R lm() weights argument being ignored when placed inside function

Question

I am trying to figure out why the following piece of code ignores the weights argument and produces simply an unweighted regression analysis. If I remove the function wrappi