How to map this wgsl code to rust-gpu code?

[LollipopFt]

Hi, I am new to shaders and I have been following along Learn WGPU tutorial, and am currently at Buffers and Indices. The current wgsl code that I am trying to port over is:

struct VertexInput {
  @location(0) position: vec3<f32>,
  @location(1) color: vec3<f32>,
};

struct VertexOutput {
  @builtin(position) clip_position: vec4<f32>,
  @location(0) color: vec3<f32>,
};

@vertex
fn vs_main(
  model: VertexInput,
) -> VertexOutput {
  var out: VertexOutput;
  out.color = model.color;
  out.clip_position = vec4<f32>(model.position, 1.0);
  return out;
}

@fragment
fn fs_main(in: VertexOutput) -> @location(0) vec4<f32> {
  return vec4<f32>(in.color, 1.0);
}

I cannot seem to link the content in the vertex buffer to vs_main, and as a result I have been unable to test if fs_main is correct either. Below is one of my rust-gpu code attempts:

#![no_std]
use spirv_std::glam::{Vec3, Vec4};
use spirv_std::spirv;

/*#[repr(C)]
#[derive(Copy, Clone, Debug, bytemuck::Pod, bytemuck::Zeroable)]
pub struct Vertex {
  position: [f32; 3],
  color: [f32; 3]
}*/

#[spirv(vertex)]
pub fn vs_main(
  #[spirv(descriptor_set = 0, binding = 0)] position: &[f32],
  #[spirv(descriptor_set = 0, binding = 1)] color: &[f32],
  #[spirv(position, invariant)] out_pos: &mut Vec4,
  out_color: &mut Vec3
) {
  let color: [f32; 3] = color.try_into().unwrap();
  *out_color = color.into();
  let pos: [f32; 3] = position.try_into().unwrap();
  let pos: Vec3 = pos.into();
  *out_pos = pos.extend(1.0);
}

#[spirv(fragment)]
pub fn fs_main(
  pos: Vec4,
  color: Vec3,
  out: &mut Vec4
) {
  *out = color.extend(1.0);
}

which gives error: invalid entry param type &[f32](expected[f32]or&mut [f32]). What is the correct way to solve this?

I had previously done the example in the earlier chapter 'The Pipeline', and successfully converted the code over.

code

#![no_std]
use spirv_std::glam::{Vec3, Vec4, Vec4Swizzles};
use spirv_std::spirv;

#[spirv(vertex)]
pub fn vs_main(
  #[spirv(vertex_index)] vtx_idx: i32,
  #[spirv(position)] clip_pos: &mut Vec4,
  vert_pos: &mut Vec3
) {
  let x = (1 - vtx_idx) as f32 * 0.5;
  let y = ((vtx_idx & 1) * 2 - 1) as f32 * 0.5;
  *clip_pos = Vec4::new(x, y, 0., 1.);
  *vert_pos = clip_pos.xyz();
}

#[spirv(fragment)]
pub fn fs_main(
  _clip_pos: Vec4,
  _vert_pos: Vec3,
  out: &mut Vec4
) {
  *out = Vec4::new(0.3, 0.2, 0.1, 1.);
}

[nazar-pc]

Essentially, with current restrictions of rust-gpu you'll not be able to convert between slices and arrays in either direction. This is supposed to change at some point, but not yet.

Similarly, you'll have a hard time using iterator trait for anything moderately complex, so will need to do manual indexing like in C/C++ quite often.

[Firestar99]

Yep, for now you'll just have to [position[i], position[i+1], position[i+2]], core::array::from_fn also doesn't work

[LollipopFt]

i see, thanks for both your help!