Relax the tolerance for radius calculation

[alarshi]

While writing a test for the WorldbuilderApp, I got an error at ln 288 in source/world.cc as
Inconsistent input. Please check whether the radius in the sperhical coordiantes is consistent with the radius of the planet as defined in the program that uses the Geodynamic World Builder. Depth = 0, radius = 6.371e+06, point = 5.77211e+06 -2.69158e+06 -166773, radius-point.norm() = -9.31323e-10
This residual radius norm is quite low and this error was circumvented using the option -ldcc by @MFraters . Maybe we can add this option as default.

[MFraters]

The tester now has -ldcc on by default, which should solve the problem I think. Let me know if you think it should be the default in all cases. My worry with that is that it becomes kind of useless, because you would need to turn this option, which no one really now about, manually on. We could improve the message to tell about a way to overwrite this assert.