Chat works but code gen doesn't

[andretost]

Description:
I installed the extension and enabled it in a new profile in VS Code. I have connected it to my local ollama instance, where I have the granite-code:8b model installed. I defined to use the granite-code:8b model for both chat and code gen. The chat function actually works, I can ask questions and get good answers. However, I am expecting to see code completion examples when typing in code in a new, say, .java file. But nothing happens. So it seems that the code gen is just not doing anything. I do not see any error messages anywhere.
Do I have to somehow enable code gen separately? Can I trigger it with some key combination? I don't see one configured in the keyboard settings for the extension.

How can I troubleshoot this? Some log that I could look at?

To Reproduce:
Steps to reproduce the behavior:

1. Install extension in new profile in VS Code.
2. Configure for ollama, model granite-code:8b, for both chat and code gen
3. Create new .java file
4. Start typing code, and wait to see code suggestions pop up.
5. Nothing happens.

Expected behavior:
I expect to see code suggestions that I can then accept using the tab key.

Environment:

• macOS Sequoia

[sreejeshtribm]

HI @andretost , We will check this and update you.

[daschnerm]

@andretost Can you please provide the logs of the output channel?

1. Set log level in settings to TRACE (wca.logLevel)
2. Click on View -> Output -> Select Granite.Code and Granite.Code Language Server on the right and copy the output of both logs.