"egui_glow requires opengl 2.0+." when trying to run egui through vscode remote ssh

[LB767]

When trying to debug or run my egui project on a remote host through VS Code's remote SSH extension, the program panics with the above error. The remote host is a Windows 10 PC with an integrated Intel UHD 630 graphics chip. I checked the supported OpenGL version and it's all fine (>4.6) and running the program locally from vscode on the actual remote works just fine, so it seems to be some weird behavior between opengl and ssh.

To Reproduce Steps to reproduce the behavior:

1. Get the remote SSH extension on vscode
2. Connect to a remote host
3. Run the egui demo

Expected behavior It should not panic.

Desktop:

• OS: Windows 11 on dev machine and Windows 10 on remote host
• Version 22

[emilk]

try the wgpu backend, or raise this error in the glutin repository

[Anny0nn]

Got the same issue but only when using hwaccel off on glow, Wgpu seems to be working just fine with hwacell off

[Colonial-Dev]

For what it's worth, this also happens inside Windows VMs, even with wgpu. (You get an unwrap panic on Wgpu(NoSuitableAdapterFound).)

If you're trying to test a cross-compiled version of your app, download this pre-built Mesa distribution. Run the systemwidedeploy batch script and select the first option (1. Core desktop OpenGL drivers) and your app should work.

[mammothcoding]

Please advice. How can I unwrap this panic in simple example?

[dscso]

I have had the same issue running in my Windows 11 VM for testing, then I changed to wgpu backend which solved the issue.

Now after updating to versions 0.28 I get the same error:

Error: Wgpu(NoSuitableAdapterFound)

unfortunately I can not roll back to version 0.25 since then I get a compiler issue:

eframe = { version = "0.25", default-features = false, features = ["persistence", "wgpu", "accesskit", "default_fonts", "glow"]}

stdout :     Blocking waiting for file lock on package cache
    Updating crates.io index
       Fetch [=================>               ] 87 complete; 75 pending      
       Fetch [=================>               ] 139 complete; 23 pending     
error: failed to select a version for the requirement `wgpu = "^0.18.0"`
candidate versions found which didn't match: 22.1.0, 22.0.0, 0.20.1, ...
location searched: crates.io index
required by package `eframe v0.25.0`
    ... which satisfies dependency `eframe = "^0.25"` of package `client-gui v0.0.0`

Edit: The 0.25.0 problem is a skill issue. Locking dependencies does the trick.

[Alcinzal]

Please advice. How can I unwrap this panic in simple example?

The error occurs in these functions: eframe::run_simple_native() or eframe::run_native(). So you can add an .expect("msg") at the end of them.

[mammothcoding]

Please advice. How can I unwrap this panic in simple example?

The error occurs in these functions: eframe::run_simple_native() or eframe::run_native(). So you can add an .expect("msg") at the end of them.

Thanks but this still doesn't solve the problem with running the application under VM. By the way, on some types of VM with good emulation of video cards this problem doesn't exist (like Proxmox).