Make output buffer copying optional

[mkeeter]

This allows us to more tightly integrate with complex GPU pipelines.

[Copilot]

Pull request overview

This PR refactors the fidget-wgpu voxel rendering API to make the GPU→CPU readback/copy step optional, enabling tighter integration with GPU-only pipelines by allowing callers to skip producing a mapped CPU buffer.

Changes:

• Split “render buffers” from “readback buffers” by introducing ImageReadBuffer and adding Context::image_buffer(&Buffers).
• Update Context::{run, run_with_vars, run_async, submit, map_image} to take &Buffers plus an explicit &mut ImageReadBuffer (or Option<&mut ImageReadBuffer> for submit*), so output copying can be omitted.
• Adjust buffer sizing checks in GenericFlexBuffer and update tests/demo CLI to the new calling pattern.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.

File	Description
fidget/tests/voxel_render.rs	Updates voxel rendering tests to use Buffers + ImageReadBuffer.
fidget-wgpu/src/voxel.rs	Introduces ImageReadBuffer, makes submit/readback optional, and updates mapping/readback flow.
fidget-wgpu/src/lib.rs	Refactors GenericFlexBuffer size checking to validate from item counts consistently.
demos/cli/src/main.rs	Updates CLI demo to use submit(..., Some(&mut out), ...) and map from ImageReadBuffer.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.