[Draft] ModifierTrait

[xeren]

Adds more control over the used lattice in InclusionBasedPointerAnalysis. More implementations can be added later.

• --program.analysis.alias=EXPERIMENTAL_FIELD_INSENSITIVE uses the singleton lattice $\lbrace\mathbb Z\rbrace$ implemented in ModifierTrait.VoidTrait. It lets the analysis be truely field-insensitive. It is not generally, but mainly, less precise than FIELD_INSENSITIVE implemented in AndersenAliasAnalysis.
• --program.analysis.alias=EXPERIMENTAL_FIELD_SENSITIVE uses the lattice $\mathbb F(\mathbb Z)+\lbrace\mathbb Z\rbrace$ of finite sets. It is implemented in ModifierTrait.Offsets and comparable with FIELD_INSENSITIVE and FIELD_SENSITIVE.
• --program.analysis.alias=EXPERIMENTAL_LINEAR_1D uses $\lbrace n+\mathbb Zs\mid n,s\in\mathbb Z\rbrace$ in ModifierTrait.SdLinear.
• --program.analysis.alias=FULL remains using $\lbrace n+\sum_{i=0}^k\mathbb Zs_k\mid k\in\mathbb N,n,s_1,\ldots,s_k\in\mathbb Z\rbrace$ which has moved to ModifierTrait.MdLinear.

[ThomasHaas]

Interesting changes, but why the changes to the IntervalAnalysis? It was a deliberate choice to make it Optional.
Apart from the fact that those changes shouldn't be part of this PR in the first place :)

[xeren]

You are right, this had nothing to do there.