Add a guard against edge case with zero span in linear guessing

[DaniGlez]

If the span of the vector (last(v) - first(v)) is 0, the current linear guessing relies on an isinf guard. This fails for one value, first(v), since the linear-guess quotient 0/0 produces a NaN instead of +Inf or -Inf. This PR aims to patch that edge case.

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

[ChrisRackauckas]

I don't want to know what real world case made you find this 😅

[DaniGlez]

It's through usage of DataInterpolations, I promise it's more sensible than one might think 😅