now we're fucking GETTING SOMEWHERE

layouts

oh my layouts

.ghci

@@ -0,0 +1,18 @@
+:set -XOverloadedStrings
+
+:set -package process
+
+:{
+import System.Exit qualified
+import System.Process qualified
+
+_reload_and_make _ = do
+    p <- System.Process.spawnCommand "make -f Makefile_happysrcs"
+    r <- System.Process.waitForProcess p
+    case r of
+        System.Exit.ExitSuccess -> pure ":reload"
+        _                       -> pure ""
+:}
+
+:def! r _reload_and_make
+

Makefile_happysrcs

@@ -0,0 +1,19 @@
+HAPPY = happy
+HAPPY_OPTS = -a -g -c
+ALEX = alex
+ALEX_OPTS = -g
+
+SRC = src
+CABAL_BUILD = dist-newstyle/build/x86_64-osx/ghc-9.6.2/rlp-0.1.0.0/build
+
+all: parsers lexers
+
+parsers: $(CABAL_BUILD)/Rlp/Parse.hs
+lexers: $(CABAL_BUILD)/Rlp/Lex.hs
+
+$(CABAL_BUILD)/Rlp/Parse.hs: $(SRC)/Rlp/Parse.y
+	$(HAPPY) $(HAPPY_OPTS) $< -o $@
+
+$(CABAL_BUILD)/Rlp/Lex.hs: $(SRC)/Rlp/Lex.x
+	$(ALEX) $(ALEX_OPTS) $< -o $@
+

README.md

@@ -12,14 +12,15 @@ $ cabal build       # Build the rlpc compiler
 $ cabal install     # Install rlpc to $PATH
 $ cabal haddock     # Build the API docs w/ Haddock
 $ make -C doc html  # Build the primary docs w/ Sphinx
+
+# run the test suite
+$ cabal test --test-show-details=direct
 ```
 
 ### Use
 ```sh
-# Compile and evaluate t.hs
-$ rlpc t.hs
-# Compile and evaluate t.hs, with evaluation info dumped to stderr
-$ rlpc -ddump-eval t.hs
+# Compile and evaluate examples/factorial.hs, with evaluation info dumped to stderr
+$ rlpc -ddump-eval examples/factorial.hs
 # Compile and evaluate t.hs, with evaluation info dumped to t.log
 $ rlpc -ddump-eval -l t.log t.hs
 # Print the raw structure describing the compiler options and die
@@ -80,7 +81,7 @@ Listed in order of importance.
     - [ ] Tail call optimisation
     - [x] Parsing rlp
 - [ ] Tests
-    - [ ] Generic example programs
+    - [x] Generic example programs
     - [ ] Parser
 
 ### December Release Plan

app/Main.hs

@@ -7,6 +7,9 @@ import Options.Applicative      hiding (ParseError)
 import Control.Monad
 import Control.Monad.Reader
 import Data.HashSet             qualified as S
+import Data.Text                (Text)
+import Data.Text                qualified as T
+import Data.Text.IO             qualified as TIO
 import System.IO
 import System.Exit              (exitSuccess)
 import Core
@@ -102,7 +105,7 @@ dshowFlags = whenFlag flagDDumpOpts do
 ddumpAST :: RLPCIO CompilerError ()
 ddumpAST = whenFlag flagDDumpAST $ forFiles_ \o f -> do
     liftIO $ withFile f ReadMode $ \h -> do
-        s <- hGetContents h
+        s <- TIO.hGetContents h
         case parseProg o s of
             Right (a,_) -> hPutStrLn stderr $ show a
             Left e -> error "todo errors lol"
@@ -110,10 +113,10 @@ ddumpAST = whenFlag flagDDumpAST $ forFiles_ \o f -> do
 ddumpEval :: RLPCIO CompilerError ()
 ddumpEval = whenFlag flagDDumpEval do
     fs <- view rlpcInputFiles
-    forM_ fs $ \f -> liftIO (readFile f) >>= doProg
+    forM_ fs $ \f -> liftIO (TIO.readFile f) >>= doProg
 
     where
-        doProg :: String -> RLPCIO CompilerError ()
+        doProg :: Text -> RLPCIO CompilerError ()
         doProg s = ask >>= \o -> case parseProg o s of
             -- TODO: error handling
             Left e -> addFatal . CompilerError $ show e
@@ -133,7 +136,7 @@ ddumpEval = whenFlag flagDDumpEval do
             where v f p h = f p h *> pure ()
 
