From e261b88a7304091b07ae0ced6ddac088e91dc8b5 Mon Sep 17 00:00:00 2001 From: Dhruv Dhamani Date: Fri, 13 Nov 2020 02:38:26 -0500 Subject: [PATCH] Update 001-resource-binding-syntax.md (#236) small typo --- docs/src/rfcs/001-resource-binding-syntax.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/src/rfcs/001-resource-binding-syntax.md b/docs/src/rfcs/001-resource-binding-syntax.md index f026bfb0a7..dcbd8f3edd 100644 --- a/docs/src/rfcs/001-resource-binding-syntax.md +++ b/docs/src/rfcs/001-resource-binding-syntax.md @@ -163,7 +163,7 @@ fn main(inputs: &ShadingInputs, indirect_lighting: &IndirectLighting) { # Suggestion -I think the most ergonomic and future proof binding method would be to have descriptors in structs, bound to the entrypoint. This allows us some nice, even more ergonomic upsides later on (when support is more widely available) where we can put data members in these structs as well. And along wit this, we can have very egonomic CPU side code as well, where we can keep shader invocation looking like a function call for a large part, instead of having to manually bind to slots again. +I think the most ergonomic and future proof binding method would be to have descriptors in structs, bound to the entrypoint. This allows us some nice, even more ergonomic upsides later on (when support is more widely available) where we can put data members in these structs as well. And along with this, we can have very egonomic CPU side code as well, where we can keep shader invocation looking like a function call for a large part, instead of having to manually bind to slots again. # Prior art @@ -238,4 +238,4 @@ RLSL also passes resource bindings as arguments to the kernel, with two template #[spirv(compute)] fn compute(compute: Compute, buffer: Buffer>) ``` - \ No newline at end of file +