a question on 'parscale'

[asas asas]

Dear lavaan users and experts,

I am reading the lav_model_estimate.R file in the package source file directory.

Can anyone explain the 'parscale' variable? In the 'objective_function' and 'gradient_function', it seems the 'x' argument is not the 'x' input to the optimization method.

Is this correct?

I could not understand the scaling before feeding 'x' to the optimizer.

Best,

Yilun