Hi guys, I'm new to Silk.NET, can anybody give me an advice on the following. I want to create a ver

Hi guys, I'm new to Silk.NET, can anybody give me an advice on the following. I want to create a very simple cross platform desktop 2D app which is to draw 2D primitives: lines, points, rectangles, polygons, circles. Which is the simplest approach to develop such an app with Silk.NET? I've spent quite time trying to find documentation or examples but haven't find anything helpful. I've already learned that Silk.NET provides bindings for SDL, GLFW and brings an abstract layer over it, but I haven't found any example on how to use any of these. So I'm confused about what is the best approach for in my case. Should I use SDL or GLFW? Does Silk.NET provide some higher level API to draw 2D primitives or I need to work directly on SDL or GLFW level? Or maybe it would be simpler to use OpenGL and use it for rendering 2D primitives? Any advice or code example is greatly appreciated!

Rrobert11 Hi guys, I'm new to Silk.NET, can anybody give me an advice on the following. I ...

Lyris the Kitori•4/12/22, 6:36 AM

i dont think GLFW even has the ability to draw primitives, and our SDL bindings dont have any of the renderer functions, the only thing our SDL bindings are used for is windowing (and thats all our GLFW abstraction does too, it only creates/manages windows), you'll have to use OpenGL/VK/D3D for rendering

Perksey•4/12/22, 6:39 AM

there are examples of how to use our high level windowing API in the examples folder

Perksey•4/12/22, 6:39 AM

these examples also show you how to render basic 2D quads using OpenGL

PPerksey these examples also show you how to render basic 2D quads using OpenGL

robert11OP•4/12/22, 7:05 AM

Ok, got it. Then I will try with OpenGL. Thanks!

LLyris the Kitori i dont think GLFW even has the ability to draw primitives, and our SDL bindings ...

robert11OP•4/12/22, 7:13 AM

I see, thanks for the explanation! Though, I'm a bit surprised because I've discovered through intellisense some methods that really looks like rendering bindings:

robert11OP•4/12/22, 7:13 AM

Here it is

Lyris the Kitori•4/12/22, 7:14 AM

oh i guess they got added

Lyris the Kitori•4/12/22, 7:14 AM

must have been when our generator got fixed

Redhacker2•4/12/22, 8:57 PM

BTW anyone know what the hell happened with that

Redhacker2•4/12/22, 8:57 PM

All of a sudden constants got added and

Redhacker2•4/12/22, 8:57 PM

was just fucking weird

Lyris the Kitori•4/12/22, 8:58 PM

i think perskey was messing with the generator

Lyris the Kitori•4/12/22, 8:58 PM

and it just generated a bunch more after a random tweak

Redhacker2•4/12/22, 8:59 PM

I know

Redhacker2•4/12/22, 8:59 PM

He pinged me LMAO

Redhacker2•4/12/22, 9:00 PM

Saying now the constants I asked about for ASSIMP were generated

Redhacker2•4/12/22, 9:00 PM

And that he had no idea what he changed to do that

RRedhacker2 And that he had no idea what he changed to do that

Perksey•4/12/22, 9:24 PM

this is still true

Perksey•4/12/22, 9:24 PM

no fucking idea how

Perksey•4/12/22, 9:24 PM

I wasn’t even touching that part of the generator

Perksey•4/12/22, 9:24 PM

and had entirely unrelated objectives

Perksey•4/12/22, 9:25 PM

and suddenly… hello constants!

