Handle edge case values for abstol and span with ITP

[dcourteville]

Checklist

• Appropriate tests were added
• Any code changes were done in a way that does not break public API
• All documentation related to code changes were updated
• The new code follows the
  contributor guidelines, in particular the SciML Style Guide and
  COLPRAC.
• Any new documentation only uses public API

Additional context

Allows ITP to handle very small abstol (and in particular abstol=0.0) and very big intervals. This is required to use it in DiffEqBase.

[oscardssmith]

why isn't this still exp2(n_h+n0)? I would expect that to be faster.

[dcourteville]

The if/else ensure that exp2(n_h) < Inf but not that exp2(n_h + n0) < Inf. So ϵ * exp2(n_h) * exp2(n0) is safer since ϵ * exp2(n_h) ≈ span / 2.

[oscardssmith]

yeah this makes sense. It is somewhat annoying that this is parameterized a little bit inefficiently, but realistically this won't cost any real performance.