Merge branch 'normalize-unevaluated-module-tables' of https://github.com/github/semantic into normalize-unevaluated-module-tables

src/Language/Java/Assignment.hs

@@ -145,13 +145,12 @@ expressionChoices =
   , continue
   , constructorDeclaration
   , explicitConstructorInvocation
-  -- , constantDeclaration
+  -- , TODO: constantDeclaration
   , doWhile
   , fieldDeclaration
   , float
   , for
   , enum
-  -- , hexadecimal
   , if'
   , interface
   , identifier
@@ -239,16 +238,6 @@ scopedIdentifier = makeTerm <$> symbol ScopedIdentifier <*> children (Expression
 
 superInterfaces :: Assignment.Assignment [] Grammar [Term]
 superInterfaces = symbol SuperInterfaces *> children (symbol InterfaceTypeList *> children(manyTerm type'))
--- a *> b
--- both of these are impure
--- getLine *> getLine
--- in half apply, they're both monadic impure actions
--- :t (<$)
--- :t (*>)
-
--- what does it mean to say monadic action? more precise term: sequence-able
--- a sequence of applicative actions can be executed left to right
--- applicative computations can't do branch and control flow; applicative computations can only compute in a direct line, monadic can compute arbitrary branches
 
 -- Declarations
 class' :: Assignment
@@ -257,19 +246,11 @@ class' = makeTerm <$> symbol ClassDeclaration <*> children (makeClass <$> many m
     makeClass modifiers identifier typeParams superClass superInterfaces = Declaration.Class (modifiers ++ typeParams) identifier (maybeToList superClass ++ superInterfaces) -- not doing an assignment, just straight up function
     classBody = makeTerm <$> symbol ClassBody <*> children (manyTerm expression)
     superClass = symbol Superclass *> children type'
-    -- matching term expression won't work since there is no node for that; it's AnonExtends
-    -- superClass = makeTerm <$> symbol SuperClass <*> children (Java.Syntax.SuperClass <$> term expression <*> type')
-    -- We'd still like to match the SuperClass node, but we don't need to create a syntax to make a term
-    -- Do you lose info by omitting the superclass term? No...
-    -- Don't need to make a term since we're not using syntax
-    -- what's the difference between using tokens: AnonExtends GenericType?
-    -- optional: when something can or can't exist and you want to produce a Maybe
--- TODO: superclass -- need to match the superclass node when it exists (which will be a rule, similar to how the type params rule matches the typeparams node when it exists)
+-- TODO: superclass
+--       need to match the superclass node when it exists (which will be a rule, similar to how the type params rule matches the typeparams node when it exists)
 --       optional, when we have a single term
 --       superInterfaces is also optional but since it produces a list, lists already have an empty value so we don't need to wrap it up in a maybe to get an empty value
 
--- define this at the top level, we may change TS grammar so that if someone wants to write a Java snippet we could assign
--- it correctly; fieldDeclaration is standalone (compared to a type, which doesn't say anything by itself)
 fieldDeclaration :: Assignment
 fieldDeclaration = makeTerm <$> symbol FieldDeclaration <*> children ((,) <$> manyTerm modifier <*> type' <**> variableDeclaratorList)
 
@@ -281,14 +262,8 @@ method = makeTerm <$> symbol MethodDeclaration <*> children (makeMethod <$> many
     methodHeader = symbol MethodHeader *> children ((,,,,) <$> (typeParameters <|> pure []) <*> manyTerm annotation <*> type' <*> methodDeclarator <*> (throws <|> pure []))
     makeMethod modifiers receiver (typeParams, annotations, returnType, (name, params), throws) = Declaration.Method (returnType : modifiers ++ typeParams ++ annotations ++ throws) receiver name params
 
--- TODO: add genericType
--- Question: should this genericType be part of type or not? Its own type because it's different structurally
-
 generic :: Assignment
 generic = makeTerm <$> symbol Grammar.GenericType <*> children(Java.Syntax.GenericType <$> term type' <*> manyTerm type')
--- when do we make a term again? - if we want to wrap something in a syntax constructor, because each piece of syntax
--- will be populated by further terms inside it. in this case, we wrap two terms in a piece of syntax.
--- Q to help decide: do we lose anything by omitting the term?
 
