Improve accuracy of frames rendered bookkeeping

[sbooth]

No description provided.

[Copilot]

Pull request overview

This PR improves the accuracy of “frames rendered” bookkeeping by pairing decoded audio chunks with explicit per-chunk metadata (sequence number, frame position/length, playback generation) and using that metadata to emit more precise render events.

Changes:

• Add a dedicated SPSC metadata ring buffer (audioMetadata_) and new chunk descriptor structs to track which decoder produced which rendered frames.
• Update seek handling to return an optional frame position and incorporate a playback generation counter to discard stale render events after seeks/cancellations.
• Rename/clarify buffer constants and add the CXXRingBuffer dependency to support metadata buffering.

Reviewed changes

Copilot reviewed 3 out of 4 changed files in this pull request and generated 4 comments.

File	Description
Sources/CSFBAudioEngine/Player/AudioPlayer.mm	Implements dual ring-buffer (audio + metadata) flow, generation-based stale-event rejection, and updated seek/render bookkeeping.
Sources/CSFBAudioEngine/Player/AudioPlayer.h	Adds decoded/rendering chunk descriptor types and new members (audioMetadata_, playbackGeneration_, renderingChunk_).
Package.swift	Adds CXXRingBuffer as a SwiftPM dependency for the metadata ring buffer.
Package.resolved	Pins CXXRingBuffer to v0.6.1.

---

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

[Copilot]

Pull request overview

Copilot reviewed 3 out of 4 changed files in this pull request and generated 4 comments.

Comments suppressed due to low confidence (1)

Sources/CSFBAudioEngine/Player/AudioPlayer.mm:1535

• If audioMetadata_.push(descriptor) fails, the code currently logs but continues writing audio. That can permanently desynchronize audio frames from their metadata (sequence number/generation), breaking frames-rendered bookkeeping. If the metadata push fails, skip writing audio for that chunk and force a drain/new generation so render state stays consistent.

                    // Write the decoded chunk descriptor to the metadata buffer
                    if (!audioMetadata_.push(descriptor)) {
                        os_log_fault(log_, "Error writing chunk descriptor: spsc::Queue::push failed");
                    }