Compilation failure for functions returning two dimensional arrays

[crupp@…]

The following model fails at compile time using r24377. It works using r23420. It seems to be caused by functions returning multidimensional arrays.

loadString("
model arrayFuncBug
  parameter Real a = 1;
  Real arr[3, 3];
equation
  arr = func(a = a);
end arrayFuncBug;

function func
  input Real a = 0.01;
  output Real arr[3, 3];
algorithm
  arr := diagonal({a ^ 2, a ^ 3, a ^ 4});
end func;
");

buildModel(arrayFuncBug);

The log file contains:

arrayFuncBug_functions.c: In function 'omc_func':
arrayFuncBug_functions.c:24: error: incompatible type for argument 3 of 'array_alloc_real_array'
C:/OpenModelica1.9.2Nightly//include/omc/c/util/real_array.h:144: note: expected 'real_array_t' but argument is of type 'modelica_real'
arrayFuncBug_functions.c:26: error: incompatible type for argument 3 of 'array_alloc_real_array'
C:/OpenModelica1.9.2Nightly//include/omc/c/util/real_array.h:144: note: expected 'real_array_t' but argument is of type 'double'
arrayFuncBug_functions.c:29: error: incompatible type for argument 3 of 'array_alloc_real_array'
C:/OpenModelica1.9.2Nightly//include/omc/c/util/real_array.h:144: note: expected 'real_array_t' but argument is of type 'double'

[crupp@…]

Just to let you know, this bug seems to be related to the diagonal function. I've since found that the issue disappears when the matrix is hard-coded.

[Mahder Alemseged Gebremedhin]

Fixed in r2726. Yes, the issue was with how diagonal was handled in simplification. It created unnecessary problems for code generation.

[Dietmar Winkler]

It doesn't make sense to keep closed ticket in the "Future" milestone that were simply forgotten to assign to the correct milestone in the past.

[Martin Sjölund]

Removing the pre1.9.4 milestone in favor of 1.9.4.