 parseProg :: RLPCOptions
-          -> String
+          -> Text
           -> Either SrcError (Program', [SrcError])
 parseProg o = evalRLPC o . (lexCore >=> parseCoreProg)
 

doc/src/commentary/gm.rst

@@ -1,16 +1,24 @@
 The *G-Machine*
 ===============
 
+The G-Machine (graph machine) is the current heart of rlpc, until we potentially
+move onto a STG (spineless tagless graph machine) or a TIM (three-instruction
+machine). rl' source code is desugared into Core; a dumbed-down subset of rl',
+and then compiled to G-Machine code, which is then finally translated to the
+desired target.
+
 **********
 Motivation
 **********
 
-Our initial model, the *Template Instantiator* (TI) was a very
-straightforward solution to compilation, but its core design has a major
-Achilles' heel, being that Compilation is interleaved with evaluation -- The
-heap nodes for supercombinators hold uninstantiated expressions, i.e. raw ASTs
-straight from the parser. When a supercombinator is found on the stack during
-evaluation, the template expression is instantiated (compiled) on the spot.
+Our initial model, the *Template Instantiator* (TI) was a very straightforward
+solution to compilation, but its core design has a major Achilles' heel, being
+that compilation is interleaved with evaluation -- The heap nodes for
+supercombinators hold uninstantiated expressions, i.e. raw ASTs straight from
+the parser. When a supercombinator is found on the stack during evaluation, the
+template expression is instantiated (compiled) on the spot. This makes
+translation to an assembly difficult, undermining the point of an intermediate
+language.
 
 .. math::
    \transrule
@@ -31,7 +39,7 @@ evaluation, the template expression is instantiated (compiled) on the spot.
    \text{where } h' = \mathtt{instantiateU} \; e \; a_n \; h \; g
    }
 
-The process of instantiating a supercombinator goes something like this
+The process of instantiating a supercombinator goes something like this:
 
 1. Augment the environment with bindings to the arguments.
 
@@ -52,13 +60,17 @@ The process of instantiating a supercombinator goes something like this
 Instantiating the supercombinator's body in this way is the root of our
 Achilles' heel. Traversing a tree structure is a very non-linear task unfit for
 an assembly target. The goal of our new G-Machine is to compile a *linear
-sequence of instructions* which instantiate the expression at execution.
+sequence of instructions* which, **when executed**, build up a graph
+representing the code.
 
 **************************
 Trees and Vines, in Theory
 **************************
 
-WIP.
+Rather than instantiating an expression at runtime -- traversing the AST and
+building a graph -- we want to compile all expressions at compile-time,
+generating a linear sequence of instructions which may be executed to build the
+graph.
 
 **************************
 Evaluation: Slurping Vines

doc/src/commentary/ti.rst

@@ -1,6 +0,0 @@
-The *Template Instantiator*
-====================================
-
-WIP. This will hopefully be expanded into a thorough walkthrough of the state
-machine.
-

doc/src/conf.py

@@ -13,7 +13,7 @@ author = 'madeleine sydney slaga'
 # -- General configuration ---------------------------------------------------
 # https://www.sphinx-doc.org/en/master/usage/configuration.html#general-configuration
 
-extensions = ['sphinx.ext.imgmath']
+extensions = ['sphinx.ext.imgmath', 'sphinx.ext.graphviz']
 
 # templates_path = ['_templates']
 exclude_patterns = []
@@ -32,6 +32,7 @@ html_theme = 'alabaster'
 imgmath_latex_preamble = r'''
 \usepackage{amsmath}
 \usepackage{tabularray}
+\usepackage{syntax}
 
