> character set I want my message encoded in and I encode it in that
> character set (using iconv) before I pass it to kannel.
>
> Why not cut out *all* the character set encoding in kannel and rely on
> the developer that is using kannel to ensure the messages are encoded
> correctly before they are passed to kannel?
>
> The thing I really like about kannel is it's simplicity.  But it could
> be simpler.  The analogy is with apache (or web servers generally) and
> CGI scripts.  All the web server does is send and receive http[s] and
> all bells and whistles go into CGI scripts.  We should do the same
> thing in kannel.  Transmit, receive and route messages but have
> external programmes do any fancy stuff like translating character
> sets.

And then why not cut also smsc-id/accept/reject, accept/deny-prefix
and sms-service keyword processing. Kannel wouldn't know about charsets,
so keyword processing won't work anymore.

And then, why don't just compile smsc_* right into your code and use it
directly ?

There's a line between functionality and simplicity. The problem is that now
we
don't know it neither we hava a team manager to define it......


>
> Worik
> --
>                                                       Worik Macky Turei
Stanton
> Whew!
[email protected]
>
Aotearoa
>
>
>