Re: Help with UTF8 needed.

[Thomas Bayen <tbayen bayen de>]

Alberto Manuel Brandão Simões schrieb:
> Hi.
> I am trying to use gtk-perl and started with a too simple example: a
> dicionary saved directly to a DB file. I'm sending the simple 86 line
> of  code in attach in case it can help.
>
> The problem is that if I use latin1 characters in the key for the DB
> file, when I retrieve it again, I got a UTF8 string, which is not
> interpreted as such by Perl.
>
> It is not simple to explain, so please test the app with the following
> test case:
>   insert 'cão' as term and 'foo bar' as data
> exit the application
>   re-enter the applicaiton

I haven't testet your application but perhaps your problem is more a
mldb-problem than a gtk-problem. My experience is that gtk is very clean
and consistent in using only utf8 strings. I am writing german programs
(with german "umlaut" äöü) for a long time and it ever has been a very
confusing theme. If you want to know my opinion: I always try to use
everywhere uft8 because this is the only real solution. It's only
drawback is the difference between byte length and char length. And this
may be a problem in some databases (That's where we are with MLDB?).


1. If you decide to use utf8 everywhere you should really say
"everywhere": In your console, your files, your editor, your sourcecode
etc. That's my way and it works well if you have set up it correctly one
time. You should read

     * perldoc utf8
     * perldoc encoding
     * perldoc perluniintro
     * perldoc perlunicode
     * perldoc perlrun (-C switch)

I don't know if MLDB ist really utf8-safe. Try it out.


2. If you want to change just the encoding you should also RTFM
mentioned above to get the needed know-how but you should read

    * perldoc Encode

too and use this. Look out for encode(), is_utf8() and friends.


> Thanks in advance.
> Best regards,
>  Alberto Simões

HTH,

Thomas Bayen