[BUG] Template deduction fails

[mike919192]

Hi. I found an instance where template deduction works in cpp but does not in cpp2. Here is the minimal example:

#include <iostream>
#include <array>
#include <complex>

//Syntax1 here
constexpr size_t N {64};
constexpr uint n {7};

//Syntax2 here
main: () -> int = {
    
    expected: std::array<double, N> = 0;

    observed: std::array<double, N> = 0;
    observed.at(n) = 1.0;

    //template deduction should work here but doesnt
    max_error: double = calc_max_error(observed, expected);
    //max_error: double = calc_max_error<N>(observed, expected);

    std::cout << max_error << "\n";

}

calc_max_error: <N: size_t> (observed: std::array<double, N>, expected: std::array<double, N>) -> double =
{
    errors: std::array<double, N> = 0;

    i: size_t = 0;
    for errors next i++ do (inout error) {
        error = std::abs(observed.at(i) - expected.at(i));
    }

    return std::max_element(errors.begin(), errors.end())*;
}

Using gcc11. Here is the compiler output

Starting build...
"../cppfront/cppfront/source/cppfront" *.cpp2 && "/usr/bin/g++" -fdiagnostics-color=always -I../cppfront/cppfront/source/ -std=c++20 -g *.cpp -o /config/workspace/hellocpp2/min_example
min_example.cpp2... ok (mixed Cpp1/Cpp2, Cpp2 code passes safety checks)

min_example.cpp2: In function ‘int main()’:
min_example.cpp2:18:37: error: no matching function for call to ‘calc_max_error(std::array<double, 64>&, std::array<double, 64>&)’
   18 |     max_error: double = calc_max_error(observed, expected);
      |                       ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~
min_example.cpp2:25:39: note: candidate: ‘template<long unsigned int N> double calc_max_error(cpp2::in<std::array<double, N> >, cpp2::in<std::array<double, N> >)’
   25 | calc_max_error: <N: size_t> (observed: std::array<double, N>, expected: std::array<double, N>) -> double =
      |                                       ^~~~~~~~~~~~~~
min_example.cpp2:25:39: note:   template argument deduction/substitution failed:
min_example.cpp2:18:37: note:   couldn’t deduce template parameter ‘N’
   18 |     max_error: double = calc_max_error(observed, expected);
      |                       ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~

When the template parameter is given the code compiles and works.

[JohelEGP]

This is resolved by #533.

[mike919192]

Great! Thanks