add ValueLattice abstraction and refactor TNum for known bits

src/main/scala/analysis/EdgeFunctionLattice.scala

@@ -24,8 +24,10 @@ trait EdgeFunction[T] extends (T => T) {
 class EdgeFunctionLattice[T, L <: Lattice[T]](val valuelattice: L) extends Lattice[EdgeFunction[T]] {
 
   val bottom: ConstEdge = ConstEdge(valuelattice.bottom)
+  def top: Nothing = ???
 
   def lub(x: EdgeFunction[T], y: EdgeFunction[T]): EdgeFunction[T] = x.joinWith(y)
+  def glb(x: EdgeFunction[T], y: EdgeFunction[T]): Nothing = ???
 
   /** Edge labeled with identity function.
     */

src/main/scala/analysis/GammaDomains.scala

@@ -4,8 +4,6 @@ import ir.*
 
 type VarGammaMap = LatticeMap[Variable, LatticeSet[Variable]]
 
-implicit val variableLatticeSetTerm: LatticeSet[Variable] = LatticeSet.Bottom()
-
 /**
  * An abstract domain that determines for each variable, a set of variables whose gammas (at
  * the start of a procedure) are "affected" this variable's gamma. This is paramaterised by