Handling of Double[] #17
Description
Tthe C# Array "Double[]" of double type is treated as pointers to array of pointers to double. For example, a[i] (in C#) would be converted to *(a->GetData(i)) instead of a->GetData(i). After adding "double" as a fundamental type in Lib/src/public/System/support/internal/TypeTraits.h, only some of the cases are handled correctly while others are still not correct.
C# Code Example:
...
// Initialize Gaussian data lists
gauss_h = new Double[fit.num_gauss];
gauss_h_cross = new Double[fit.num_gauss];
gauss_p = new Double[fit.num_gauss];
gauss_p_init = new Double[fit.num_gauss];
for (i = 0; i < fit.num_gauss; i++)
{
vec.error_position[i] = Math.Abs(gauss_p_init[i] - gauss_p[i]) >= extend.peak_diff_cutoff;
}
// Determine the cross-normalized peak heights
for (i = 0; i < fit.num_gauss; i++)
{
vec.normalized_cross_height[i] = (CROSS_COEFF * metric_cross * Math.Abs(gauss_h[i])) + (metric * Math.Abs(gauss_h_cross[i]));
vec.normalized_cross_height[i] /= (CROSS_COEFF * metric_cross) + metric;
}
...
C++ output:
...
Array* array = new Array((int)(fit->num_gauss));
Array* array2 = new Array((int)(fit->num_gauss));
Array* array3 = new Array((int)(fit->num_gauss));
Array* array4 = new Array((int)(fit->num_gauss));
for (uint num = 0u; num < (uint)(fit->num_gauss); num += 1u) {
vec->error_position->SetData((int)(num), Math::Abs(*(array4->GetData((int)(num))) - *(array3->GetData((int)(num)))) >= extend->peak_diff_cutoff);
}
for (uint num = 0u; num < (uint)(fit->num_gauss); num += 1u) {
vec->normalized_cross_height->SetData((int)(num), ((5.0 * metric_cross) * Math::Abs(array->GetData((int)(num)))) + (metric * Math::Abs(array2->GetData((int)(num)))));
vec->normalized_cross_height->SetData((int)(num), (5.0 * metric_cross) + metric);
}
...