Comment 15 for bug 1167036

Mikael: I think that the periodicity isn't being (fully?) enforced when the "periodic" in the .geo file are removed, maybe because corresponding points are too far apart. The test I supplied then appears to work, but for the wrong reason. If I use the meshes for a dynamic simulation (shallow water equations), I get nonsense (blow-up from the edges) when I attempt to run a dynamic simulation with it. Indeed, I only get sensible results on my 16b mesh if the periodic lines are as they are written (e.g. changing -1 and -2 to 1 and 2 doesn't fix it!)

Garth: will try...