| Safe Haskell | Safe-Inferred |
|---|
DDC.Core.Llvm.Metadata.Graph
- newtype UG a = UG (Dom a, Rel a)
- newtype DG a = DG (Dom a, Rel a)
- orientUG :: (Show a, Ord a) => UG a -> DG a
- partitionDG :: Eq a => DG a -> [Tree a]
- newtype Tree a = Tree (Dom a, Rel a)
- sources :: Eq a => a -> Tree a -> [a]
- anchor :: Eq a => a -> Tree a -> Tree a
- type Dom a = [a]
- type Rel a = a -> a -> Bool
- fromList :: Eq a => [(a, a)] -> Rel a
- toList :: Dom a -> Rel a -> [(a, a)]
- transClosure :: Eq a => Dom a -> Rel a -> Rel a
- transOrient :: (Show a, Ord a) => UG a -> DG a
- aliasMeasure :: Eq a => Rel a -> Partitioning a -> Int
- isTree :: Dom a -> Rel a -> Bool
Graphs and Trees for TBAA metadata
An undirected graph.
Instances
| Show a => Show (UG a) |
A directed graph.
partitionDG :: Eq a => DG a -> [Tree a]Source
Partition a DG into the minimum set of (directed) trees
An inverted tree (with edges going from child to parent)
Instances
| Show a => Show (Tree a) |
Quickcheck Testing ONLY
transClosure :: Eq a => Dom a -> Rel a -> Rel aSource
Find the transitive closure of a binary relation using Floyd-Warshall algorithm
transOrient :: (Show a, Ord a) => UG a -> DG aSource
Transitively orient an undireted graph
Using the algorithm from Lex-BFS and partition refinement, with applications to transitive orientation, interval graph recognition and consecutive ones testing, R. McConnell et al 2000
In the case where the transitive orientation does not exist, it simply gives some orientation
note: gave up on modular decomposition, this approach has very slightly worse time complexity but much simpler
aliasMeasure :: Eq a => Rel a -> Partitioning a -> IntSource
Calculate the aliasing induced by a set of trees this includes aliasing within each of the trees and aliasing among trees.