Bug 9056 – More precise error messages when function arguments are wrong

This is wrong code (it's not meaningful code because it's a full reduction of a much longer program): struct Foo {} void bar(in Foo* f4, in ref Foo f5, in uint m, ref const(Foo)* f6, ref double x, ref uint n) {} void main() { Foo f1; Foo* f2, f3; double x = 0; uint n = 0; bar(f2, f1, 0, f3, x, n); } DMD 2.061alpha gives: test.d(13): Error: function test.bar (const(Foo*) f4, ref const(Foo) f5, const(uint) m, ref const(Foo)* f6, ref double x, ref uint n) is not callable using argument types (Foo*,Foo,int,Foo*,double,uint) The function bar() has six arguments. Generally it's better to avoid writing functions with six arguments, especially in languages that don't have "named arguments", but once in a while this happens. In bar() the arguments are tagged in various complex ways. Given such error message it is not immediate to see that this is a good way to fix the code: struct Foo {} void bar(in Foo* f4, in ref Foo f5, in uint m, ref const(Foo)* f6, ref double x, ref uint n) {} void main() { Foo f1; Foo* f2; const(Foo)* f3; double x = 0; uint n = 0; bar(f2, f1, 0, f3, x, n); } Probably during the compilation the type system of the D compiler has more information. So in my opinion here the compiler should give a more precise error message, that tells what arguments are not acceptable. In this case the problem is just in the f6 argument of bar(). So I suggest an error message similar to (or better): test.d(13): Error: function test.bar(const(Foo*) f4, ref const(Foo) f5, const(uint) m, ref const(Foo)* f6, ref double x, ref uint n) is not callable using argument types (Foo*,Foo,int,Foo*,double,uint). The incompatible arguments are: f6. (Even better is to explain why they are not compatible, but this probably is too much complex to do, so it's better leave this to a future enhancement request.)

I've mostly addressed this with a pull for Issue 15613, but I'll leave open because there's also the issue of passing the wrong number of arguments.

Ignore that, I was thinking of Issue 16165. *** This issue has been marked as a duplicate of issue 15613 ***