Why derive from a mature base?
A Wayland compositor is not just “where windows go”. It includes DRM/KMS, input, protocols, XWayland, outputs, rendering, compatibility and an impressive number of ways to ruin a Saturday.
Felunyx authorship does not depend on rebuilding all of it. It lives in workspace, layout, stack, focus, rule, restoration and system-integration policy.
What stays close to upstream
Generic infrastructure should stay as close as practical to the mature foundation. Felunyx keeps its delta measurable and sends generic fixes upstream when that is sustainable.
- hardware and DRM/KMS;
- input/render/output plumbing;
- standard Wayland protocols;
- XWayland lifecycle;
- generic compatibility.
What remains Felunyx
Floating, Snap, Tiling, Rolling, Monocle, stacks, dynamic workspaces, deterministic transitions, rules and restoration remain Felunyx policy. The shell is still planned in Qt/QML with state and policy outside presentation.
And cosmic-comp?
`pop-os/cosmic-comp` is the preferred first prototype foundation because it preserves the Rust + Smithay lineage and already carries mature compositor infrastructure.
That does not mean “Felunyx uses COSMIC”. Phase 8 must prove that useful infrastructure can be separated from unnecessary COSMIC-specific policy and services at a sustainable sync cost. If not, the foundation decision remains open.
Current status
This is a Phase 8 prototype strategy, not a shipped desktop. KDE remains the initial reference and the native desktop does not block the first useful distro release.