Add "in" for IntervalBox

[dpsanders]

Fixes #240.

Defines ∈ for IntervalBox correctly.

cc @Kolaru

[Kolaru]

Looks globally good to me, I only have some minor improvement propositions:

• Defining in for Vector as well, so that [0, 0] in IntervalBox(-1..1, -1..1) does not error.
• Testing a case where v in X is false.

[dpsanders]

OK, will define for AbstractVector{T}.

I realised that the problem is that the type T inside the vector should not be allowed to be an interval though? Or maybe that's not important.

[dpsanders]

The reason for restricting to SVector{N} was that then you guarantee that the vector is of the correct length.

Also, is there any particular reason why you are using standard Julia vectors instead of SVectors?

[codecov-io]

Codecov Report

Merging #244 into master will increase coverage by 0.06%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #244      +/-   ##
==========================================
+ Coverage    82.7%   82.76%   +0.06%     
==========================================
  Files          24       24              
  Lines        1139     1143       +4     
==========================================
+ Hits          942      946       +4     
  Misses        197      197

Impacted Files	Coverage Δ
src/IntervalArithmetic.jl	50% <ø> (ø)	⬆️
src/multidim/intervalbox.jl	90.9% <100%> (+0.58%)	⬆️
src/intervals/complex.jl	72.54% <100%> (+0.54%)	⬆️
src/intervals/set_operations.jl	100% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4835e16...07490f6. Read the comment docs.

[coveralls]

Coverage increased (+0.06%) to 82.765% when pulling 07490f6 on fix_membership_intervalbox into 4835e16 on master.

[Kolaru]

I realised that the problem is that the type T inside the vector should not be allowed to be an interval though? Or maybe that's not important.

Technically you want T to be a type that can be a bound for an interval. Disallowing Interval is good but does not cover all cases. Doing so probably requires something like what is proposed in #185.

Also, is there any particular reason why you are using standard Julia vectors instead of SVectors?

Only when testing stuff in REPL as standard vectors are quicker to type.

[dpsanders]

Yes I agree that something more is required, but I think this is as far as this PR needs to go. Do you agree?

[Kolaru]

Yes, the "more" should be dealt with in a separate PR (I'll look in renaming/polishing my proposal soon).

The current PR looks good to me, by the way.

[lbenet]

Is this ready for reviewing/merging? I actually agree with @Kolaru's view that this is already ok.

[dpsanders]

Yes, in my opinion it's ready to merge.

[lbenet]

Thanks!