[QUESTION] Support skipping instantiation in libdoc or fallback to static keyword discovery when constructor fails

[violette]

Hi,

I'm working on a large project with custom Robot libraries that have complex Python dependencies. Some of these libraries run logic in their init methods, which can raise exceptions when certain environments or services aren’t available.

This prevents the language server from indexing or navigating their keywords... which is super annoying as the whole robot is not working even is the .py file (which contain the keywords) is just next to it -> as it follows the hierarchy and finish to fail...

Is there a way to skip class instantiation during keyword discovery? I looked through the documentation but didn’t find anything — maybe I missed it. If not, is this something you plan to support?

Something like:
"robotcode.robot.env": { "ROBOTCODE_ANALYZING_LEVEL": "1" }
would be super nice.

Thanks,

[d-biehl]

sorry, there is no way to skip class initialization or to do static analyzing of libraries, that's the way how libraries in robot framework works.

But did you read this?: https://robotframework.org/robotframework/latest/RobotFrameworkUserGuide.html#detecting-is-robot-framework-running

this works in most cases for normal robot libraries, if you have dynamic libraries that changes the keywords depending on arguments, variables and so on, you need maybe another approach.

What type of libraries do you have?