Readd cache

[ThanatosGit]

Ok, I lied in the other PR. I just looked into the code. We return a generator via yield in the function.
It's a pretty bad idea to use a cache on a generator because it will return the same generator multiple times but we iterate through it, so the generator is already exhausted after first loop.
Or short: This code is only executed once and not for all required scenarios.

    for asset in editor.get_asset_names_in_folder("actors/props/systemmechdna"):
        editor.ensure_present_in_scenario(scenario_name, asset)

I switched to return a list instead.
Should have seen this when I initially reviewed it. Stupid me, sorry. :(

But better to not include for the next release.

Fix #435

[Miepee]

Can you show the performance change between main / this branch?

[dyceron]

Hmm, my testing shows it to be about the same?

[ThanatosGit]

Ok

[dyceron]

I didn't mean for you to close this. I was hoping you would test this to confirm.