RE: [RFI] octstr_recode

Sorry - I didn't understand : with which setting it will recode UCS2 to
utf-8 , with mo-record set to true or to false ?

--
Oded Arbel
m-Wise Inc.
[email protected]

When you know absolutely nothing about the topic, make your forecast by
asking a carefully selected probability sample of 300 others who don't
know the answer either.
	-- Edgar R. Fiedler


> -----Original Message-----
> From: Bruno David Rodrigues [mailto:[email protected]]
> Sent: Wednesday, March 06, 2002 4:42 PM
> To: Kannel-devel (E-mail)
> Subject: Re: [RFI] octstr_recode
> 
> 
> Ok. Before we think about real unicode, as we usually inject 
> iso-8859-1 when
> we are sending a message, if (and only if) you set mo-recode = true in
> smsbox
> group, smsbox will try to recode the ucs2 to iso-8859-1 when 
> it receive a
> message.
> 
> As usually, if you don't want it, just don't set mo-recode 
> and everything
> will
> be just like before.
> 
> The patch was commited to cvs yesterday...
> 
> ----- Original Message -----
> From: "Oded Arbel" <[email protected]>
> To: "Andreas Fink" <[email protected]>
> Cc: <[email protected]>
> Sent: Wednesday, March 06, 2002 9:40 AM
> Subject: RE: [RFI] octstr_recode
> 
> 
> > >>  Well what I'm trying to say is that Kannel should 
> support Unicode
> > >receiving but
> > >>  keyword matching in smsbox against unicode is quite a headake.
> > >
> > >  not so - just convert it to utf-8 and strstr() it.
> >
> >
> > Yes but how do you type a keyword in utf-8 in the config file if the
> > config file is plain ascii?
> 
> Just type it using an utf-8 aware editor.
> 
> > the target is to match non ascii
> > keywords. If the keyword would be ASCII then we woulndt have any
> > Unicode.
> 
> of course you would - you forget that the ASCII chars in ISO-8859 and
> single-byte chars in UTF-8 are bitwise identical. if you 
> write a keyword
> using "low ISO-8859-1" (characters where only 7 bits are in 
> use) and try
> to match it to the same word written in UTF-8, you'll have a 
> match. OTOH
> - if you wrote the keyword in mutli-byte UTF-8, one would 
> expect that to
> match this keyword one would have to send a SM in unicode.
> note: ASCII (American Standard Code for Information Interchange) is
> defined as 7 bit and often used as 8bit by adding a 0 bit at the front
> of each char.
> 
> i.e. - as long as you use on the "low" range (0-127) characters of
> ISO-8859-1 in the config file, it doesn't matter if its ASCII 
> or UTF-8.
> it only matter if you attempt to use weirder stuff (not sure where in
> ISO-8859-8 are the accented chracters, but I guess that some 
> of these do
> fall in the "high" range). when that is the case- we should pick a
> coding and stick to it : I'd suggest to go with UTF-8, otherwise we
> either don't have unicode support, or we have a headache ;-)
> 
> --
> Oded Arbel
> m-Wise Inc.
> [email protected]
> 
> "A Meltdown? One of those annoying buzzwords. We prefer to think of it
> as an
> unrequested fission surplus!"
> -- Montegue Burns, (The Simpsons)
> 
> 
> 
> 
>