Round bounds for ZeroOne variables to avoid CPLEX warning

[odow]

Closes #311

[codecov]

Codecov Report

Merging #318 into master will increase coverage by 0.32%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #318      +/-   ##
==========================================
+ Coverage   54.08%   54.40%   +0.32%     
==========================================
  Files           7        7              
  Lines        2703     2722      +19     
==========================================
+ Hits         1462     1481      +19     
  Misses       1241     1241              

Impacted Files	Coverage Δ
src/MOI/MOI_wrapper.jl	89.35% <100.00%> (+0.15%)	⬆️

---

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 8bf4a84...86e67c2. Read the comment docs.

[odow]

Barring objections, I'll leave this open for another 48 hours, then merge.

[blegat]

Not sure this is relevant but it seems you are handling cases where both bounds and the ZeroOne constraint are added.
These cases should be an error, see https://github.com/jump-dev/MathOptInterface.jl/blob/0423a6c61c6f47e3baf81642dd60e6c7a3c9e405/src/Utilities/model.jl#L482-L496

[odow]

These cases should be an error

Most definitely not. Where did we discuss this??? People add binary variables with bounds all the time.

[blegat]

It seems it was discussed in
jump-dev/MathOptInterface.jl#570 (comment)
and then implemented in
jump-dev/MathOptInterface.jl#732
Looking at the LOWER_BOUND_MASK and UPPER_BOUND_MASK in more details, it seems that we consider that ZeroOne don't set bound (the bounds are 0 and 1, we know it from the existence of the ZeroOne flag in the mask so we don't need to store the bounds anywhere so it does not conflict with other set that need their bounds to be stored.

[odow]

Correct. ZeroOne does not set bounds.