 \newcommand{\transrule}[2]
     {\begin{tblr}{|rrrlc|}

doc/src/references/rlp-grammar.rst

@@ -0,0 +1,67 @@
+The Complete Syntax of rl'
+==========================
+
+WIP.
+
+Provided is the complete syntax of rl' in (pseudo) EBNF. {A} represents zero or
+more A's, [A] means optional A, and terminals are wrapped in 'single-quotes'.
+
+.. math
+   :nowrap:
+
+   \setlength{\grammarparsep}{20pt plus 1pt minus 1pt}
+   \setlength{\grammarindent}{12em}
+   \begin{grammar}
+       <Decl> ::= <InfixDecl>
+       \alt <DataDecl>
+       \alt <TypeSig>
+       \alt <FunDef>
+
+       <InfixDecl> ::= <InfixWord> `litint' <Name>
+       <InfixWord> ::= `infix'
+       \alt `infixl'
+       \alt `infixr'
+
+       <DataDecl> ::= `data' `conname' {}
+
+   \end{grammar}
+
+.. code-block:: bnf
+
+   Decl      ::= InfixDecl
+               | DataDecl
+               | TypeSig
+               | FunDef
+
+   InfixDecl ::= InfixWord 'litint' Operator
+   InfixWord ::= 'infix'
+               | 'infixl'
+               | 'infixr'
+
+   DataDecl  ::= 'data' 'conname' {'name'} '=' Data
+   DataCons  ::= 'conname' {Type1} ['|' DataCons]
+
+   TypeSig   ::= Var '::' Type
+   FunDef    ::= Var {Pat1} '=' Expr
+
+   Type      ::= Type1 {Type1}
+               -- note that (->) is right-associative,
+               -- and extends as far as possible
+               | Type '->' Type
+   Type1     ::= '(' Type ')'
+               | 'conname'
+
+   Pat       ::= 'conname' Pat1 {Pat1}
+               | Pat 'consym' Pat
+
+   Pat1      ::= Literal
+               | 'conname'
+               | '(' Pat ')'
+
+   Literal   ::= 'litint'
+
+   Var ::= 'varname'
+         | '(' 'varsym' ')'
+   Con ::= 'conname'
+         | '(' 'consym' ')'
+

examples/constDivZero.hs

@@ -0,0 +1,3 @@
+k x y = x;
+main = k 3 ((/#) 1 0);
+

examples/factorial.hs

@@ -0,0 +1,7 @@
+fac n = case (==#) n 0 of
+    { 1 -> 1
+    ; 0 -> (*#) n (fac ((-#) n 1))
+    };
+
+main = fac 3;
+

examples/sumList.hs

@@ -0,0 +1,9 @@
+nil = Pack{0 0};
+cons x y = Pack{1 2} x y;
+list = cons 1 (cons 2 (cons 3 nil));
+sum l = case l of
+    { 0      -> 0
+    ; 1 x xs -> (+#) x (sum xs)
+    };
+main = sum list;
+

rlp.cabal

@@ -22,37 +22,52 @@ library
                     , TI
                     , GM
                     , Compiler.RLPC
+                    , Compiler.RlpcError
+                    , Compiler.JustRun
                     , Core.Syntax
                     , Core.Examples
                     , Core.Utils
                     , Core.TH
+                    , Core.HindleyMilner
+                    , Control.Monad.Errorful
+                    , Rlp.Syntax
+                    -- , Rlp.Parse.Decls
+                    , Rlp.Parse
+                    , Rlp.Parse.Associate
+                    , Rlp.Lex
+                    , Rlp.Parse.Types
 
     other-modules:    Data.Heap
                     , Data.Pretty
                     , Core.Parse
                     , Core.Lex
-                    , Control.Monad.Errorful
                     , Core2Core
-                    , RLP.Syntax
+                    , Control.Monad.Utils
 
     build-tool-depends: happy:happy, alex:alex
 
     -- other-extensions:
     build-depends:    base ^>=4.18.0.0
-                    , containers
-                    , microlens
-                    , microlens-th
-                    , mtl
-                    , template-haskell
                     -- required for happy
-                    , array
-                    , data-default-class
-                    , unordered-containers
-                    , hashable
-                    , pretty
-                    , recursion-schemes
-                    , megaparsec
-                    , text
+                    , array >= 0.5.5 && < 0.6
+                    , containers >= 0.6.7 && < 0.7
+                    , template-haskell >= 2.20.0 && < 2.21
+                    , pretty >= 1.1.3 && < 1.2
+                    , data-default >= 0.7.1 && < 0.8
+                    , data-default-class >= 0.1.2 && < 0.2
+                    , hashable >= 1.4.3 && < 1.5
+                    , mtl >= 2.3.1 && < 2.4
+                    , text >= 2.0.2 && < 2.1
+                    , megaparsec >= 9.6.1 && < 9.7
+                    , microlens >= 0.4.13 && < 0.5
+                    , microlens-mtl >= 0.2.0 && < 0.3
+                    , microlens-platform >= 0.4.3 && < 0.5
+                    , microlens-th >= 0.4.3 && < 0.5
+                    , unordered-containers >= 0.2.20 && < 0.3
+                    , recursion-schemes >= 5.2.2 && < 5.3
+                    , data-fix >= 0.3.2 && < 0.4
+                    , utf8-string >= 1.0.2 && < 1.1
+                    , extra >= 1.7.0 && < 2
 
