Debugger say "-1.#IND" what does that mean?

[RV74]

hello guys, MMF built-in debugger reported me a strange value "-1#IND" while I was using it to trace variables (or Alterable values in MMF term).

That variable stores the distance from enemies to player, here's how I update the value:

Always
- Enemies: Set Alterable Value L To 
   Sqr(
    (X( "AUX::PlayerSprite" )-X( "Group.Enemies" )) pow 2) +
    (Y( "AUX::PlayerSprite" )-Y( "Group.Enemies" )) pow 2)
   )

where AUX::PlayerSprite is the role controlled by player.


Thanks!

[RV74]

Okay, I finally figured out that I've made a stupid mistake.
The "+" symbol in my expression is mistyped as "-" therefore MMF attempt to calculate the square of a negative number, that is not a real number.

So I guess that "-1.#IND" reported by debugger means the number is NULL or unable to be presented.

[DavidN]

Interesting nonetheless - I hadn't seen that before.

[Joshtek]

So I guess that "-1.#IND" reported by debugger means the number is NULL or unable to be presented.

Correct. -1.#IND is similar to a null/NaN (not a number) and roughly translates to "an indefinite (negative?) (number?)".

To quote Wikipedia, "no negative number can have a real square root".

[LazyCoder]

But putting -4 ^ 0.5 gives -2...

I just thought that was interesting.

[Anders]

-4 ^ 0.5 is evaluated as -(4^0.5) not as (-4)^0.5

[Dynasoft]

Wow, that's the worst example of Clickteam's occasional operator precedence errors I've seen yet

[Corentin]

never realized this ! This should be fixed, anyone trying to use some maths could get such an error and it would be very hard to debug... (of course you could use Abs(Pow0.5 but this seems stupid)

[Anders]

Mi TI-Interactive application calculates with the same operator precedence rules as MMF2. It's not an error.

I think the order is:
^ (pow)
% (mod)
* /
+ -

[Dynasoft]

Except that in this case, it's the unary "-" not the binary "-", and the precedence should be different. The unary "-" (and unary +, if supported) should be acted on before any other operators, or even interpreted as part of the number: (-4) not -(4).

Btw, for those who don't know Clickteam's last operator precedence mistake was the * and / operators in Jamagic, interpreting a*b/c*d as (a*b)/(c*d) instead of ((a*b)/c)*d or a*(b/c)*d.