From 8a97277f206db5ddbe6b287de19f5c6169693600 Mon Sep 17 00:00:00 2001
From: iris <iris@hyperhive>
Date: Sun, 31 May 2026 22:19:26 +0200
Subject: [PATCH] fix(#540): pin weston socket + inject WAYLAND_DISPLAY
 globally

Services started by systemd in a gui-enabled container didn't have
WAYLAND_DISPLAY set, so Wayland clients couldn't find the compositor.
libwayland would fall back to a headless display or error out, leaving
apps running invisibly while the VNC session showed a blank weston
desktop (the double-screen problem).

Fix in weston-vnc.nix:
- Pass --socket=wayland-0 to weston so the socket name is
  deterministic (weston normally picks any free wayland-N name).
- Set WAYLAND_DISPLAY=wayland-0 and XDG_RUNTIME_DIR=/run/user/0 as
  global environment.variables gated on hyperhive.gui.enable, so
  every service in the container inherits them automatically.
- Update gui.json to include wayland_display for tooling that reads it.

Update docs/gotchas.md with the rationale and pointer to #540.
---
 docs/gotchas.md              | 14 ++++++++++++--
 nix/templates/weston-vnc.nix | 29 +++++++++++++++++++++++++++--
 2 files changed, 39 insertions(+), 4 deletions(-)

diff --git a/docs/gotchas.md b/docs/gotchas.md
index 8fa69869..22c2a766 100644
--- a/docs/gotchas.md
+++ b/docs/gotchas.md
@@ -264,9 +264,19 @@ connects to the compositor at `127.0.0.1:<vnc_port>`.
   `[15900, 16799]`. Mirrors the agent web-UI port pattern from
   `docs/gotchas.md::Web UI ports collide on hash` — same FNV-1a
   constant, different range. The compositor's startup script writes
-  `/etc/hyperhive/gui.json = {"vnc_port":N,"auth":"none"}` so the
-  harness reads the port at runtime; no nix-side / harness-side hash
+  `/etc/hyperhive/gui.json = {"vnc_port":N,"auth":"none","wayland_display":"wayland-0"}`
+  so the harness reads the port at runtime; no nix-side / harness-side hash
   duplication.
+- **Fixed Wayland socket name (`--socket=wayland-0`)**: weston is
+  launched with `--socket=wayland-0` so the socket path is
+  deterministic. `harness-base.nix` exports `WAYLAND_DISPLAY=wayland-0`
+  and `XDG_RUNTIME_DIR=/run/user/0` as global system environment
+  variables (gated on `hyperhive.gui.enable`) so every systemd service
+  in the container inherits them. Without this, services starting
+  Wayland clients could not find the compositor — libwayland falls
+  back to a headless display or errors out, the app "works" on a
+  second invisible display, and the VNC session shows a blank weston
+  desktop (#540 double-screen).
 - **VNC bind address**: weston's VNC backend has no CLI
   bind-address flag (unlike the RDP backend's `--address`), so the
   listener binds `0.0.0.0`. The harness relay only connects via
diff --git a/nix/templates/weston-vnc.nix b/nix/templates/weston-vnc.nix
index 11f90308..1102ef1d 100644
--- a/nix/templates/weston-vnc.nix
+++ b/nix/templates/weston-vnc.nix
@@ -78,9 +78,11 @@
           done
           VNC_PORT=$((15900 + hash % 900))
 
-          # Marker file the harness reads at startup.
+          # Marker file the harness reads at startup. Also records the
+          # fixed Wayland socket name (`wayland-0`) so other tooling can
+          # read it without inspecting the env-var injection below.
           ${pkgs.coreutils}/bin/mkdir -p /etc/hyperhive
-          ${pkgs.coreutils}/bin/printf '{"vnc_port":%d,"auth":"none"}\n' \
+          ${pkgs.coreutils}/bin/printf '{"vnc_port":%d,"auth":"none","wayland_display":"wayland-0"}\n' \
             "$VNC_PORT" > /etc/hyperhive/gui.json || true
 
           # --disable-transport-layer-security: skips the VeNCrypt TLS
@@ -90,11 +92,21 @@
           WESTON_INI=$(${pkgs.coreutils}/bin/mktemp /tmp/weston-XXXXXX.ini)
           ${pkgs.coreutils}/bin/printf '[core]\nidle-time=0\n\n[vnc]\nauth-method=none\n' > "$WESTON_INI"
 
+          # --socket=wayland-0: pin the compositor's Wayland socket name
+          # to `wayland-0` (weston default is to pick any free name such
+          # as `wayland-1`). Pinning lets the WAYLAND_DISPLAY=wayland-0
+          # global env injection below (see environment.variables) take
+          # effect unconditionally — any Wayland client launched by any
+          # systemd service in this container automatically connects to
+          # this compositor instead of failing or starting a second
+          # isolated display. Closes #540 (double-screen: VNC shows blank
+          # weston desktop while services render on a different seat).
           exec ${pkgs.weston}/bin/weston \
             --config="$WESTON_INI" \
             --backend=vnc-backend.so \
             --renderer=pixman \
             --port="$VNC_PORT" \
+            --socket=wayland-0 \
             --disable-transport-layer-security
         '';
         Restart = "on-failure";
@@ -102,6 +114,19 @@
       };
     };
 
+    # Expose the compositor's socket to every process in the container
+    # so Wayland clients started by any systemd service (e.g. bitburner
+    # launched via a `systemd.services.*` declaration in agent.nix) can
+    # find the compositor without per-service wiring. Without these vars
+    # a service starting a Wayland client would either fail to connect
+    # (libwayland falls back to creating a headless display) or open a
+    # second compositor entirely — both cause the VNC session to show a
+    # blank desktop while apps appear to "work" elsewhere (#540).
+    environment.variables = {
+      WAYLAND_DISPLAY = "wayland-0";
+      XDG_RUNTIME_DIR = "/run/user/0";
+    };
+
     # weston on the agent's interactive PATH so claude can run Wayland
     # clients / weston-info against the compositor.
     environment.systemPackages = [ pkgs.weston ];