Unable to perform scalar operations

[aloysiousBenoy]

For some reason scalar multiplication is giving me error When I add in the following line of code the code fails to compile

 k_process_noise = 10 * k_intermediate;

I get the following error

 no match for 'operator*' (operand types are 'int' and 'BLA::Matrix<1, 1>')
     k_process_noise = 10 * k_intermediate;
                       ~~~^~~~~~~~~~~~~~~~
In file included from Arduino/libraries/BasicLinearAlgebra/BasicLinearAlgebra.h:142:0,
               Arduino/libraries/BasicLinearAlgebra/impl/BasicLinearAlgebra.h:12:35: note: candidate: template<class MatAType, class MatBType, int MatARows, int MatACols, int MatBCols, class DType> BLA::Matrix<MatARows, MatBCols, DType> BLA::operator*(const BLA::MatrixBase<MatAType, MatARows, MatACols, DType>&, const BLA::MatrixBase<MatBType, MatACols, MatBCols, DType>&)
 Matrix<MatARows, MatBCols, DType> operator*(const MatrixBase<MatAType, MatARows, MatACols, DType> &matA,
                                   ^~~~~~~~
Arduino/libraries/BasicLinearAlgebra/impl/BasicLinearAlgebra.h:12:35: note:   template argument deduction/substitution failed:
/xiao_test/xiao_test.ino:68:28: note:   mismatched types 'const BLA::MatrixBase<MatAType, MatARows, MatACols, DType>' and 'int'
     k_process_noise = 10 * k_intermediate;
                            ^~~~~~~~~~~~~~
Arduino/libraries/BasicLinearAlgebra/BasicLinearAlgebra.h:142:0,
                xiao_test/xiao_test.ino:5:
Arduino/libraries/BasicLinearAlgebra/impl/BasicLinearAlgebra.h:158:27: note: candidate: template<class MatType, int Rows, int Cols, class DType> BLA::Matrix<Rows, Cols, DType> BLA::operator*(const BLA::MatrixBase<MatType, Rows, Cols, DType>&, DType)
 Matrix<Rows, Cols, DType> operator*(const MatrixBase<MatType, Rows, Cols, DType> &mat, const DType k)
                           ^~~~~~~~
Arduino/libraries/BasicLinearAlgebra/impl/BasicLinearAlgebra.h:158:27: note:   template argument deduction/substitution failed:
xiao_test/xiao_test.ino:68:28: note:   mismatched types 'const BLA::MatrixBase<MatType, Rows, Cols, DType>' and 'int'
     k_process_noise = 10 * k_intermediate;

[tomstewart89]

Hey @aloysiousBenoy , I'm pretty sure that line will compile if you modify it like so:

 k_process_noise = k_intermediate * 10.0f;

The problem is that you're using an integer datatype for the multiplier whereas BLA::Matrixs expect a multiplier of the same type as their elements, which by default is float.

Hope that helps! Let me know if you have any more trouble

[aloysiousBenoy]

That solved the issue !! Thanks a lot ! Is this documented/ mentioned somewhere or is this a very obvious thing?

[tomstewart89]

No problem! I'm glad that did the trick.

It's not super obvious and it's also not explicitly documented anywhere. That said, pretty much every function in this library has a unit test that demonstrates how it works so they might be a useful reference if you run into any more compiler errors. In this case the corresponding test is here: https://github.com/tomstewart89/BasicLinearAlgebra/blob/782afcee029544fe5b1dc25265ab8169ab2d0db4/test/test_arithmetic.cpp#L106

Hope that helps! Good luck with your project!

[bobthefairycat]

I guess my recent opened issue has to do with this as well. Somehow, v3.7.0 may have been doing implicit type casting which I guess is dangerous? Wouldn't it default to the type of the left-hand-side being assigned to (i.e. casting to higher resolution should be allowed)?

In our case, it seems that casting int -> float should be permissible (and indeed it is in v 3.7.0).