     hs-source-dirs:   src
     default-language: GHC2021
@@ -64,11 +79,12 @@ executable rlpc
     -- other-extensions:
     build-depends: base ^>=4.18.0.0
                  , rlp
-                 , optparse-applicative
-                 , microlens
-                 , microlens-mtl
-                 , mtl
-                 , unordered-containers
+                 , optparse-applicative >= 0.18.1 && < 0.19
+                 , microlens >= 0.4.13 && < 0.5
+                 , microlens-mtl >= 0.2.0 && < 0.3
+                 , mtl >= 2.3.1 && < 2.4
+                 , unordered-containers >= 0.2.20 && < 0.3
+                 , text >= 2.0.2 && < 2.1
 
     hs-source-dirs:   app
     default-language: GHC2021
@@ -84,7 +100,9 @@ test-suite rlp-test
                  , rlp
                  , QuickCheck
                  , hspec            ==2.*
+                 , microlens
     other-modules: Arith
                  , GMSpec
+                 , Core.HindleyMilnerSpec
     build-tool-depends: hspec-discover:hspec-discover
 

src/Compiler/JustRun.hs

@@ -0,0 +1,46 @@
+{-|
+Module      : Compiler.JustRun
+Description : No-BS, high-level wrappers for major pipeline pieces.
+
+A collection of wrapper functions to demo processes such as lexing, parsing,
+type-checking, and evaluation. This module intends to export "no-BS" functions
+that use Prelude types such as @Either@ and @String@ rather than more complex
+types such as @RLPC@ or @Text@.
+-}
+module Compiler.JustRun
+    ( justLexSrc
+    , justParseSrc
+    , justTypeCheckSrc
+    )
+    where
+----------------------------------------------------------------------------------
+import Core.Lex
+import Core.Parse
+import Core.HindleyMilner
+import Core.Syntax                      (Program')
+import Compiler.RLPC
+import Control.Arrow                    ((>>>))
+import Control.Monad                    ((>=>))
+import Data.Text                        qualified as T
+import Data.Function                    ((&))
+import GM
+----------------------------------------------------------------------------------
+
+justLexSrc :: String -> Either RlpcError [CoreToken]
+justLexSrc s = lexCoreR (T.pack s)
+             & fmap (map $ \ (Located _ _ _ t) -> t)
+             & rlpcToEither
+
+justParseSrc :: String -> Either RlpcError Program'
+justParseSrc s = parse (T.pack s)
+               & rlpcToEither
+    where parse = lexCoreR >=> parseCoreProgR
+
+justTypeCheckSrc :: String -> Either RlpcError Program'
+justTypeCheckSrc s = typechk (T.pack s)
+                   & rlpcToEither
+    where typechk = lexCoreR >=> parseCoreProgR >=> checkCoreProgR
+
+rlpcToEither :: RLPC e a -> Either e a
+rlpcToEither = evalRLPC def >>> fmap fst
+

src/Compiler/RLPC.hs

@@ -13,9 +13,12 @@ errors and the family of RLPC monads.
 {-# LANGUAGE DeriveGeneric, DerivingStrategies, DerivingVia #-}
 module Compiler.RLPC
     ( RLPC
-    , RLPCT
+    , RLPCT(..)
     , RLPCIO
     , RLPCOptions(RLPCOptions)
+    , RlpcError(..)
+    , IsRlpcError(..)
+    , rlpc
     , addFatal
     , addWound
     , MonadErrorful
@@ -24,6 +27,9 @@ module Compiler.RLPC
     , evalRLPCT
     , evalRLPCIO
     , evalRLPC
+    , addRlpcWound
+    , addRlpcFatal
+    , liftRlpcErrs
     , rlpcLogFile
     , rlpcDebugOpts
     , rlpcEvaluator
@@ -42,6 +48,7 @@ import Control.Exception
 import Control.Monad.Reader
 import Control.Monad.State      (MonadState(state))
 import Control.Monad.Errorful
+import Compiler.RlpcError
 import Data.Functor.Identity
 import Data.Default.Class
 import GHC.Generics             (Generic)
@@ -93,7 +100,6 @@ evalRLPCIO o m = do
         -- TODO: errors
         Left e -> throwIO e
         Right a -> pure a
-    
 
 data RLPCOptions = RLPCOptions
     { _rlpcLogFile     :: Maybe FilePath
@@ -115,13 +121,24 @@ data Severity = Error
 -- temporary until we have a new doc building system
 type ErrorDoc = String
 
