Merge remote-tracking branch 'origin/master' into floats-scientific

src/Analysis/Abstract/Caching.hs

@@ -80,7 +80,7 @@ instance ( Corecursive t
          , Semigroup (Cell (LocationFor v) v)
          )
          => MonadAnalysis t v (CachingAnalysis t v) where
-  evaluateTerm = foldSubterms $ \e -> do
+  analyzeTerm e = do
     c <- getConfiguration (embedSubterm e)
     -- Convergence here is predicated upon an Eq instance, not α-equivalence
     cache <- converge (\ prevCache -> do

src/Analysis/Abstract/Dead.hs

@@ -79,6 +79,6 @@ instance ( Corecursive t
          , Semigroup (Cell (LocationFor v) v)
          )
          => MonadAnalysis t v (DeadCodeAnalysis t v) where
-  evaluateTerm = foldSubterms (\ term -> do
+  analyzeTerm term = do
     revive (embedSubterm term)
-    eval term)
+    eval term

src/Analysis/Abstract/Evaluating.hs

@@ -76,4 +76,4 @@ instance ( Evaluatable (Base t)
          , Semigroup (Cell (LocationFor v) v)
          )
          => MonadAnalysis t v (Evaluation t v) where
-  evaluateTerm = foldSubterms eval
+  analyzeTerm = eval

src/Control/Abstract/Analysis.hs

@@ -1,9 +1,18 @@
-{-# LANGUAGE FunctionalDependencies #-}
+{-# LANGUAGE DefaultSignatures, FunctionalDependencies #-}
 module Control.Abstract.Analysis where
 
+import Prologue
+
 -- | A 'Monad' in which one can evaluate some specific term type to some specific value type.
 --
 --   This typeclass is left intentionally unconstrained to avoid circular dependencies between it and other typeclasses.
 class Monad m => MonadAnalysis term value m | m -> term, m -> value where
-  -- | Evaluate a term to a value using the semantics of the current analysis. This should always be used instead of explicitly folding 'eval' over subterms, except in 'MonadAnalysis' instances themselves.
-  evaluateTerm :: term -> m value
+  -- | Analyze a term using the semantics of the current analysis. This should generally only be called by definitions of 'evaluateTerm' and 'analyzeTerm' in this or other instances.
+  analyzeTerm :: SubtermAlgebra (Base term) term (m value)
+
+  -- | Evaluate a term to a value using the semantics of the current analysis.
+  --
+  --   This should always be called instead of explicitly folding either 'eval' or 'analyzeTerm' over subterms, except in 'MonadAnalysis' instances themselves.
+  evaluateTerm :: MonadAnalysis term value m => term -> m value
+  default evaluateTerm :: (MonadAnalysis term value m, Recursive term) => term -> m value
+  evaluateTerm = foldSubterms analyzeTerm