[Nitro] [Og] Og with Sequel
Mark Van De Vyver
mvyver at gmail.com
Fri Nov 2 18:56:53 EDT 2007
On Nov 3, 2007 9:11 AM, Trans <transfire at gmail.com> wrote:
> On Nov 1, 6:46 pm, "Mark Van De Vyver" <mvy... at gmail.com> wrote:
> > Hi,
> > I'm still in the process of working on a DBI adapter.
> > One thing I'm currently working on is trying to leverage off Sequel's
> > "non-Model" code.
> > Primarily to get to the Sequel::Dataset functionality and any other
> > useful features I come across along the way.
> > I've contact Sharon Rosen and she seemed more than happy to have
> > Sequel used in more contexts.
> > I'm working my way through trying to get Sequel specs working, and
> > well as trying to implement some of the other changes/issues
> > discussed on this list (mainly the lower level method naming - which
> > turns out to be fortuitous because otherwise there would have been
> > some name clashes with Sequel).
> > I'm making notes and would hope to get to a point where the relevant
> > sequel modules/classes can be dropped into Og with as little effort as
> > possible.
> > At the moment I'm keeping as much of the changes in the
> > "./adapter/dbi" area as possible.
> > Naturally some things are starting to 'leak' out, mainly related to
> > options handling and this I had to do to be able to write some focused
> > specs.
> > Anyway, I suppose I'm asking for some feedback on:
> > Does this idea (Og+ Sequel) resonate with the community?
> I'm not sure I understand. But I may misunderstand Sequel. Aren't
> Sequel and Og two different approaches to the same task? How can they
> work together? If you are creating a DBI adapter, what does Sequel
> have to do with that?
They really differ on their approach to *::Model.
It is possible to share much of the SQL adapter code (in Sequel this
is Sequel::Database), SQL generation (Sequel seems ahead here).
I'm most interested in 'piggy-backing' on the Sequel::Dataset, which
Og does not have as advanced. There is ready built/tested
Sequel::Dataset functionality that seems fine to me, and it seems to
make sense to try and share these semantics (Og::Dataset) - both
projects can feedback to each other on this.
Given that Dataset is what I'd like to have in common, I'm trying at
the moment to share some of the low level DbiAdapter/Database methods
(e.g it seems to have better pooling/threading management). This is
mostly below the user level, but it seems it can be done.
It is early days yet but it looks promising.
So this is more about trying to share the Dataset interface, and
trying to implement what it takes to get that done.
Hope that is clearer?
> Nitro-general mailing list
> Nitro-general at rubyforge.org
More information about the Nitro-general