Fix linear_gradient float precision bug.

This was reported by:
https://github.com/ArthurSonzogni/FTXUI/issues/998

Indeed, the `t` interpolation factor, which is itself interpolated might become
slightly larger than 1.0. This is due to the float precision.
This was supposedly handled, but there was an off-by-one error in the check.

Along the way, fix a bug found by a fuzzer.

Bug: https://github.com/ArthurSonzogni/FTXUI/issues/998
Fixed: https://github.com/ArthurSonzogni/FTXUI/issues/998