 methodInvocation :: Assignment
 methodInvocation = makeTerm <$> symbol MethodInvocation <*> children (Expression.Call [] <$> (callFunction <$> expression <*> optional (token AnonDot *> expression)) <*> (argumentList <|> pure []) <*> emptyTerm)
@@ -318,14 +293,9 @@ interface = makeTerm <$> symbol InterfaceDeclaration <*> children (normal <|> an
     annotationType = symbol AnnotationTypeDeclaration *> children (Declaration.InterfaceDeclaration [] <$> identifier <*> annotationTypeBody)
     annotationTypeBody = makeTerm <$> symbol AnnotationTypeBody <*> children (many expression)
     interfaceMemberDeclaration = symbol InterfaceMemberDeclaration *> children (term expression)
-    -- we won't make a term because we have a choice of a bunch of things
 
 package :: Assignment
--- package = makeTerm <$> symbol PackageDeclaration <*> children (Java.Syntax.Package <$> someTerm expression)
-package = do
-  loc <- symbol PackageDeclaration -- location which is calling the symbol API
-  c <- children $ do Java.Syntax.Package <$> someTerm expression
-  pure (makeTerm loc c) -- pure is re-wrapping it back into the outer context, which in this case is Assignment (ie., the return type of the function)
+package = makeTerm <$> symbol PackageDeclaration <*> children (Java.Syntax.Package <$> someTerm expression)
 
 enum :: Assignment
 enum = makeTerm <$> symbol Grammar.EnumDeclaration <*> children (Java.Syntax.EnumDeclaration <$> term identifier <*> manyTerm enumConstant)
@@ -515,12 +485,10 @@ constructorDeclaration = makeTerm <$> symbol ConstructorDeclaration <*> children
       constructor modifiers (typeParameters, identifier, formalParameters) = Java.Syntax.Constructor modifiers typeParameters identifier formalParameters -- let partial application do its thing
 
 typeParameters :: Assignment.Assignment [] Grammar [Term]
-typeParameters = symbol TypeParameters *> children (manyTerm typeParam) -- this produces a list, which is what we need to return given by the type definition
+typeParameters = symbol TypeParameters *> children (manyTerm typeParam)
   where
-    typeParam = makeTerm <$> symbol Grammar.TypeParameter <*> children (Java.Syntax.TypeParameter <$> manyTerm annotation <*> term identifier <*> (typeBound <|> pure [])) -- wrapping up all three of those fields so we need to makeTerm (producing a term here)
+    typeParam = makeTerm <$> symbol Grammar.TypeParameter <*> children (Java.Syntax.TypeParameter <$> manyTerm annotation <*> term identifier <*> (typeBound <|> pure []))
     typeBound = symbol TypeBound *> children (manyTerm type')
-    -- manyTerm typeParam made sense because each type Parameter was wrapped up into a Grammar.TypeParameter node, dissimilar
-    -- to superInterfaces
 
 annotation :: Assignment
 annotation = makeTerm <$> symbol NormalAnnotation <*> children (Java.Syntax.Annotation <$> term expression <*> (elementValuePairList <|> pure []))
@@ -540,14 +508,6 @@ formalParameters = manyTerm parameter
     parameter = makeTerm <$> symbol FormalParameter <*> children (makeAnnotation <$> manyTerm modifier <*> type' <* symbol VariableDeclaratorId <*> children identifier)
     makeAnnotation [] type' variableName = Type.Annotation variableName type'
     makeAnnotation modifiers type' variableName = Type.Annotation variableName (makeTerm1 (Java.Syntax.TypeWithModifiers modifiers type'))
--- know when we are in a functor context and fmap is all gravy
--- we're just wrapping stuff up in data, we aren't building a pattern (assignment) so we aren't in an applicative context
--- when in an applicative context, we're also in a functor context (ie., defining how fmap will work over it)
--- sometimes it is nice to be able to say you're in an applicative context without refering to any particular applicative instance
-
--- constantDeclaration :: Assignment
--- constantDeclaration = makeTerm <$> symbol ConstantDeclaration <*>
 
 castExpression :: Assignment
 castExpression = makeTerm <$> symbol CastExpression <*> children (flip Type.Annotation <$> type' <*> term expression)
--- term expression, because we can deal with comments