Consider a hypothesis test of difference of means for two independent populations x1 and x2. Suppose that both sample sizes are greater than 30 and that you know standard deviation 1 but not standard deviation 2. Is it standard practice to use the normal distribution or a Student's t distribution?