Subject | Re: [Firebird-Java] Is it safe to use UNICODE_FSS? |
---|---|
Author | David Johnson |
Post date | 2005-09-05T21:32:54Z |
For unicode, I actually set the database to CHARSET=NONE, and then
define columns in terms of the byte count (not the character count).
This becomes an issue when you need to use database side collations, but
works well when you can use the database for searches and do collations
on the Java side.
The issue with all multibyte character sets in versions prior to 2 is
that there is a presumption in one of the C buffer allocation routines
that character count = byte count. I am not clear about whether the
issues are with a server side function or a client DLL function. I
understand that the issues with multibyte character sets are resolved in
Firebird 2.
I suspect that the optimum path for migrating from FB1.5.x to FB2 will
be through the JDBC drivers themselves.
define columns in terms of the byte count (not the character count).
This becomes an issue when you need to use database side collations, but
works well when you can use the database for searches and do collations
on the Java side.
The issue with all multibyte character sets in versions prior to 2 is
that there is a presumption in one of the C buffer allocation routines
that character count = byte count. I am not clear about whether the
issues are with a server side function or a client DLL function. I
understand that the issues with multibyte character sets are resolved in
Firebird 2.
I suspect that the optimum path for migrating from FB1.5.x to FB2 will
be through the JDBC drivers themselves.
On Mon, 2005-09-05 at 11:25 -0300, Federico Tello Gentile wrote:
> I'm using Firebird 1.5.1, with the latest JayBird JDBC driver (1.5.5).
> I
> created the database with UNICODE_FSS default charset. My application
> is
> a pure java 1.5 running with tomcat. So far I can insert and
> retrieve
> non ascii values form the db and they seem to show up as I entered
> them.
> I know UNICODE_FSS is broken and it has problems, but I'm not quite
> sure
> which problems are they.
> On a previous project I did, I ended up using ISO8859_1, but I
> remember
> some people recomending me keeping UNICODE_FSS.
>
> I know the JDBC driver handles unicode well, as Java is great at it,
> but
> should I keep using it in this database with FB 1.5.1? I don't plan
> to
> transliterate neither to ASCII nor to any other charset. All my input
> is
> UTF-8 and my output should be UTF-8, too.
>
> All I want when FB 2.0 comes out is to be able to create a script
> with
> all the insert statements in my DB and run then in FB 2.0 without
> losing
> the UTF-8 characters.
>
> Thanks.
>
>
>
>
>
> ______________________________________________________________________
> YAHOO! GROUPS LINKS
>
> 1. Visit your group "Firebird-Java" on the web.
>
> 2. To unsubscribe from this group, send an email to:
> Firebird-Java-unsubscribe@yahoogroups.com
>
> 3. Your use of Yahoo! Groups is subject to the Yahoo! Terms of
> Service.
>
>
> ______________________________________________________________________
>