[kvon]

I'm having trouble trying to create an array of float or double. I'm using JDE 4.1.0.185. What happens is when I set a breakpoint any time after this array declaration has executed, the debugger freezes. Not sure if this is something I'm doing wrong, or if it's an issue with the JDE itself...??

Strange thing is, when I change the array to any form of an int or long, it's fine. But with float or double, the debugger chokes. I can also use a float or double variable, it's just the array that causes a problem. Tried using strictfp (long shot, but don't know what else to try), that didn't fix it.

Here's the code:

public class myClass {
// any of these variations cause the problem
private double [] C = new double [4];
private double [][] C = new double [4][4];
private float [] C = new float [4];
private float [][] C = new float [4][4];

// this is fine:
private int [][] C = new int [4][4];

TIA

[bemshaswing]

I've never experienced this myself, but if you're really having problems you could just move to the Objects insteand of primitives
ex: private Double[] C = new Double[4];

[kvon]

Originally Posted by bemshaswing

...you could just move to the Objects insteand of primitives
ex: private Double[] C = new Double[4];

This sounded like a fantastically simple idea at first blush - as in "why didn't I think of this". It may only serve as a workaround for me though - I'm doing an algorithm that's fairly heavy on the number crunching (matrix determinants), and it makes the code pretty ugly to have to do lots of math using "variable.doubleValue()". It may also prove to be a bottleneck in processing time.

At any rate, thanks for the temporary fix, but I'd sure like to know why this throws the debugger for a loop. Starting to think more & more it's RIMs JDE that's the problem.

[bemshaswing]

$ to donuts it's the JDE. Try downloading 4.2.1 and working through there. If that works and you specifically need to build using the 4.1.0.185, just have both share the same workspace. This way you can work and debug using the 4.2.1, then build your deployments with the 4.1.x

[kvon]

Actually, I DO need to build using 4.1.0 (that's another sore point - for a different message thread), so trying 4.2.1 sounds like a good idea, if not maybe a little painful. Gives me a couple of other ideas too - I'll try debugging on the device, see if that makes a difference. Also, I've never shopped around other 3rd party debuggers, still fairly new to BB development - can anyone recommend a couple of the good ones? It might be time to upgrade...

Thanks!

[Skipper_Joe]

just an idea:
maybe you define rapc property "MicroEdition-Configuration: CLDC-1.0" instead of "MicroEdition-Configuration: CLDC-1.1"?
CLDC1.0 doesn't support float/double types.

[kvon]

Originally Posted by Skipper_Joe

just an idea:
maybe you define rapc property "MicroEdition-Configuration: CLDC-1.0" instead of "MicroEdition-Configuration: CLDC-1.1"?
CLDC1.0 doesn't support float/double types.

The thing that's funny about this is that the problem specifically is with the debugger, not the code. The code runs just fine, including the float or double calculations. It's just when I try to use the (RIM JDE) debugger to "watch" the double values -- everything freezes up. Tried debugging on the device too, same result.