Crossplay with pokopia-native: protocol comparison and options (research)

Đang mở
  • Có gì không ổn?
#505 0 bình luận 0 reaction 0 người được giao Xem trên GitHub

Chưa có ai nhận issue này.

Đánh giá

Độ khó
5/5
Thời gian dự kiến
Hơn một tuần
Mức phù hợp với người mới
20/100
Loại issue
Tài liệu
Độ rõ ràng
Cần làm rõ
Mức độ hoạt động
Sôi nổi
Công nghệ
cpp
Lĩnh vực
networking

Hướng nghiên cứu

Start by reading the issue’s comparison of the Link Play stacks and its A/B/C options; it identifies no repository files or tests to inspect. The work is to decide whether crossplay is wanted and select an approach, after resolving the listed protocol, identity, security, discovery, and licensing questions.

Do mô hình lập chỉ mục viết ra từ nội dung của issue.

Mô tả

in progress multiplayer tracking

Read-only comparison of the Link Play ("Local connection") stacks in Dittobyte/Recompia and xFractals/pokopia-native (MIT, clone at 4a72d23, 1.3.0), and what crossplay would take.

Bottom line

The two games would interoperate at the game level (same game, same nn::ldn HLE, same LocalCommunicationVersion 301, same virtual UDP socket on port 49530), but the transport, discovery and identity/auth layers are entirely different designs. Crossplay today is impossible without one side implementing the other's protocol (A), a bridge (B), or an agreed shared protocol (C). The room-code format is already shared; the key derivation is not.

What matches, and what does not
  • Matches: nn::ldn emulation and the NetworkInfo/NodeInfo the game sees; the LocalCommunicationVersion; the virtual UDP socket on port 49530.
  • Differs: transport and message framing, LAN discovery, and the join/identity checks. Their build has a build-ID + mods-fingerprint gate with no analogue here; this side's session-variant has none there.
  • Room code: same format, different KDF ("pokopia-relay room v1" vs "pokopia-coop room v1").
  • Security: their internet play authenticates every frame/datagram (HMAC, replay-protected); this side authenticates only the join and trusts the post-join data path.
Options

A. Our runtime speaks their protocol as a second Link Play mode. Needs their wire protocol end to end, a Linux mDNS/DNS-SD backend (theirs has none), their identity (game main build ID + mods fingerprint), and their room-code KDF. Cost: high — effectively a parallel transport stack (~3–4k lines). Risk: their Link Play is unreleased and may change.

B. A bridge process. A helper that speaks this relay protocol on one side and their coop protocol on the other, mapping one room to one session and forwarding NetworkInfo + game datagrams, and answering both identity checks. Cost: moderate; reuses both stacks. Adds a hop and a support burden, and needs a NAT/mDNS answer.

C. A shared protocol agreed by both projects. Both adopt one versioned discovery + transport + auth format. Highest coordination, lowest long-term duplication. Their Link Play is unreleased and their roadmap is their own.

Risks
  1. Their Link Play is "not released"; protocol, ports, DNS-SD keys and KDF may change.
  2. Room-code KDF mismatch: same code, different keys.
  3. No shared identity gate: a compatible mode must define a common compat gate or drop their check (weakening their cheat refusal).
  4. Security asymmetry: pick one threat model.
  5. License boundary: MIT is GPL-compatible, but needs attribution and a decision on vendoring vs reimplementing.
  6. mDNS on Linux: any option that uses their discovery needs a Linux backend.
Open questions
  • Is crossplay with pokopia-native wanted at all, and at what priority (LAN first, internet later)?
  • Is MIT code in the GPL tree acceptable (vendoring their coop module) vs a clean-room reimplementation vs a bridge?
  • Is the game's main build ID available at runtime to supply their identity check, and would the mods-fingerprint scheme be adopted or replaced with the session-variant scheme?
  • Which room-code KDF should win?
  • Preferred option A / B / C, given the runtime is Windows-built but runs under Proton and any added Linux mDNS work lands in code that cannot be built locally today.
Ngôn ngữ chính
C++
Star
224
Fork
0
Không có pull request nào được merge trong 30 ngày

Chuẩn bị môi trường

Bắt đầu từ đâu

  1. Đọc hết issue, rồi đọc hướng dẫn đóng góp của dự án.
  2. Bình luận trên issue rằng bạn sẽ nhận — tránh hai người làm cùng một việc.
  3. Fork repository và làm thay đổi trên một nhánh.
  4. Mở pull request có tham chiếu số hiệu của issue.

Issue khác của featureJosh/Recompia

Tất cả issue của featureJosh/Recompia

Issue tương tự

Thêm issue về C++

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.