ibatis-user-java mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Clinton Begin" <clinton.be...@gmail.com>
Subject Re: Interface binding
Date Sun, 07 Sep 2008 01:05:54 GMT
>> The 3.0 whiteboard discussion also mentions interface binding. Is it
likely that this functionality will return in a future version?

Yes, absolutely.  This is an important part of the improved 3.0 API.

>> The 2.1.7 source code only calls the static method
MapperProxy.newMapperProxy(), so perhaps I could compile/externalize the
missing classes. Is this feasable?

I really wanted it in the 2.x series, but unfortunately I ran out of time
and I'm not sure that it could be done well enough and consistent enough to
call it a full fledged feature.  That said, I think the implementation
that's currently in there was only a few days worth of work.  You could
probably roll your own pretty quickly.

  * Use namespaces for the fully qualified class name.
  * Name the statements after methods.
  * Generate a dynamic proxy at runtime to  bind the interface to the SQL
Map.

Unfortunately you'll still have to specifically map out all of the parameter
types etc... this is where 3.0 has the advantage.  There's enough
information in the method signature to do away with a lot of the XML
required to map most statements.

It is indeed something I want to do.  FWIW:  IBM did build something like
this, not sure where it is now.  At one point this was supposed to be a new
ease of use API for JDBC, but it never happened.

Cheers,
Clinton



On Sat, Sep 6, 2008 at 4:51 PM, Ronald Borman <r.borman@chello.nl> wrote:

> I got started with iBATIS this week. I bought 'iBATIS in Action' (finishing
> the last chapter tomorrow), did some tests and browsed through the mailing
> lists. It looks like a compact, comprehensible and fun to use framework.
> Kudos to the team!
>
> On the incubator mailing list I found a May 2005 announcement which caught
> my attention: interface binding (
> http://www.mail-archive.com/ibatis-user-java@incubator.apache.org/msg02403.html).
> It reminded me of an IBM DeveloperWorks article about Generic DAOs and I was
> eager to try it out. Unfortunately SqlMapClientImpl.getMapper() hasn't
> survived (although it is still listed in the online dev Javadoc).
>
> The 2.1.7 source code only calls the static method
> MapperProxy.newMapperProxy(), so perhaps I could compile/externalize the
> missing classes. Is this feasable?
>
> The 3.0 whiteboard discussion also mentions interface binding. Is it likely
> that this functionality will return in a future version?
>
> Thanks
>
>
> Ronald
>

Mime
View raw message