-class Diagnostic e where
-    errorDoc :: e -> ErrorDoc
-
 instance (Monad m) => MonadErrorful e (RLPCT e m) where
     addWound = RLPCT . lift . addWound
     addFatal = RLPCT . lift . addFatal
 
+liftRlpcErrs :: (IsRlpcError e, Monad m)
+             => RLPCT e m a -> RLPCT RlpcError m a
+liftRlpcErrs m = RLPCT . ReaderT $ \r ->
+    mapErrors liftRlpcErr $ runRLPCT >>> (`runReaderT` r) $ m
+
+addRlpcWound :: (IsRlpcError e, Monad m) => e -> RLPCT RlpcError m ()
+addRlpcWound = addWound . liftRlpcErr
+
+addRlpcFatal :: (IsRlpcError e, Monad m) => e -> RLPCT RlpcError m ()
+addRlpcFatal = addWound . liftRlpcErr
+
+rlpc :: (Monad m) => ErrorfulT e m a -> RLPCT e m a
+rlpc = RLPCT . ReaderT . const
+
 ----------------------------------------------------------------------------------
 
 instance Default RLPCOptions where

src/Compiler/RlpcError.hs

@@ -0,0 +1,15 @@
+module Compiler.RlpcError
+    ( RlpcError(..)
+    , IsRlpcError(..)
+    )
+    where
+----------------------------------------------------------------------------------
+import Control.Monad.Errorful
+----------------------------------------------------------------------------------
+
+data RlpcError = RlpcErr String -- temp
+    deriving Show
+
+class IsRlpcError a where
+    liftRlpcErr :: a -> RlpcError
+

src/Control/Monad/Errorful.hs

@@ -6,6 +6,7 @@ module Control.Monad.Errorful
     , runErrorfulT
     , Errorful
     , runErrorful
+    , mapErrors
     , MonadErrorful(..)
     )
     where
@@ -63,3 +64,10 @@ instance (Monad m) => Monad (ErrorfulT e m) where
             Right (a,es) -> runErrorfulT (k a)
             Left e       -> pure (Left e)
 
+mapErrors :: (Monad m) => (e -> e') -> ErrorfulT e m a -> ErrorfulT e' m a
+mapErrors f m = ErrorfulT $ do
+    x <- runErrorfulT m
+    case x of
+        Left  e      -> pure . Left  $ f e
+        Right (a,es) -> pure . Right $ (a, f <$> es)
+

src/Control/Monad/Utils.hs

@@ -0,0 +1,21 @@
+module Control.Monad.Utils
+    ( mapAccumLM
+    )
+    where
+----------------------------------------------------------------------------------
+import Data.Tuple               (swap)
+import Control.Monad.State
+----------------------------------------------------------------------------------
+
+-- | Monadic variant of @mapAccumL@
+
+mapAccumLM :: forall m t s a b. (Monad m, Traversable t)
+           => (s -> a -> m (s, b))
+           -> s
+           -> t a
+           -> m (s, t b)
+mapAccumLM k s t = swap <$> runStateT (traverse k' t) s
+    where
+        k' :: a -> StateT s m b
+        k' a = StateT $ fmap swap <$> flip k a
+

src/Core/Examples.hs

@@ -8,14 +8,13 @@ module Core.Examples
     ( fac3
     , sumList
     , constDivZero
+    , idCase
     ) where
 ----------------------------------------------------------------------------------
 import Core.Syntax
 import Core.TH
 ----------------------------------------------------------------------------------
 
--- TODO: my shitty lexer isn't inserting semicolons
-
 letrecExample :: Program'
 letrecExample = [coreProg|
     pair x y f = f x y;
@@ -181,30 +180,39 @@ constDivZero = [coreProg|
         main = k 3 ((/#) 1 0);
     |]
 
