Allow obtaining custom implementation from wgpu api types

[sagudev]

Connections
Fixes #7533

Description
When implementing custom beckend, one was not be able to extract actual custom type from descriptors. This is now possible by introducing as_custom method on each wgpu API type (and Dispatch* type) that are analogues to as_hal. They try to return actual concrete custom type that is stored inside custom backend and return None if backend is not custom (or user wanted wrong T).

Testing
Example is modified to cover obtaining original custom type from dynamic custom backend type.

Squash or Rebase? Squash, but it's reviewable per commit.

Checklist

• Run cargo fmt.
• Run taplo format.
• Run cargo clippy --tests. If applicable, add:
  • --target wasm32-unknown-unknown
• Run cargo xtask test to run tests.
• If this contains user-facing changes, add a CHANGELOG.md entry.

[raphaelhetzel]

I already use the changes in this pr to run the compute hello_world example inside wasmtime (custom backend passing the calls to wgpu running on the host).
So i can confirm that this is needed and useful.
Thanks @sagudev.

[cwfitzgerald]

Small note but otherwise fine. Just wanted to clue you in I would like to have a unified interface for as_hal, as_custom, and a potential future as_webgpu. I'm not sure exactly what this would look like, but I think having a single entry point would probably good for api consistency. This is fine to have for the mean time however.