Comments on: Weekend Fun: ActiveObjects Testability

By: Chuck

Chuck — Tue, 18 Nov 2008 08:36:06 +0000

The problem i was having is that the comment table gets created but the documentation table is not. anyway i will take this to the user lists. Cheers for the help

By: Daniel Spiewak

Daniel Spiewak — Tue, 18 Nov 2008 08:00:03 +0000

@Chuck

The code you gave should work pretty much as you would expect. In other words, I think the answer to your question is “yes”.

The best place for these sorts of questions is on the AO users list: users [at] activeobjects.dev.java.net

By: Chuck

Chuck — Tue, 18 Nov 2008 07:56:52 +0000

Hello Daniel, I find the ActiveObjects orm very interesting. I am running into a problem with the migrate method. Does ActiveObjects support the case where an entity has several onetomany relationships or even a onetomany and manytomany relationship in the same entity. eg.

public interface Event extends Entity{

public String getName();
public void setName(String name);

public int getNumber();
public void setNumber(int number);

public boolean getComplete();
public void setComplete(boolean complete);

@OneToMany
public Comment[] getComments();

@OneToMany
public Documentation[] getDocumentation();

}

hope you can make sense of my question. thanks alot

By: Glen

Glen — Mon, 19 May 2008 00:28:55 +0000

>It seems to run some sort of metrics based on things like constructors, global state,
>serializability, method side-effects and so on. I think that it’s trying to quantify how easy it is
>to isolate the effects of a certain class from the rest of the library, something which makes
>unit testing easier. That’s just a guess though.

If so it still needs some tweaking in my opinion. The class it pointed out had very few external dependencies. What probably triggered it were some of the conditional logic however most of it was fairly localized and I’m surprised it would have considered it a problem.

By: Daniel Spiewak

Daniel Spiewak — Mon, 19 May 2008 00:22:12 +0000

It seems to run some sort of metrics based on things like constructors, global state, serializability, method side-effects and so on. I think that it’s trying to quantify how easy it is to isolate the effects of a certain class from the rest of the library, something which makes unit testing easier. That’s just a guess though.

One thing I forgot to mention was that I tried to run it against the Scala standard library, but it crashed with an exception. I would have been interested to see the results of that examination, considering how Scala encourages immutable data and the limiting of side-effects. Considering the metrics they seem to be using, most idiomatic Scala libraries would probably score quite well.

By: Glen

Glen — Mon, 19 May 2008 00:17:25 +0000

Not sure how accurate this is. I ran it through some code and the class that was marked as the worst offender was one I would have considered fairly easily tested.