Graceful handling of infinity

[dpsanders]

Currently the following hangs in an infinite calculation:

julia> using IntervalArithmetic, IntervalRootFinding

julia> roots(x -> x^2 - x, -Inf..Inf)

since it cannot prove that it can exclude the regions at infinity.

We should explicitly check for infinite intervals and stop bisecting at some point, say when the result is above 1e200.

[KronosTheLate]

I don't have anything to contribute with in terms of solutions, but as a user, being able to check from -Inf to Inf is what I actually want to do most of the time, and so this would be a great feature.