Re: A question of UTF-8

[Christian Borup <borup borup com>]

On l?, 2003-06-07 at 23:52, René Seindal wrote:
> I have an application that is a frontend for a database which then
> drives a web-site.
>
> The database is in ISO-8859-1 with loads of European 8-bit characters.
>
> Gtk2 and perl 5.8 prefers utf8.

Thats not quite true. Gtk2 requires utf8, perl5.8 mostly tolerates it...

> The only thing I do to confront the situation is that on saving data to
> the database, I convert everything to ISO-8859-1 using
> Encode::encode('ISO_8859-1', $data).

> Apparently it works, mostly due to the magic worked by perl 5.8, I
> guess.

That should be enough. What gtk-perl-xs does is to upgrade the strings
to utf8 on every call (if needed). Likewise every string returned from
gtk-perl-xs is marked as utf8 (just setting the flag - no upgrade
needed).

If you are working in a iso-8859-1 locale encode('iso-8859-1' ...)
should be a non-op if the string is not marked as utf8.

Btw you really should be using 8859-15 some of you users might want to
use a Euro sign (€) not available in 8859-1.

> Do anybody have experience with these things?
> Are there pitfalls I should be aware of?
>
> The last thing I want is to find most of my database doubly encoded to
> utf8 :-(

Been there - it sucks if you get it wrong.

> The database is Mysql 4.0.13.  Mysql 4.1 supports utf8 directly, which
> should make life a lot easier.

I have big hopes for that too. I too am stuck with mysql at work.
Someone should do a gtk-perl-xs based DBI shell, that would be sweet.

./borup