[Typo3-dev] Porting to PostgreSQL (WAS: How to integrate non-MySQL ... )

[Frank Joerdens]

On Fri, Mar 26, 2004 at 05:50:27PM +0100, Martin T. Kutschker wrote:
[]
> There is a set of standard SQL 
> datatypes which many database systems implement (but sometimes with a 
> different name).  Many database systems have proprietary extensions (eg 
> Mysql has unisgned integers and others). For the DBAL layer you will 
> have to create a set of datatype requirements. *)

Has anyone done this yet? To port the core, it'd be necessary, I think,
to know why a tinyblob is a tinyblob and not a blob etc. (why a
varchar(10) is not a varchar(11) etc.); i.e. if this implies a
constraint and wasn't done for e.g. performance/storage considerations
...

> An even bigger problem are different functions supported. Again a set of 
> "allowed" functions should be made. **)

Presumably you're not referring to server-side functions (the core
doesn't use any server-side logic, as far as I could make out?!) ...
then this would be the DBAL's business to translate from one to the
other, and which functions are common to all supported databases. Has
anyone decided which DBAL you're going to use with Typo3?

> And of course there are always slightly different SQL syntaxes, 
> sometimes extensions of the standard, some histoic incompatibilites. ***)

Again, it hinges on your particular DBAL technology decision.

This thread here lists some of the things to keep in mind:

http://www.phpkitchen.com/article.php?story=20030120094236904

For instance, an important advantage that Pear MDB offers over Adodb is
DB schema indepence with different RDBMS's:

--- begin quote from the thread linked above ---

So your criticism is party correct, but we are working on it to make
solutions that surpass ADODB in a lot of key areas. Nevertheless keep in
mind that ADODB also has a different design philosophy. For example the
author does not think it is worthwhile making DB schemas RDBMS
independent. And he offers reasons for that. MDB does allow you to
define schemas in XML so that you can use the same schema for different
RDBMS. You can even export data from one RDBMS to another. At the same
time ADODB offers caching and html output which are handled by separate
Classes in PEAR.

---- end quote from the thread linked above ----


Regards,

Frank