-corePrelude :: Module Name
-corePrelude = Module (Just ("Prelude", [])) $
-    -- non-primitive defs
-    [coreProg|
+idCase :: Program'
+idCase = [coreProg|
         id x = x;
-        k x y = x;
-        k1 x y = y;
-        s f g x = f x (g x);
-        compose f g x = f (g x);
-        twice f x = f (f x);
-        fst p = casePair# p k;
-        snd p = casePair# p k1;
-        head l = caseList# l abort# k;
-        tail l = caseList# l abort# k1;
-        _length_cc x xs = (+#) 1 (length xs);
-        length l = caseList# l 0 length_cc;
-    |]
-    <>
-    -- primitive constructors need some specialised wiring:
-    Program
-        [ ScDef "False" [] $ Con 0 0
-        , ScDef "True" [] $ Con 1 0
-        , ScDef "MkPair" [] $ Con 0 2
-        , ScDef "Nil" [] $ Con 1 0
-        , ScDef "Cons" [] $ Con 2 2
-        ]
+
+        main = id (case Pack{1 0} of
+            { 1 -> (+#) 2 3
+            })
+    |]
+
+-- corePrelude :: Module Name
+-- corePrelude = Module (Just ("Prelude", [])) $
+--     -- non-primitive defs
+--     [coreProg|
+--         id x = x;
+--         k x y = x;
+--         k1 x y = y;
+--         s f g x = f x (g x);
+--         compose f g x = f (g x);
+--         twice f x = f (f x);
+--         fst p = casePair# p k;
+--         snd p = casePair# p k1;
+--         head l = caseList# l abort# k;
+--         tail l = caseList# l abort# k1;
+--         _length_cc x xs = (+#) 1 (length xs);
+--         length l = caseList# l 0 length_cc;
+--     |]
+--     <>
+--     -- primitive constructors need some specialised wiring:
+--     Program
+--         [ ScDef "False" [] $ Con 0 0
+--         , ScDef "True" [] $ Con 1 0
+--         , ScDef "MkPair" [] $ Con 0 2
+--         , ScDef "Nil" [] $ Con 1 0
+--         , ScDef "Cons" [] $ Con 2 2
+--         ]
 

