Fixes potentially outdated value of getVal

[mmghannam]

Without this, the value could be outdated as it depends on calling getBestSol before.

[Joao-Dionisio]

Can you add a test that fails without this, please?

I'm a bit worried about scripts that call getVal very often, though. Sure, it's a tiny tiny additional effort, but people using things like getVarDict on big models might feel the difference. We would be creating a Solution object for every variable.

Can we somehow circumvent this? A terrible idea I just had was to attach an even handler to every model, that catches BESTSOLFOUND and the execmethod calls getBestSol.

[Joao-Dionisio]

This is cracking me up. If not raise error, then raise value error.
We were the ones who added this, Mo ahah

[mmghannam]

Can you add a test that fails without this, please?

I'm a bit worried about scripts that call getVal very often, though. Sure, it's a tiny tiny additional effort, but people using things like getVarDict on big models might feel the difference. We would be creating a Solution object for every variable.

Can we somehow circumvent this? A terrible idea I just had was to attach an even handler to every model, that catches BESTSOLFOUND and the execmethod calls getBestSol.

We could maybe prevent it in the solving stage, and raise and error guiding the user to the other methods.

[Joao-Dionisio]

what was it that we decided was the better way to deal with this @mmghannam ?

[marouane-f]

if not stage_check or self._bestSol.sol == NULL and SCIPgetStage(self._scip) != SCIP_STAGE_SOLVING:

The check allows to call getVal during solving, if a solution is available. Is this intended behavior?

[Joao-Dionisio]

@mmghannam I remember we had a discussion about this, with a better way to do this. Do you remember what it was?

I'm not too fond of preventing it in the solving stage, since the intended behavior seems clear, and crashing the user's code seems somewhat frustrating.

[mmghannam]

I unfortunately don't remember. I would also lean on the side of allowing it in the solving stage. Regarding the eventhandler idea, this would add an overhead to all models, which is not so nice. But either this solution or the eventhandler, any overhead is worth not returning an outdated value.