On Tue, 20 Dec 2005, Thomas Bruederli wrote:
Martin Marques wrote:
I found some lines in this file. Particularly the ones that give information on the table names and sequence names in the case of PostgreSQL.
Are they really necesary? Shouldn't we use the default table names that are in the SQL/* files?
To avoid naming conflicts in shared databases, RoundCube offers the ability to use other table names than the suggested ones. You don't have to change them, it's just a POSSIBILITY and the default settings should match with the SQL/* files. What's the disadvantage to make table names configurable?
Curious about it. But now that I remember, Horde/IMP also had configurable table names. It's OK for me.
And also, does RC use the PG sequences explicitly?
Yes, these are used to retrieve the record ID after an INSERT operation. If you don't use Postgres you can ignore or even delete them because other databases offer functions like last_insert_id.
Using the $db->nextId($seq_name), right? ;-)
Never had to use it in my code, so I had to read the PEAR pgsql.php code.
Thomas, I'm sending you some changes to the postgres.initial.sql. Sorry for the delay.
It contains:
- Put PK and FK in the table definition for better reading
- Eliminate DEFAULT clause in the user_id fields that reference to the
users table, because it can bring foreign key violations (PG sequences start at 1, so there will not be a 0 value to reference).
I'm also tracking down some DB errors. I'll give you a better report latter.
P.D.: is there any way to configure RC so that it doesn't print the errors in the browser?
-- 18:02:25 up 4 days, 9:57, 5 users, load average: 1.59, 1.57, 1.62
Lic. Martín Marqués | SELECT 'mmarques' || Centro de Telemática | '@' || 'unl.edu.ar'; Universidad Nacional | DBA, Programador, del Litoral | Administrador