fix: destroy_files function memory leak

[johanneshorner]

This pr should make the destroy_files function free the memory correctly, that was earlier allocated in the workspace_files function.

Automatic tests with memory analysis could be very useful. Haven't actually set up something like that on my own, but i can look into it.

@vhyrro Maybe you can give your opinion on the todos i added. If I were to use this library in my application, I think I would want the library to crash if I passed in a value that didn't really make sense.

[vhyrro]

First of all, thanks!

Automatic tests with memory analysis could be very useful. Haven't actually set up something like that on my own, but i can look into it.

Yeah, this would be super cool. I can try setting something up like this in a bit.

Maybe you can give your opinion on the todos i added. If I were to use this library in my application, I think I would want the library to crash if I passed in a value that didn't really make sense.

I agree that it would be nice to crash with a message in the create_* functions and all other functions that allocate or create new objects, but I don't believe we should crash in the destroy functions. To quote one of the points in the zig zen: "Resource allocation may fail; resource deallocation must succeed".

[johanneshorner]

That zig zen quote does make sense. I added asserts in the create_* functions.

[vhyrro]

Thank you yet again :D