Oops, I did it again

Last year there was the simulation where the dark matter content of a galaxy was suddenly and brutally removed. Consequently, half of it exploded.

This one is similar but less intentional. I'm trying to get the gas to have self-gravity, so we don't have to simulate it with a fixed, analytic potential (which is unrealistic except in the case that the disc is stable). Got that working, but I had problems compiling the code with the analytic gravity as well, so I just commented it out. Et voila...

In last year's simulation we not only had dark matter and gas but also stars. In centre of the galaxy, the stellar mass dominates. Problem is in this simulation, there aren't any stars, and the mass of the gas is negligible - not nearly enough to hold it together. So this time it's not just the inner regions but the entire gas disc which flies apart. Since the boundary conditions are periodic (anything that hits the edge re-appears on the opposite side), what you get is an expanding gas disc that collides with itself. And behold, all hell breaks loose.