DMZ 1.7.1 (DataMapper OverZealous Edition)

[eluser]OverZealous[/eluser]
@rideearthtom
You can already do that: just include the "join_<name>" column names when using to_json. You should always specify the column names, anyway, so you know exactly what is being sent.

As for your question above, check this post for some pointers. That's basically the only way to make it work.

@Mareshal
Not sure exactly what you mean. I don't build DMZ based on other, existing ORMs. stensi based his original design on someone who based their design on (I assume) the RoR DataMapper pattern.

I based my changes on adding features that I find useful, or others have requested many times. I'm not building toward a goal, per se, but rather, building up. At this point, feature-wise, I'm trying my best to not add more to the base library, as it is already very, very big. The reason 1.7 included so many new core features is that all but a few of them (namely, the paging functions) required changes in the library itself to work. And I figured the paging tools were useful enough that I should have them as part of the core.

I guess that I've been working with ideas for my own ORM library for a long time. When I was a Java developer, the big ORM was Hibernate. But I could never get started, because it felt like you needed a couple hundred XML files to get it working. So, to keep from writing SQL code, I have to write a ton of XML code. And still write a ton of Java to glue everything... Yeah, that's soooo much better. (A lot has changed since then, I know.)

Anyway, when I started using DataMapper, I needed certain things (self-relationships, multiple relationships, join table fields, etc). When stensi was unable to work on DM for a while, I took over (I did discuss this with him some), with DMZ.

Hoepfully that helps explain somewhat.