src/Core/HindleyMilner.hs

@@ -0,0 +1,220 @@
+{-|
+Module      : Core.HindleyMilner
+Description : Hindley-Milner type system
+-}
+{-# LANGUAGE LambdaCase #-}
+module Core.HindleyMilner
+    ( Context'
+    , infer
+    , check
+    , checkCoreProg
+    , checkCoreProgR
+    , TypeError(..)
+    , HMError
+    )
+    where
+----------------------------------------------------------------------------------
+import Lens.Micro
+import Lens.Micro.Mtl
+import Data.Maybe               (fromMaybe)
+import Data.Text                qualified as T
+import Data.HashMap.Strict      qualified as H
+import Data.Foldable            (traverse_)
+import Compiler.RLPC
+import Control.Monad            (foldM, void)
+import Control.Monad.Errorful   (Errorful, addFatal)
+import Control.Monad.State
+import Control.Monad.Utils      (mapAccumLM)
+import Core.Syntax
+----------------------------------------------------------------------------------
+
+-- | Annotated typing context -- I have a feeling we're going to want this in the
+-- future.
+type Context b = [(b, Type)]
+
+-- | Unannotated typing context, AKA our beloved Γ.
+type Context' = Context Name
+
+-- TODO: Errorful monad?
+
+-- | Type error enum.
+data TypeError
+    -- | Two types could not be unified
+    = TyErrCouldNotUnify Type Type
+    -- | @x@ could not be unified with @t@ because @x@ occurs in @t@
+    | TyErrRecursiveType Name Type
+    -- | Untyped, potentially undefined variable
+    | TyErrUntypedVariable Name
+    | TyErrMissingTypeSig Name
+    deriving (Show, Eq)
+
+-- TODO:
+instance IsRlpcError TypeError where
+    liftRlpcErr = RlpcErr . show
+
+-- | Synonym for @Errorful [TypeError]@. This means an @HMError@ action may
+-- throw any number of fatal or nonfatal errors. Run with @runErrorful@.
+type HMError = Errorful TypeError
+
+-- TODO: better errors. Errorful-esque, with cummulative errors instead of
+-- instantly dying.
+
+-- | Assert that an expression unifies with a given type
+--
+-- >>> let e = [coreProg|3|]
+-- >>> check [] (TyCon "Bool") e
+-- Left (TyErrCouldNotUnify (TyCon "Bool") (TyCon "Int#"))
+-- >>> check [] (TyCon "Int#") e
+-- Right ()
+
+check :: Context' -> Type -> Expr' -> HMError ()
+check g t1 e = do
+    t2 <- infer g e
+    void $ unify [(t1,t2)]
+
+-- | Typecheck program. I plan to allow for *some* inference in the future, but
+-- in the mean time all top-level binders must have a type annotation.
+checkCoreProg :: Program' -> HMError ()
+checkCoreProg p = scDefs
+            & traverse_ k
+    where
+        scDefs = p ^. programScDefs
+        g = gatherTypeSigs p
+
+        k :: ScDef' -> HMError ()
+        k sc = case lookup scname g of
+             Just t  -> check g t (sc ^. _rhs)
+             Nothing -> addFatal $ TyErrMissingTypeSig scname
+             where scname = sc ^. _lhs._1
+
+-- | @checkCoreProgR p@ returns @p@ if @p@ successfully typechecks.
+checkCoreProgR :: Program' -> RLPC RlpcError Program'
+checkCoreProgR p = do
+    liftRlpcErrs . rlpc . checkCoreProg $ p
+    pure p
+
+-- | Infer the type of an expression under some context.
+--
+-- >>> let g1 = [("id", TyVar "a" :-> TyVar "a")]
+-- >>> let g2 = [("id", (TyVar "a" :-> TyVar "a") :-> TyVar "a" :-> TyVar "a")]
+-- >>> infer g1 [coreExpr|id 3|]
+-- Right TyInt
+-- >>> infer g2 [coreExpr|id 3|]
+-- Left (TyErrCouldNotUnify (TyVar "a" :-> TyVar "a") TyInt)
+
+infer :: Context' -> Expr' -> HMError Type
+infer g e = do
+    (t,cs) <- gather g e
+    -- apply all unified constraints
+    foldr (uncurry subst) t <$> unify cs
+
+-- | A @Constraint@ between two types describes the requirement that the pair
+-- must unify
+type Constraint = (Type, Type)
+
+-- | Type of an expression under some context, and gather the constraints
+-- necessary to unify. Note that this is not the same as @infer@, as the
+-- expression will likely be given a fresh type variable along with a
+-- constraint, rather than the solved type.
+--
+-- For example, if the context says "@id@ has type a -> a," in an application of
+-- @id 3@, the whole application is assigned type @$a0@ and the constraint that
+-- @id@ must unify with type @Int -> $a0@ is generated.
+--
+-- >>> gather [("id", TyVar "a" :-> TyVar "a")] [coreExpr|id 3|]
+-- (TyVar "$a0",[(TyVar "a" :-> TyVar "a",TyInt :-> TyVar "$a0")])
+
+gather :: Context' -> Expr' -> HMError (Type, [Constraint])
+gather = \g e -> runStateT (go g e) ([],0) <&> \ (t,(cs,_)) -> (t,cs) where
+    go :: Context' -> Expr' -> StateT ([Constraint], Int) HMError Type
+    go g = \case
+        Lit (IntL _)  -> pure TyInt
+        Var k         -> lift $ maybe e pure $ lookup k g
+            where e = addFatal $ TyErrUntypedVariable k
+        App f x       -> do
+            tf <- go g f
+            tx <- go g x
+            tfx <- uniqueVar
+            addConstraint tf (tx :-> tfx)
+            pure tfx
+        Let NonRec bs e -> do
+            g' <- buildLetContext g bs
+            go g' e
+        -- TODO letrec, lambda, case
+
+    buildLetContext :: Context' -> [Binding']
+                    -> StateT ([Constraint], Int) HMError Context'
+    buildLetContext = foldM k where
+        k :: Context' -> Binding' -> StateT ([Constraint], Int) HMError Context'
+        k g (x := y) = do
+            ty <- go g y
+            pure ((x,ty) : g)
+
+uniqueVar :: StateT ([Constraint], Int) HMError Type
+uniqueVar = do
+    n <- use _2
+    _2 %= succ
+    pure (TyVar . T.pack $ '$' : 'a' : show n)
+
+addConstraint :: Type -> Type -> StateT ([Constraint], Int) HMError ()
+addConstraint t u = _1 %= ((t, u):)
+
+-- | Unify a list of constraints, meaning that pairs between types are turned
+-- into pairs of type variables and types. A useful thought model is to think of
+-- it as solving an equation such that the unknown variable is the left-hand
+-- side.
+
+unify :: [Constraint] -> HMError Context'
+unify = go mempty where
+    go :: Context' -> [Constraint] -> HMError Context'
+
+    -- nothing left! return accumulated context
+    go g [] = pure g
+
+    go g (c:cs) = case c of
+        -- primitives may of course unify with themselves
+        (TyInt,   TyInt)            -> go g cs
+
+        -- `x` unifies with `x`
+        (TyVar t, TyVar u) | t == u -> go g cs
+
+        -- a type variable `x` unifies with an arbitrary type `t` if `t` does
+        -- not reference `x`
+        (TyVar x, t)                -> unifyTV g x t cs
+        (t, TyVar x)                -> unifyTV g x t cs
+
+        -- two functions may be unified if their domain and codomain unify
+        (a :-> b, x :-> y)          -> go g $ (a,x) : (b,y) : cs
+
+        -- anything else is a failure :(
+        (t,u)                       -> addFatal $ TyErrCouldNotUnify t u
+
+    unifyTV :: Context' -> Name -> Type -> [Constraint] -> HMError Context'
+    unifyTV g x t cs | occurs t   = addFatal $ TyErrRecursiveType x t
+                     | otherwise  = go g' substed
+        where
+            g' = (x,t) : g
+            substed = cs & each . both %~ subst x t
+
+            occurs (a :-> b) = occurs a || occurs b
+            occurs (TyVar y)
+                | x == y     = True
+            occurs _         = False
+
+gatherTypeSigs :: Program b -> Context b
+gatherTypeSigs p = p ^. programTypeSigs
+                 & H.toList
+
+-- | The expression @subst x t e@ substitutes all occurences of @x@ in @e@ with
+-- @t@
+--
+-- >>> subst "a" (TyCon "Int") (TyVar "a")
+-- TyCon "Int"
+-- >>> subst "a" (TyCon "Int") (TyVar "a" :-> TyVar "a")
+-- TyCon "Int" :-> TyCon "Int"
+
+subst :: Name -> Type -> Type -> Type
+subst x t (TyVar y) | x == y  = t
+subst x t (a :-> b)           = subst x t a :-> subst x t b
+subst _ _ e                   = e
+

