/g/ project: specifications and compiler of a modern, simple

just mix the best parts of D and C++ desu

>>60631568
dumb blepeposter

>>60631568
Revive Cyclone

Lets not make a new C based language but instead make a new compiler+Language with assembly.

>>60631568
>without bringing in OOP paradigm
kys

File: 1495187431708.gif (5KB, 200x175px) Image search: [Google]

5KB, 200x175px

>>60631568
>simple and modern compiled language to extend C
Would it be compatible with C libraries?

File: 1494708116053.png (338KB, 558x493px) Image search: [Google]

338KB, 558x493px

>>60632085
>OOP
>not shit

>>60632129

OOP is brilliant bro, it's basically the best of procedural and functional programming brought together.

>>60632007
Which are, in your opinion, the best parts of D and C++?
>>60632085
Everything that can be done in OOP can be done with struct.
I was thinking something like this:

struct x {
    int foo(int, char);
}

x.foo(a,b);
//which is equivalent to
//foo(&x, a,b);

It is just a clean way to organise function and make cleaner code.
>>60632093
No, libraries need to be written.
But with polymorfic data types, we can make standards libraries that include common data structures independent of the data types.

>>60632222

Here's my model for how member functions would work in my ideal language.
Rather than the three way distinction of C++'s static, non-static non-virtual and non-static virtual you have a simple two way distinction. Static/nonstatic. Or if you prefer, virtual/nonvirtual.

struct Foo
{
    void doThing(Foo& this) { }  // a static function
};

Foo myFoo;
myFoo.doThing(); // resolves to Foo::doThing(myFoo);

UFCS sugar makes nonstatic nonvirtuals redundant and this particular style avoids polluting the global namespace with free functions. So I'd like to get rid of them to make things cleaner.

>>60631568
literally "i'll make the logo and be the ideas guy". lurk more, idiot

>>60631568
>having symbolic modules instead header/implementation, because that's the only non-retarded way to include files
>have runtime execution, AST API and special annotations for metaprogramming like in JAI

also >>60632085 or at least add another sane way for runtime polymorphic dispatch
Pro-tip: Library based anything like gObject, moose or clos isn't.

fuck off to >>>/bant/

>operator overloading aka ad hoc polymorphism
fuck off m8
atleast make it type classes instead

File: retardation.jpg (12KB, 473x356px) Image search: [Google]

12KB, 473x356px

>>60632505

>overloading = overriding

File: 1493936129463.jpg (97KB, 1024x768px) Image search: [Google]

97KB, 1024x768px

>>60632514
It overrides the regular function of an operator

>>60633169

It's not overriding if it's not a runtime resolution. Overloading is a compiletime convenience feature.

>>60633309
>It's not overriding if it's not a runtime resolution
Hence why "Java performance" is an oxymoron.

>>60633355

What does Java have to do with anything?
You're sounding schizophrenic, anon.

File: 1494200269193.jpg (6KB, 202x159px) Image search: [Google]

6KB, 202x159px

>>60633375
Maybe because different people are answering to you, retard

>>60633375
Override is from Java. C++ uses overloading instead.

>>60633438

Overriding is definitely a part of C++ senpai, it's the exact reason why virtual functions exist.