Refactor native video decoder for code reuse

[emilk]

What

• Part of Native H.264 #7606

This refactors the native decoder so that we can easily slot in a different native decoder (ffmpeg!).

The innermost trait is now SyncDecoder which gets pushed a chunk, and blocks while producing frames (or errors). Around that is an AsyncDecoder that runs the SyncDecoder on a background thread to produce a non-blocking interface.
Finally in re_renderer there is the NativeDecoder that wraps SyncDecoder and handles texture uploads.

There is a lot of code moved, but very little code actually changed. It's just another layer of abstraction introduced.

Checklist

• I have read and agree to Contributor Guide and the Code of Conduct
• I've included a screenshot or gif (if applicable)
• I have tested the web demo (if applicable):
  • Using examples from latest main build: rerun.io/viewer
  • Using full set of examples from nightly build: rerun.io/viewer
• The PR title and labels are set such as to maximize their usefulness for the next release's CHANGELOG
• If applicable, add a new check to the release checklist!
• If have noted any breaking changes to the log API in CHANGELOG.md and the migration guide

• PR Build Summary
• Recent benchmark results
• Wasm size tracking

To run all checks from main, comment on the PR with @rerun-bot full-check.

[Wumpf]

well threaded in 😁
We're starting to see a nice hierarchy of decoder wrangling logic. Let's hope we don't regret this and have to add holes through these nice abstractions, but so far I'm very optimistic on that!

[Wumpf]

found it odd at first that all NativeDecoders host a SyncDecoder internally. But from the looks of it this gonna hold up at least for a a few weeks 😁 this is what we're dealing with :)