you know,i would actually love to doc silk.net if i actually knew what 99% of the functions actually

SSpider-Kyle you know,i would actually love to doc silk.net if i actually knew what 99% of th...

R

Redhacker2•8/12/22, 12:23 AM

use intellisense :P, oh wait, intellisense can actually crash your IDE when talking 12000 overloads

RRedhacker2 IMHO there should be 2 overloads for Silk.Net 3.0 per existing overload

A

Aqua•8/12/22, 12:23 AM

So have 24000 overloads for that one function now?

A

Aqua•8/12/22, 12:23 AM

Needs more

AAqua So have 24000 overloads for that one function now?

R

Redhacker2•8/12/22, 12:23 AM

no

SSpider-Kyle you know,i would actually love to doc silk.net if i actually knew what 99% of th...

P

Perksey•8/12/22, 12:24 AM

ftr we don't expect anyone to document the generated functions

R

Redhacker2•8/12/22, 12:24 AM

2 overloads per function (with the original function overloads)

A

Aqua•8/12/22, 12:24 AM

Ah
Right

R

Redhacker2•8/12/22, 12:24 AM

Fully safe, fully unsafe no more

R

Redhacker2•8/12/22, 12:25 AM

the question is how should the safe stuff look

R

Redhacker2•8/12/22, 12:25 AM

like you could have multiple "safe" overloads

R

Redhacker2•8/12/22, 12:25 AM

which should you generate

A

Aqua•8/12/22, 12:26 AM

Only thing we need now is

using static Silk.NET.INSANITY;

public class MyClass {

    public void thing() {
        glGenBuffer(GL_VERTEX_BUFFER);
    }

}

using static Silk.NET.INSANITY;

public class MyClass {

    public void thing() {
        glGenBuffer(GL_VERTEX_BUFFER);
    }

}

using static Silk.NET.INSANITY;

public class MyClass {

    public void thing() {
        glGenBuffer(GL_VERTEX_BUFFER);
    }

}

using static Silk.NET.INSANITY;

public class MyClass {

    public void thing() {
        glGenBuffer(GL_VERTEX_BUFFER);
    }

}

Etc

A

Aqua•8/12/22, 12:26 AM

I screwed that up

RRedhacker2 the question is how should the safe stuff look

P

Perksey•8/12/22, 12:26 AM

well that's why i was thinking of the

PointerLike

PointerLike

PointerLike

PointerLike stuff

P

Perksey•8/12/22, 12:27 AM

one parameter type that multiple types implicitly convert to

P

Perksey•8/12/22, 12:27 AM

the main issue i can already see is ref lifetime though and implicit ref conversions

PPerksey well that's why i was thinking of the `PointerLike` stuff

R

Redhacker2•8/12/22, 12:27 AM

hmm

R

Redhacker2•8/12/22, 12:27 AM

interesting

P

Perksey•8/12/22, 12:27 AM

actually ref lifetimes would be easy

P

Perksey•8/12/22, 12:27 AM

but the implicit conversion of a ref to a type is i think impossible

P

Perksey•8/12/22, 12:27 AM

so it'd have to be

new(ref x)

new(ref x)

new(ref x)

new(ref x) instead of

ref x

ref x

ref x

ref x

A

Aqua•8/12/22, 12:28 AM

Lifetimes? I thought that was only a rust thing

AAqua Lifetimes? I thought that was only a rust thing

R

Redhacker2•8/12/22, 12:28 AM

most certainly not

P

Perksey•8/12/22, 12:28 AM

rust is just the thing that makes lifetimes explicit

P

Perksey•8/12/22, 12:28 AM

you have implicit lifetimes in C#

P

Perksey•8/12/22, 12:28 AM

you also have scary things called "GC holes" as well

R

Redhacker2•8/12/22, 12:28 AM

it is a thing in all languages, some do better at hiding it

A

Aqua•8/12/22, 12:28 AM

I see

R

Redhacker2•8/12/22, 12:29 AM

RAII for instance in C++

P

Perksey•8/12/22, 12:29 AM

int* x = default;
{
    int y;
    x = &y;
}
// y is out of scope thus x's "lifetime" has been exceeded

int* x = default;
{
    int y;
    x = &y;
}
// y is out of scope thus x's "lifetime" has been exceeded

int* x = default;
{
    int y;
    x = &y;
}
// y is out of scope thus x's "lifetime" has been exceeded

int* x = default;
{
    int y;
    x = &y;
}
// y is out of scope thus x's "lifetime" has been exceeded

P

Perksey•8/12/22, 12:30 AM

same with

Span<int> x = default;
{
    int y = default;
    x = MemoryMarshal.AsSpan(ref y, 1);
}
// y is out of scope thus x's "lifetime" has been exceeded

Span<int> x = default;
{
    int y = default;
    x = MemoryMarshal.AsSpan(ref y, 1);
}
// y is out of scope thus x's "lifetime" has been exceeded

Span<int> x = default;
{
    int y = default;
    x = MemoryMarshal.AsSpan(ref y, 1);
}
// y is out of scope thus x's "lifetime" has been exceeded

Span<int> x = default;
{
    int y = default;
    x = MemoryMarshal.AsSpan(ref y, 1);
}
// y is out of scope thus x's "lifetime" has been exceeded

P

Perksey•8/12/22, 12:30 AM

though i realised the easy solution would be just make

PointerLike

PointerLike

PointerLike

PointerLike itself a

ref struct

ref struct

ref struct

ref struct that can track its own refs

P

Perksey•8/12/22, 12:35 AM

ah yep it's illegal https://sharplab.io/#v2:EYLgxg9gTgpgtADwGwBYA0AXEBDAzgWwB8ABAJgEYBYAKGIAYACY8gOgCUBXAOwwEt8YLAJI8YUCAAcAymIBuvMDFwBuGvSatOPfoIDCEfBN4AbMTKjzFKmmoDMDWADMGuDFA5gMDAAoReoqAAZXgBrGAAeABUAPgYAdwALMRgGSJAGbnxsLmwAcxgAExoAbxoGcoYJKF5ZbAwUqQlsqNiAfVwmrlVqCqZ7ZiQGGAQJYwVeL0kxOugfPwDgsJaACidUhgQASjLeioBeWK4YOLn/eqDQiJiGYoZ2zoY9hgBZGHxoAE9n7ChcBOxjCxdLA6jBGtlVjBnAg0AxyJsGABfbqImhAA===

SharpLab

C#/VB/F# compiler playground.

KKai Please use an data source instead of the foor loops and use a XUnit theory

R

Redhacker2•8/12/22, 12:58 AM

question, do you want me to write an IEnumerable<(int,int,int, int)>, if so is the behavior to be like the previous for loop, and I imagine you would want both tests to be done in such a manner?

L

Lyris the Kitori•8/12/22, 1:56 AM

okay got everything written down for live backend switching
https://github.com/orgs/Furball-Engine/projects/1/views/1
so hopefully without needing a restart i can switch between OpenGL, GLES, Veldrid, DX11, etc.

L

Lyris the Kitori•8/12/22, 1:56 AM

gonna chew through this one by one

L

Lyris the Kitori•8/12/22, 3:45 AM

ah great version number https://i.beyleyisnot.moe/68oAM2g.png