undo times vector change

[coveralls]

Pull Request Test Coverage Report for Build 884375509

• 3 of 5 (60.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 63.361%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/validatedODEs.jl	3	5	60.0%
<!--	Total:	3	5	60.0%	-->

Totals
Change from base Build 883408675:	0.0%
Covered Lines:	773
Relevant Lines:	1220

---

💛 - Coveralls

[lbenet]

I don't think this change is necessary, as I explained in this comment. The current (master) behavior is more consistent with respect to indexing, in such a way that the returned time at step n corresponds to the origin of the domain of the n-th entry of the solution, while before, it was the n-1 time entry.

[mforets]

but then, the length should be n+1? (or don't return the first reachable set at the time point t0). otherwise, we don't know by looking only at tTM when does the flowpipe end as i showed in this example

[lbenet]

The length is actually maxsteps+1 (see here). The extra 1 comes from the initial condition, which i kept because it's sometimes handy to use directly the TaylorModel1 of that initial condition, as we do it in the tests (see here).

See #117 to see how everything becomes more consistent.

[mforets]

i still think that it is unexpected to just see two zeros in the example above instead of [0, 0, 0.1] (because validated_integ returns a view of only the maxsteps terms of tTM, not the full maxsteps+1) but ok :thumbsup: