core: Add Scout-compatible FLM telemetry system

[gpeacock]

Description

Adds an opt-in telemetry backend to Ruffle core that records frame, render, input, network, and memory metrics and serializes them in Adobe Scout's FLM (AMF3-based) format. Desktop builds can also stream telemetry live over TCP to a running Scout instance while still buffering the full session for write_flm on exit.

• New core/src/telemetry.rs: span/counter recording and FLM serialization
• desktop: --telemetry CLI flag to opt in and configure output/streaming
• core: player/context wiring to record frame, render (including bitmap decompression as rend.buildbits), input, network, and memory spans
• Documentation and a Python FLM reader/decoder under tests/tests/telemetry/ to aid verification and debugging

See tests/tests/telemetry/TELEMETRY_USAGE.md for full usage details.

Testing

Added tests/tests/telemetry/test_complete.rs and test_spans_gen.rs, which exercise recording a full session and validate the emitted .flm output. amf3reader.py / telemetry.py provide an independent Python-based decoder used to cross-check the binary format against Scout's expectations.

Locally verified: cargo check -p ruffle_core, cargo fmt --check -p ruffle_core, and cargo clippy -p ruffle_core -- -D warnings all pass clean.

Checklist

• I, a human, have self-reviewed this PR and fully understand the changes within.
• I have made or updated tests where possible.
• All of my commits are properly scoped, compile successfully, and pass all tests.
• This PR does not make sense to split up into smaller PRs.
• An LLM was involved in the authoring of this code.