Not able to start using Avalonia in VSCode on Linux

[nsk7even]

Hey there, I am a professional C# developer, using Microsoft Visual Studio on Windows at business.
Privately I only use Linux. Today I started to try out Avalonia for VSCode to build a Linux UI application.

Sadly, my experience with Avalonia is quite bad until now:

1. Installed net8
2. Installed Avalonia for VSCode
3. dotnet new install Avalonia.Templates
4. Create Avalonia Project
5. Avalonia MVVM App
6. Open
7. Start Debugging
8. Select C# in popup
9. Select C# <projectname> in popup
10. Hello world example running
11. Terminal error CS0103 of Views/MainWindow.axaml.cs(9,9): the name "InitializeComponent" does not exists (compiles anyways)
12. Console is flooded with [..] Skipped loading symbols.[..] messages
13. Inserted "moduleLoad": false for logging in launch.json to stop this
14. Console shows ObjectDisposedExceptions at stop

So far so good (except the ugly terminal outputs). But it seems that this example is quite of precompiled/prepared, because as soon as I am making changes, the hell begins:

1. Text completion does not work (see request textDocument/completion failed AvaloniaVSCode#113 and VScode Avalonia XAML autocompletion doesn't work AvaloniaVSCode#110)
2. Downgraded Avalonia for VSCode to 0.29
3. Text completion works now
4. Made change to MainWindow.axaml: wrapped TextBlock from example into a Grid
5. Build failure: CSC : error AXN0002: Dots aren't allowed in type names Line 20, position 14. [.... /dev/test/test.csproj]

Well. Starting with troubleshooting after troubleshooting when just trying to set up the toolchain is not what my intent was.

Am I doing something completely wrong?
Is it too early for beginners to start using Avalonia UI on linux?

---

Sorry, just recognised that the build failure was due to a syntax error. This wasn't visualized in the UI...

[timunie]

Rider can maybe run on Linux. Dev experience is way better there imo when it comes to Avalonia

[nsk7even]

Yes, Rider provides a far better experience, I tried it out in parallel.
But it is much slower and costs money. On a faster machine and when using it regularly, it is the better option for sure.