Dynamic time series regression with varying intervals/missing values in forecast

I am trying to run a dynamic regression y_t ~ Lag(y_t)+ x_t+e_t, but were my observations are separated by irregular spans of time (thus, the variance of the error is not constant from one observation to the other, but proportional to the length of the interval). I have seen in @robjhyndman and athanasopoulos book, that the function Arima of the forecast package 'handles missing data'. Does that mean that it will incorporate this type of heterocestadicity?

Thank you in advance!


Referred here by Forecasting: Principles and Practice, by Rob J Hyndman and George Athanasopoulos