[Dbix-class] branch status

[Carl Franks]

On 21/12/06, John Napiorkowski <jjn1056 at yahoo.com> wrote:
>
> --- Carl Franks <fireartist at gmail.com> wrote:
>
> > On 20/12/06, Matt S Trout <dbix-class at trout.me.uk>
> > wrote:
> > > While I remember: can everybody with an active
> > branch off -current sing
> > > out and tell us
> >
> > "columns_info_for"
> >
> > > (1) what it's designed for
> >
> > Detailed column specs (as much as each database will
> > provide).
> > Primarily designed for use by form validators - to
> > allow automatic
> > length / datatype constraints based on the column
> > type.
> >
> > > (2) where it's up to
> >
> > mysql support.
> > enum handling needs changed to how sqlfairy does it.
> >
> > > (3) when you expect it to be ready for audit
> > before merge to -current
> >
> > Need to find time to learn how dbix-class does tests
> > - and write some.
> > Would be good to also add support for at least
> > SQL::Lite, to see if
> > chosen column attribute names make sense outside of
> > mysql.
>
> Carl,
>
> Seems like some of this could overlap with my needs,
> since I created a 'bind_attribute_by_data_type' method
> in package 'DBIx::Class::Storage::DBI' where I let a
> storage driver provide the required binding attributes
> for a particular column type.  Right now I just have
> some dumb mapping for the Pg driver.  Basically what
> this method does is take a look at the data_type
> attribute out of col_info and returns the correct
> hashref that DBI->bind_param wants.
>
> How is you new 'columns_info_for' method accessed?

columns_info_for() isn't a new method, it's just overridden, to allow
the db engine to provide more detailed info.

> Right now my code rewrites the insert and update
> methods to pass the entire result source so that I can
> do: $source->column_info($column)->{data_type} and get
> something meaningful I can map to.

It doesn't change {data_type}
it adds extra attributes, such as 'default_value', 'is_nullable',
'is_unsigned', the max/min range for integers + floats, the set of
values for an enum, etc...

> Is part of what you are trying to do potentially able
> to return the bind constant?  I know that Oracle has
> that information in it's tables somewhere.  I'll see
> if Postgresql does as well.

Don't know anything about that :)

Carl