A hyprland script for a laptop-external-monitor setup, toggling between which is in use

hyprland-monitor-toggle.sh

#! /usr/bin/env sh

# A hyprland script for a laptop-external-monitor setup, toggling between which is in use

# Launch at startup to make hyprland disable the internal monitor if an external monitor is detected and enabled
# Additionally it's called with a keybind to switch between a laptop monitor and an external display
# Ideally the conditional monitor behaviour was instead done directly in hyprland.conf, but I'm not sure whether that's possible
#
# Relevant info:
# - hyprctl monitors: identifies currently enabled monitors
# - hyprctl monitors all: identifies ALL connected monitors - including those not in use
#
# Suggested use:
# Add this line somewhere after the regular monitor configuration in hyprland.conf:
# exec = /path/to/hyprland-monitors-toggle.sh
# Add a keybind to run this script on demand:
# bind =,SomeKeyHere, exec, /path/to/hyprland-monitors-toggle.sh

move_all_workspaces_to_monitor() {
  TARGET_MONITOR="$1"

  hyprctl workspaces | grep ^workspace | cut --delimiter ' ' --fields 3 | xargs -I '{}' hyprctl dispatch moveworkspacetomonitor '{}' "$TARGET_MONITOR"

  # Previous approach
  #hyprctl swapactiveworkspaces $EXTERNAL_MONITOR $INTERNAL_MONITOR
}

# TODO: Detect these instead of hardcoding them
INTERNAL_MONITOR="ADD YOUR INTERNAL MONITOR NAME HERE"
EXTERNAL_MONITOR="ADD YOUR EXTERNAL MONITOR NAME HERE"

NUM_MONITORS=$(hyprctl monitors all | grep --count Monitor)
NUM_MONITORS_ACTIVE=$(hyprctl monitors | grep --count Monitor)

# For initial startup if you use hyprland's default monitor settings:
# Turn off the laptop monitor if it + another monitor is active
if [ "$NUM_MONITORS_ACTIVE" -ge 2 ] && hyprctl monitors | cut --delimiter ' ' --fields 2 | grep --quiet ^$INTERNAL_MONITOR; then
    # Doing this I hopefully end up on workspace 1 on the external monitor rather than 2 at startup
    move_all_workspaces_to_monitor $EXTERNAL_MONITOR
    hyprctl keyword monitor "$INTERNAL_MONITOR, disable"
    # Alternate fix to ensure I start on workspace 1
    #hyprctl dispatch workspace 1
    exit
fi

# For dynamically toggling which monitor is active later via a keybind
if [ "$NUM_MONITORS" -gt 1 ]; then # Handling multiple monitors
  if hyprctl monitors | cut --delimiter ' ' --fields 2 | grep --quiet ^$EXTERNAL_MONITOR; then
    hyprctl keyword monitor $INTERNAL_MONITOR,preferred,0x0,1
    move_all_workspaces_to_monitor $INTERNAL_MONITOR
    hyprctl keyword monitor "$EXTERNAL_MONITOR, disable"
  else
    hyprctl keyword monitor $EXTERNAL_MONITOR,preferred,0x0,1
    move_all_workspaces_to_monitor $EXTERNAL_MONITOR
    hyprctl keyword monitor "$INTERNAL_MONITOR, disable"
  fi
else  # If the external monitor is disconnected without running this script first, it might become the case that no monitor is on - therefore turn on the laptop monitor!
    hyprctl keyword monitor $INTERNAL_MONITOR,preferred,0x0,1
    move_all_workspaces_to_monitor $INTERNAL_MONITOR
fi

Thanks1

I just did a rewrite that will hopefully solve that issue, but strangely during testing I also noticed that the monitor switch - or at least the version above - occasionally crashes either Hyprland itself or XWayland. Very strange.
XWayland crashes should no longer bring Hyprland down with it in the upcoming version, though, but still.

  # Jun 08 12:37:38 cpu kernel: traps: Hyprland[54550] trap invalid opcode ip:64776e17ed94 sp:7fff2fcec150 error:0 in Hyprland[64776e17c000+338000]
  # Jun 08 12:37:38 cpu systemd[1]: Started Process Core Dump (PID 55366/UID 0).
  # Jun 08 12:37:39 cpu systemd-coredump[55367]: [🡕] Process 54550 (Hyprland) of user 1000 dumped core.

It's something in these three lines (beside the comment) that causes it:

    hyprctl keyword monitor "$INTERNAL_MONITOR, preferred, 0, auto"
    # Move workspaces from the external monitor to the internal
    hyprctl swapactiveworkspaces $EXTERNAL_MONITOR $INTERNAL_MONITOR
    hyprctl keyword monitor "$EXTERNAL_MONITOR, disable"

swapactiveworkspaces is there due to sometimes seeing a crash when there were windows on multiple workspaces.

swapactiveworkspaces 1 0 (or 0 1, dont know which direction hyperland uses) instead of their names maybe ?

keep in mind it completely swaps workspaces.. see:

i think you're better off moving all (hidden) windows to a different workspace?
maybe with this: https://hyprland-community.github.io/pyprland/lost_windows.html

Good point, while it probably would work in my case since I only have one monitor open at a time it doesn't seem ideal.
I've put up a new version which moves them via the dispatch moveworkspacetomonitor instead.

Also changed how the new monitor is enabled, which seems to've fixed the issue with the occasional crashes.

I had some issues where the script stopped randomly working and enabled my internal monitor again. Not sure why exactly but i have been working on a different implementation using Hyprlands IPC socket (hyprctl-like requests) and socket2 (live events).

I have not managed to make the hyprctl-like requests work for socket. For example monitors works exactly like it does with hyprctl but workspaces returns nothing. What i did make work however was listing to socket2 events, so this would eliminate the need for manual reloading of the config for you script. Here is a example:

# Handle socket events from hyprland socket2
handleEvents() {
  case $event in
  monitoradded* | monitoraddedv2*)
    moveAllWorkspacesToMonitor "$EXTERNAL_MONITOR"
    hyprctl keyword monitor "$INTERNAL_MONITOR disable"
    ;;
  monitorremoved*)
    hyprctl keyword monitor "$INTERNAL_MONITOR,preferred,0x0,2"
    moveAllWorkspacesToMonitor "$INTERNAL_MONITOR"
    ;;
  *)
    // ignore all other events or use them somehow else
    ;;
  esac
}

# Listen to live events from hyprland socket2 continuously
socat -U - UNIX-CONNECT:$XDG_RUNTIME_DIR/hypr/$HYPRLAND_INSTANCE_SIGNATURE/.socket2.sock | while read -r line; do handleEvents "$line"; done