| ??? 03/04/10 08:35 Read: times |
#173825 - Debugger != Simulator Responding to: ???'s previous message |
Munish Kumar said:
... in the debugger (sic), all XRAM locations are initialized to 0
No, it's not the Debugger which does that - that's the Simulator. I think there should be atlest an option for this in simulator whether to fill uninitialized location with random data in order to get a closer approximation of the real world. But the real-world data may not be "random" - there may well be features of the specific hardware that tend to make it start with a certain set of values. Those values may be "random" in that you couldn't predict them in advance of first seeing the particular unit, but might be "constant" for that particular unit. And different for all other units. Or it could be the case that 98% of units start with RAM at all zeros - so you'd only notice this bug on the 2%. Or your first Batch might all start up at zeros - so you'd only notice this bug on a later batch where that wasn't the case. Or it could be that 98% of the RAM starts at zero - so you'd only notice this bug if your array happened to fall immediately before one of the 2% of "non-zero" locations. Clearly, you can't simulate all this stuff! As Erik so often says, the fact that you've "tested" it and not noticed any problems does not prove that there are no problems! Another option, of course, is to have your startup code zero all RAM for you... |
