Fix support for MOI.TimeLimitSec

[odow]

x-ref jump-dev/MathOptInterface.jl#2256

[codecov]

Codecov Report

Patch coverage: 100.00% and project coverage change: +1.90% 🎉

Comparison is base (3a5bab7) 80.10% compared to head (1fe0dc2) 82.01%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #146      +/-   ##
==========================================
+ Coverage   80.10%   82.01%   +1.90%     
==========================================
  Files           3        3              
  Lines         372      378       +6     
==========================================
+ Hits          298      310      +12     
+ Misses         74       68       -6     

Files Changed	Coverage Δ
src/MOI_wrapper/MOI_wrapper.jl	83.28% <100.00%> (+0.51%)	⬆️

... and 2 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[odow]

cc @matbesancon there are a bunch of solvers like this that had slightly varying implementations of TimeLimitSec. This PR is arguably a bug fix, but it is potentially breaking in that querying TimeLimitSec will now return nothing when it used to return -1.0. Do you think that's okay?

[matbesancon]

I would say that's fine. There are always problems with getting unset attribute values, because it often requires the wrapper to store what has been set or not, which is quite cumbersome.
In this case one could argue that code that breaks was relying on incorrect solver behavior / UB in some way