strange memory leak

[Tobias Burnus]

Hi Mattias,

mattias alveteg wrote:
> Yes, you're definitely right, I used dim= in the wrong way. I'm a bit
> surprised that neither GFortran nor Absoft noticed that at compile time.
> Thanks for pointing that out.

Well, the problem is: It is not known at compile time.
You use:

| subroutine EvaluateStateError(states1,error)
| real(rk),dimension(:,:),intent(in) :: states1
| real(rk) :: Diff(size(states1,1),size(states1,2))
| integer :: i(1),j(1)

The i(1) is correct if size(states1,2) is one. When compiling the module
there is no possibility to find out which size the argument to
EvaluateStateError will have.

The only possibility to check this is at run time. And gfortran catches
this with -fbounds-check. (It is actually a good idea to use the check
options of the compiler, if the program crashes.)

Tobias