random_steps does not have a lonlat option and calculates wrong x2_ y2_

[RonEfrat]

I am not sure if it is really an issue or perhaps on purpose, or maybe I am missing something?
When using random_steps on a steps_xyt object that was made with lonlat = T, the function still calculates the random steps' x2_ and y2_ the same way, leading to points that are not the right distance from x1_, y1_.

EDIT: Another strange thing that happens when using lonlat = T is that the turning angles used in random_steps are really strange, even though I give it the sl_distr, ta_distr, rand_sl and rand_ta myslef