I'm curious about this. As I understand it, Habari uses PHP Data
Objects (PDO) as an abstraction layer for 'database independence'.

Obviously each supported DB platform needs to have hooks in the
installer process and a schema/upgrade files to be actively maintained
but shouldn't anything that uses PDO to manipulate data in the
database work for all database platforms. Isn't that the whole point ?

Presumably if we use raw SQL in Habari core (with any DB specific SQL
clauses), then that code needs DB specific versions although that
strikes me as a bad thing.

Do we need some wrapper functions for DB specific things like
sequences ?

Clearly, we can not police every single plugin to enforce that all DB
platforms are fully supported but that's a secondary issue.

I have a little knowledge of Oracle and I would be willing to help get
an Oracle port up and running (there is also an Oracle schema created
by Chris Meller in existence that would need sync'ing up) but I'm a
little hazy about the precise scope and magnitude of the required code
changes.

Obviously, the more DB platforms supported by Habari the better -
MySQL and SQLite is merely one more than another leading blogging
platform.

The fact we don't currently have Postgres users doesn't mean there
isn't a potential market out there.
--
Andy

On Apr 8, 4:14 am, Chris Meller <[email protected]> wrote:
> I brought this up on IRC just now and seemed to get a big "we don't care"
> vibe from the participants in the conversation, so I wanted to bring it up
> here as well.
>
> About a week before our surprise release, ticket #924 was created as a task
> item and assigned a blocker priority on the 0.6 milestone. Its goal was to
> remove Postgres from the release before it was packaged and shipped. This
> never happened...
>
> Postgres is totally and utterly broken in 0.6. Any changes that were made to
> support ACL were never tested. In fact, there might even still be a typo I
> noticed in the schema.sql file that never got fixed. It never should have
> gone out with the release, but we've all got it.
>
> If we're at all concerned with the quality image we want Habari to project,
> this would be a big harpoon in its side. Postgres is displayed in the
> installer, there are files there, there's a SQL schema file... it just won't
> work after I finish the installer. Oh, turns out Postgres isn't supported in
> this release and everyone knew that... Then why is it included?
>
> This bothers me for many reasons, not the least of which is that we're
> advocating the use of Trac and then ignore a blocker (3, actually) assigned
> to the milestone. If we're going to use Trac we need to start actually using
> it, because this should never happen...
>
> Anyway, add that rant to the list of major no-no's to keep in mind while
> we're revising our release checklist.
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/habari-dev
-~----------~----~----~----~------~----~------~--~---

Reply via email to