Updat.she docs, use entrypoint.sh

.github/workflows/docker-test.yml

@@ -21,14 +21,13 @@ jobs:
 
     - name: Run commands in Docker container
       run: |
-        docker run --rm -v ${{ github.workspace }}:/workspace --user root \
-         -e HOST_UID=$(id -u) -e HOST_GID=$(id -g) test-image /bin/bash -c "
-          usermod -u \$HOST_UID builder && groupmod -g \$HOST_GID builder &&
-          chown -R \$HOST_UID:\$HOST_GID /home/builder &&
-          su builder -l -c 'cd /workspace &&
+        docker run --rm -v ${{ github.workspace }}:/workspace \
+          -e HOSTUID=$(id -u) -e HOSTGID=$(id -g) test-image \
+          '/bin/bash -l -c "
+          cd /workspace &&
             bun install &&
             bun run codegen &&
             bun run meson-setup.clang-release &&
-            bun run vs.example
-          '"
+            meson compile -C build/ vs:executable
+          "'
   

docker/Dockerfile

@@ -37,4 +37,5 @@ RUN \
     echo "export PATH=\"\$BUN_INSTALL/bin:\$PATH\"" >> $HOME/.profile
 
 USER root
-CMD ["bash", "-l"]
+ADD entrypoint.sh /entrypoint.sh
+ENTRYPOINT ["/entrypoint.sh"]

docker/README.md

@@ -2,7 +2,7 @@
 
 ## Using docker-compose
 
-You may build and test your changes by using the compose file:
+You may build your changes by using the compose file:
 
     docker-compose -f ./docker/docker-compose.yml up
 
@@ -12,6 +12,7 @@ While in the repository root, to enter the environment and be presented with a
 shell:
 
     docker run -it --rm -u builder \
+      -e HOSTUID=$(id -u) \
       -v $PWD:/workspace ghcr.io/KaruroChori/vs-fltk:build-env
 
 This will mount your current directory as */workspace* inside the container.
@@ -28,4 +29,81 @@ You can pull the image manually:
 If you use `docker-compose` or `docker run ...` the image will be pulled
 automatically the first time.
 
+## Running the app from the container
 
+Note the following instructions were adapted from instructions generated by
+ChatGPT. The X11 instructions were tested on Manjaro Linux. Use at your own
+risk. Please update these docs if you have any information about accuracy or
+security-related considerations.
+
+To run an application from a Docker container that requires access to the host
+machine's graphics (such as a GUI application), you'll need to share certain
+resources from the host with the container. This involves passing through the
+host's display (X11 or Wayland, depending on your system) and sharing the
+necessary devices. Here's a general guide:
+
+### For X11-based systems (e.g., most Linux desktop environments)
+
+1. **Install dependencies on the host:**
+   Make sure you have `xhost` installed, which allows you to control access to
+   the X server.
+
+    sudo apt install x11-xserver-utils  # For Ubuntu/Debian
+    sudo pacman -S xorg-xhost  # For Arch-based
+
+
+2. **Allow access to X server:**
+   Before running the container, allow Docker to access your X server:
+
+    xhost +local:docker
+
+
+3. **Run the container with access to the display:**
+   When starting the container, pass the display environment variable
+   (`$DISPLAY`) and mount the X11 socket:
+
+```bash
+docker run -it --rm -u builder \
+  -v $PWD:/workspace \
+  --env="DISPLAY" \
+  --env="QT_X11_NO_MITSHM=1" \
+  --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" \
+  ghcr.io/KaruroChori/vs-fltk:build-env
+```
+
+   - `--env="DISPLAY"`: Passes the display from the host to the container.
+   - `--env="QT_X11_NO_MITSHM=1"`: Some applications may need this to avoid shared memory issues.
+   - `--volume="/tmp/.X11-unix:/tmp/.X11-unix:rw"`: Shares the X11 socket.
+
+4. **Run the graphical application inside the container:**
+   Once inside the container, running the app should open its GUI on your host's screen.
+
+### For Wayland-based systems (e.g., some modern Linux environments):
+
+1. **Share the Wayland display:**
+   If your system uses Wayland, you need to pass the Wayland display and devices:
+
+```bash
+docker run -it --rm -u builder \
+  -v $PWD:/workspace \
+  --env="WAYLAND_DISPLAY=$WAYLAND_DISPLAY" \
+  --volume="/run/user/$(id -u)/wayland:/run/user/$(id -u)/wayland" \
+  --device=/dev/dri \
+  ghcr.io/KaruroChori/vs-fltk:build-env
+```
+
+   - `--env="WAYLAND_DISPLAY=$WAYLAND_DISPLAY"`: Passes the Wayland display variable.
+   - `--volume="/run/user/$(id -u)/wayland:/run/user/$(id -u)/wayland"`: Shares the Wayland display socket.
+   - `--device=/dev/dri`: Provides access to GPU devices (for rendering).
+
+### Additional considerations:
+
+- **GPU access**: If your application requires GPU acceleration (e.g.,
+OpenGL), you may also need to pass through GPU devices to the container. For
+example, with NVIDIA, you can use the NVIDIA Docker runtime:
+
+   docker run --gpus all -it ghcr.io/KaruroChori/vs-fltk:build-env
+
+- **X11 access security**: Allowing Docker to connect to your X server with `xhost +local:docker` is not secure. After you're done, revoke access:
+
+    xhost -local:docker

docker/entrypoint.sh

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+OLDPWD=$PWD
+
+if [ -z "HOSTUID" ]; then
+  echo "HOSTUID is not set."
+  exit 1
+fi
+
+if [ -z "HOSTGID" ]; then
+  echo "HOSTGID is not set."
+  exit 1
+fi
+
+usermod -u $HOSTUID builder
+groupmod -g $HOSTGID builder
+chown -R $HOSTUID:$HOSTGID /home/builder
+su builder -c "PATH=/home/builder/.local/bin:$PATH && cd $OLDPWD && $1"