[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: National charset support

Francisco Ruiz wrote:

>     I've tried to create some new objects with cn attributes that
> include spanish characters and accents, however, what I get when I issue
> an ldapsearch command is
>             cn: NOT ASCII

Yes.  Rather disgusting isn't it?

> and looking at the dbm file I see and strange row of characters like
>             cn: KLNLKSD981&KSIELSA$OKER

You should rather see:

	cn:: SXNpZHJvIEJvbGHxb3MK

That is the base64 representation of "Isidro Bolaños" in ISO 8859-1 or:

	cn:: SXNpZHJvIEJvbGHDsW9zCg==

that is the base64 representation of same in the UTF-8 encoding of ISO
10646 or Unicode.

> where I should see:
>             cn: Isidro Bolaños
>     Could you help me to figure out what the problem is?  TIA

The problem is that the clients are unable to translate between the
wire encoding and the charset used in your locale.  I am working on
this, but cannot be done without understanding the schema.  So for the
last weeks I have been working on schema issues and these are starting
to come together now.  I'll be back to charsets soon.

So for the time being you will have to leave with it.  Netscape will
want your databases in UTF-8 today and you will want that in the future
anyway.  On CGI scripts you may want to use 'text/html; charset=UTF-8'
in Content-Type and Netscape will do the translation for you both on
output and on input.  But you will have problems with other things like
nss_ldap or your own programs.