Window uses wrong offset in dual monitor setup

[MykBamberg]

The changes made in 8a192ad cause the window to have an incorrect offset on some monitors depending on their resolution.

Tested on labwc 0.9.1

[Kirottu]

The window position works fine for me on niri, with multiple different resolutions, scales and monitors. Can you share the specific monitor configurations (resolution, orientation, scale, etc.) where this happens and the output of Anyrun when you run it with this patch:

diff --git a/anyrun/src/main.rs b/anyrun/src/main.rs
index bd49278..2fb5743 100644
--- a/anyrun/src/main.rs
+++ b/anyrun/src/main.rs
@@ -636,6 +636,12 @@ fn activate(app: &gtk::Application, runtime_data: Rc<RefCell<RuntimeData>>) {
         .monitor_at_window(window.window().as_ref().unwrap())
         .unwrap();
 
+    println!(
+        "{}, {}",
+        monitor.geometry().width(),
+        monitor.geometry().width()
+    );
+
     let width = runtime_data
         .borrow()
         .config

[MykBamberg]

Monitors (kanshi config)

profile {
    output "DP-1" mode 2560x1440@60Hz
    output "DP-2" mode 1920x1080@60Hz
}

The window is only offset on DP-2.

The output of the patched executable on either monitor is

2560, 1440

I changed the second .width() to .height()

[MykBamberg]

The same behavior is also present on Sway and Hyprland.

[Kirottu]

I have also replicated this on Sway, for now best option is to revert back to the version before this change while I figure out why this happens.

[Kirottu]

Can you test out the branch relm4? I found a way that seemed to fix the monitor selection issue on GTK4 so it is more or less a full port of Anyrun to relm4. Some features are missing, and plugins need to be recompiled with it, but monitor selection should work.

[MykBamberg]

The branch does fix the issue.

[Kirottu]

The relm4 rewrite is now merged, so this is resolved.