Bug fix for Failed0 error checking functions

[rafmudaf]

Bug fix in repeated Failed0 error checking functions

This pull request corrects a bug in the Failed0 error checking functions. Since the bug is repeated exactly, I'm assuming it was present when the function was introduced and then spread by copy / paste throughout. Therefore, while this is likely not causing any issues at the moment, I'm submitting the change to fix it so that it does not continue to spread.

Currently, most of the Failed0 functions check for whether the global error, usually ErrStat, is not 0. However, these functions are primarily used after array allocations with the error status going into the local error, usually ErrStat2. See the example below. In the current Failed0, the if-statement to set the error status and message is only entered if the global error is already non-zero, but it is intending to check for whether the local error is set by the allocation.

! In FAST_Farm_Subs.f90

! Farm_InitWD
ALLOCATE(farm%WD(farm%p%NumTurbines),STAT=ErrStat2);  if (Failed0('Wake Dynamics data')) return;

contains

   ! check for failed where /= 0 is fatal
   logical function Failed0(txt)
      character(*), intent(in) :: txt
      if (errStat /= 0) then
         ErrStat2 = ErrID_Fatal
         ErrMsg2  = "Could not allocate memory for "//trim(txt)
         call SetErrStat(errStat2, errMsg2, errStat, errMsg, RoutineName)
      endif
      Failed0 = errStat >= AbortErrLev
      if(Failed0) call cleanUp()
   end function Failed0

There are two more changes included:

• When I changed the Failed0 functions, I also used a consistent case for the error status and error message variables to make it easier to scan for these words. Note that a few instances of Failed0 already had the correct error checking but inconsistent case. I did not modify those.
• The AeroDyn Inflow C Binding library was not copying the Fortran error to the C error in Failed0. Even if the error was caught by the Fortran module, it wouldn't have been sent to the calling code. This is fixed here.

[andrew-platt]

Thanks for fixing that!