[Feature Request] Inline initialisation for enum structs

[sirdigbot]

Very often when using enum structs, it's necessary to have an init() method to set all of the values:

enum struct ABC
{
    int a;
    int b;
    int c;

    void Init()
    {
        a = 1;
        b = 2;
        c = 3;
    }
}

This is not only clunkier to use

ABC x;
x.Init()

but it's also unnecessary code duplication, and introduces a potential mistake in that you can forget to call Init().
The alternative is far simpler and potentially safer:

enum struct ABC
{
    int a = 1;
    int b = 2;
    int c = 3;
}

[Wend4r]

After SourcePawn 1.11, it is possible to do differently by

ABC x = {1, 2, 3};

[Wend4r]

I use ABC::Init() for my decl optimizations (combined with a stack compression), otherwise you can do without it

[sirdigbot]

After SourcePawn 1.11, it is possible to do differently by

ABC x = {1, 2, 3};

That's not a viable substitute because it requires you to specify the default values each time, which is not necessarily what you want.
It doesn't tell you which values correspond to which members, and if the enum struct were to change it could silently cause breakage.

[sirdigbot]

I use ABC::Init() for my decl optimizations (combined with a stack compression), otherwise you can do without it

If you can do without the call to init then that's not the kind of use case I'm talking about here.

[Wend4r]

It doesn't tell you which values correspond to which members, and if the enum struct were to change it could silently cause breakage.

I also foresee the need to use the name resolution engine (sizeof() and constant expressions). With the current spcomp structure, it would be easier to release it into something like a constructor. There is a principle in OOP that it is not always good to do complex fields initialization in the constructor (like create SM handle), it may be necessary to add an additional requirement only in a constant expression (default as 0).

[sirdigbot]

A constructor might be an option but I feel like it's a more heavy-handed feature request than just "when I create an instance of an enum struct use different initial values or call an inlined init function to set them"