Allow savestate creation/loading during netplay (in-memory) #2

New Issue

pbwraith · 2026-08-25T06:03:44+02:00

pbwraith commented

2026-08-25 06:03:44 +02:00

Repo: gopher64-netplay (fork of https://github.com/gopher64/gopher64)
Context: Full scope doc at upstream analysis — see commit history. Netplay uses GGRS over WebRTC (matchbox_socket), reliable channel carries NetplayMessage { name, data } in 16KB chunks. Savestates serialize full Device via postcard + compress.

Goal: Enable savestate ops while a netplay session is active, without disk I/O.

Background:

src/savestates.rs create_savestate does disk write in a spawned tokio task; load_savestate reads compressed file
src/device/vi.rs:103 guards process_savestates with if device.netplay.is_none() — states are refused during netplay
process_requests in netplay.rs already receives GGRS SaveGameState/LoadGameState requests but stubs them (lines 271-285)

Acceptance:

Add SavestateData::save_to_memory() -> Vec<u8> (postcard + compress, no disk) and SavestateData::load_from_memory(&[u8]) (decompress + deserialize)
Add create_savestate_in_memory(device) -> Option<SavestateData> — clones device like create_savestate does, returns data instead of writing
Add load_savestate_from_data(device, &SavestateData) — refactor the existing std::mem::swap block in load_savestate into a shared function both paths use
The device-clone statics (DEVICE_CLONE, SAVES_CLONE) must be reused, not duplicated
Remove/bypass the vi.rs:103 guard ONLY for the netplay-savestate path (keep offline behavior identical)
Tests: create → load round-trip via memory, verify Device state matches (sample: cpu regs, rdram bytes, vi regs). Test during an active (mocked) netplay session.

Verification: cargo test passes. Offline savestate behavior unchanged.

**Repo:** gopher64-netplay (fork of https://github.com/gopher64/gopher64) **Context:** Full scope doc at upstream analysis — see commit history. Netplay uses GGRS over WebRTC (matchbox_socket), reliable channel carries `NetplayMessage { name, data }` in 16KB chunks. Savestates serialize full Device via postcard + compress. **Goal:** Enable savestate ops while a netplay session is active, without disk I/O. **Background:** - `src/savestates.rs` `create_savestate` does disk write in a spawned tokio task; `load_savestate` reads compressed file - `src/device/vi.rs:103` guards `process_savestates` with `if device.netplay.is_none()` — states are refused during netplay - `process_requests` in netplay.rs already receives GGRS `SaveGameState`/`LoadGameState` requests but stubs them (lines 271-285) **Acceptance:** - Add `SavestateData::save_to_memory() -> Vec<u8>` (postcard + compress, no disk) and `SavestateData::load_from_memory(&[u8])` (decompress + deserialize) - Add `create_savestate_in_memory(device) -> Option<SavestateData>` — clones device like `create_savestate` does, returns data instead of writing - Add `load_savestate_from_data(device, &SavestateData)` — refactor the existing `std::mem::swap` block in `load_savestate` into a shared function both paths use - The device-clone statics (`DEVICE_CLONE`, `SAVES_CLONE`) must be reused, not duplicated - Remove/bypass the `vi.rs:103` guard ONLY for the netplay-savestate path (keep offline behavior identical) - Tests: create → load round-trip via memory, verify `Device` state matches (sample: cpu regs, rdram bytes, vi regs). Test during an active (mocked) netplay session. **Verification:** `cargo test` passes. Offline savestate behavior unchanged.

pbwraith referenced this issue

2026-08-25 06:03:45 +02:00

Session save — capture netplay session state to disk (host) #3

pbwraith referenced this issue

2026-08-25 06:03:45 +02:00

Session resume — reload state and broadcast to peers (host+join) #4

pbwraith referenced this issue

2026-08-25 06:03:46 +02:00

Desync auto-resync — host sends state on desync event #6

pbwraith referenced this issue

2026-08-25 06:03:46 +02:00

Desync auto-resync — client receives and loads state #7