Lyris the Kitori•4/12/22, 9:25 PM

i think you are just so good at programming that the sheer aura of your code scares other code into working

Perksey•4/12/22, 9:25 PM

more like the sheer stench

Redhacker2•4/12/22, 9:34 PM

LOL Silktouch needs a fresh coat of fire

Kai•4/12/22, 10:01 PM

working on that

unit test all the things

Alexx999•4/13/22, 1:45 AM

Hi all,
I've noticed that OpenCL bindings are pretty much broken due to the fact that some enums need to be 64 bit, which leads to methods such as i.e. clGetDeviceIDs being toast due to wrong argument size

I swooped in hoping for an easy fix but oh boy was this rabbit hole deep

Perksey•4/13/22, 8:47 AM

@alexx999_ can you open an issue and I'll look into this for the May update

PPerksey @alexx999_ can you open an issue and I'll look into this for the May update

Alexx999•4/13/22, 8:50 AM

well, I still plan to submit a PR

an easy dirty fix would be to typemap particular types to ulong
and the hard way is to group enums first and typemap to correct base enum

Perksey•4/13/22, 8:51 AM

In fact if I remember correctly that’s the only way without reinventing the OpenCL converter and/or adding additional metadata

Perksey•4/13/22, 8:51 AM

from what I remember the cl.xml was pretty crap

Perksey•4/13/22, 8:52 AM

but I do welcome your enthusiasm to contribute! It is truly appreciated and a rare occurrence for us

PPerksey from what I remember the cl.xml was pretty crap

Alexx999•4/13/22, 8:57 AM

actually it's OK now aside from couple typos and oddities, in fact I've been able to get the enums "grouped", correctly marked as deprecated and correctly marked as having ulong base types

Perksey•4/13/22, 8:58 AM

Nice! Thank you so much!

Perksey•4/13/22, 8:59 AM

then I’m guessing this’ll be the biggest improvement to our CL bindings in a while

Perksey•4/13/22, 8:59 AM

@kaijellinghaus look at that someone other than me can navigate the mess that is BuildTools!

Kai•4/13/22, 9:03 AM

Honestly impressed

PPerksey @kaijellinghaus look at that someone other than me can navigate the mess that is...

Alexx999•4/13/22, 9:52 AM

yeah, tbh I've started typing a message saying that "I can see why there's not too many contributions"

Alexx999•4/13/22, 9:53 AM

that is, because no sane person would want to figure out that LINQ spaghetti
too bad that I love spaghetti... and LINQ...

Alexx999•4/13/22, 10:26 AM

there's still some open questions, especially around that "extensions" separation (those are not shipped, right?)
but I guess that can be left as-is for now

now I need to wrangle the typemap to point to the new enums and probably even try to fix those stupid tagged arg lists

Kai•4/13/22, 10:35 AM

You mentioned you marked the old enums as deprecated? That's probably unnecessary. If any code that used them was broken anyways its fine to just break

KKai You mentioned you marked the old enums as deprecated? That's probably unnecessar...

Alexx999•4/13/22, 10:50 AM

actually... I fixed entire stack so it's marked as obsolete in the output files even with correct "obsolete in" versions

Alexx999•4/13/22, 10:50 AM

https://github.com/dotnet/Silk.NET/pull/881

GitHub

Impelement OpenCL enum generation by Alexx999 · Pull Request #881 ·...

This should add some basic enum generation support for OpenCL aside from the umbrella CLEnum type.
The initial motivation is that some of the enum types have to be ulong-based causing OpenCL API ca...

Alexx999•4/13/22, 10:50 AM

there you go

Alexx999•4/13/22, 10:57 AM

there's some room for improvement but I guess this should do as a starting point

Alexx999•4/13/22, 10:58 AM

it's always better to make small PRs

AAlexx999 yeah, tbh I've started typing a message saying that "I can see why there's not t...

Perksey•4/13/22, 1:19 PM

haha, the rest of silk.net isn't that bad and we do get contributions there

Perksey•4/13/22, 1:19 PM

but BuildTools is the darkest corner of the codebase

Perksey•4/13/22, 1:19 PM

the approach in early Silk.NET history was literally "if it works then it's fine, insta merge"

PPerksey the approach in early Silk.NET history was literally "if it works then it's fine...

Alexx999•4/13/22, 1:20 PM

usual story, that's what the word "legacy code" means