Ray March
Ray March
Naga seems to treat `gl_VertexIndex` as a `uint` when the spec says it is an `int`. The following vertex shader compiles successfully on other shader compilers. ```glsl #version 450 void...
Hi @aldeka, Thank you so much for including my ferris 3d model on your site! 😊 I had lots of fun making it! I noticed that clicking on the image...
Right now, the WGSL storage layout is the default layout applied to structs when `#[derive(GpuLayout)]` is used. For consistency, and to allow maybe adding a `#[gpu_repr(uniform)]` in the future, the...
Replace the `Index` and `IndexMut` implementations of `PoolRefMut` with fallible indexing, so that no panics can ever happen from a generation check failing. That way we can be more confident...
the trait `sm::CpuAligned` is not public and not implemented for `!Sized + sm::CpuLayout` structs. as of now, the following code doesn't compile ```rust #[derive(sm::CpuLayout)] #[repr(C)] struct Unsized { a: f32,...
create an example that shows the usage of `shame_wgpu`'s `bind_group!` macro.
There should be an example crate that shows how to set up hot-reloading. Maybe even 2 crates - [ ] hot reloading via a separate binary crate that sends the...
the `WorkGroup::thread_id` docs say ```rust /// the 1-dimensional numeric index of a thread within its workgroup. /// /// It lies in the range `shame::zero()..self.thread_grid_size` /// /// this value is different...
Wgsl has subgroup operations now https://www.w3.org/TR/WGSL/#subgroup-ops we should add them to the `shame` api ### Steps - [ ] define expr enum `SubgroupOperations` containing all the functions - [ ]...
Wgsl has quad operations now https://www.w3.org/TR/WGSL/#quad-builtin-functions we should add them to the `shame` api ### Steps - [ ] define expr enum `QuadOperations` containing all the functions as variants -...