struct and union about gmqcc HOT 2 OPEN

I'm not an advocate of this form of name at the end, (C does the typedef struct { ... } name) without explicit typedef it look strange because scope-end shouldn't represent the name of the struct/union type. struct name { contents } optional_object_name; makes more sense. (also I don't suggest "struct" or "union" to be used to create an object of that struct/union type, see struct sockaddr_in in C for example ... opposed to just sockaddr_in (in C++)).

from gmqcc.

Exactly, you define the type at a global level:

struct StructName {
    contents;
};

and that's it, and that's the only form we should have. StructName becomes a type, usable the same way as float or entity, we do not use typedef. We also do not use struct StructName varname; to declare a variable.

However, if we want to not require prototyping for types as well, this poses a slight problem. See #4, where it would be solved with a let keyword. (It's just a proposal, we could keep the var keyword, but I don't want to keep the local keyword id1-qc has. There's no need to have separate keywords for the global scope and a local function scope.)

from gmqcc.