Re: [Jython-dev] java.sql.Date instead of datetime.date?

[Charlie G. <cha...@gm...>]

On 7/18/06, Alan Kennedy <jyt...@xh...> wrote:
> Whereas the jython SVN trunk currently has a pure-python datetime.py?
> Where did that come from? The jython SVN browser doesn't provide access
> to file history.

http://fisheye3.cenqua.com/browse/jython/trunk/jython/Lib/datetime.py

Looks like Frank added it.

>  > I am wondering if anybody is already looking into this, or if I
>  > should go ahead and look into changing
>  > com.ziclix.python.sql.DataHandler to create python datetime instances.
>
> While the pythonic datetime option is preferable, I think, it would be
> nice if this were somehow configurable, so that existing code relying on
> java.sql.Date does not break, and could be configured to provide the
> legacy behaviour.

This may be a boneheaded idea, but could we write an implementation of
datetime.date in Java that extends java.sql.Date?  I know it's
impossible right now since it has to extend PyObject, but I don't see
any reason we couldn't extract an IPyObject interface from PyObject
that replaces PyObject as the argument to all internal Jython methods.
 PyObject would just be the default implementation of IPyObject.  With
all the code generation we're already doing for the new style classes
it'd be easy to copy the existing PyObject implementation into classes
implementing IPyObject so they don't actually have to recreate
PyObject's implementation.  Having datetime.date be the same as
java.util.Date would be nice for general usage in Java in addition to
keeping legacy Jython code running.

Charlie