src/Core/Lex.x

@@ -3,8 +3,10 @@
 Module      : Core.Lex
 Description : Lexical analysis for the core language
 -}
+{-# LANGUAGE OverloadedStrings #-}
 module Core.Lex
     ( lexCore
+    , lexCoreR
     , lexCore'
     , CoreToken(..)
     , SrcError(..)
@@ -15,13 +17,17 @@ module Core.Lex
     where
 import Data.Char (chr)
 import Debug.Trace
+import Data.Text            (Text)
+import Data.Text            qualified as T
+import Data.String          (IsString(..))
 import Core.Syntax
 import Compiler.RLPC
+import Compiler.RlpcError
 import Lens.Micro
 import Lens.Micro.TH
 }
 
-%wrapper "monad"
+%wrapper "monad-strict-text"
 
 $whitechar = [ \t\n\r\f\v]
 $special   = [\(\)\,\;\[\]\{\}]
@@ -68,6 +74,7 @@ rlp :-
     "{"                     { constTok TokenLBrace }
     "}"                     { constTok TokenRBrace }
     ";"                     { constTok TokenSemicolon }
+    "::"                    { constTok TokenHasType }
     "@"                     { constTok TokenTypeApp }
     "{-#"                   { constTok TokenLPragma `andBegin` pragma }
 
@@ -80,7 +87,7 @@ rlp :-
     "where"                 { constTok TokenWhere }
     "Pack"                  { constTok TokenPack } -- temp
 
-    "\\"                    { constTok TokenLambda }
+    "\"                     { constTok TokenLambda }
     "λ"                     { constTok TokenLambda }
     "="                     { constTok TokenEquals }
     "->"                    { constTok TokenArrow }
@@ -90,7 +97,7 @@ rlp :-
     @varsym                 { lexWith TokenVarSym }
     @consym                 { lexWith TokenConSym }
 
-    @decimal                { lexWith (TokenLitInt . read @Int) }
+    @decimal                { lexWith (TokenLitInt . read @Int . T.unpack) }
 
     $white                  { skip }
     \n                      { skip }
@@ -134,10 +141,11 @@ data CoreToken = TokenLet
                | TokenLBrace
                | TokenRBrace
                | TokenSemicolon
+               | TokenHasType
                | TokenTypeApp
                | TokenLPragma
                | TokenRPragma
-               | TokenWord String
+               | TokenWord Text
                | TokenEOF
                deriving Show
 
@@ -155,11 +163,11 @@ data SrcErrorType = SrcErrLexical String
 
 type Lexer = AlexInput -> Int -> Alex (Located CoreToken)
 
-lexWith :: (String -> CoreToken) -> Lexer
-lexWith f (AlexPn _ y x,_,_,s) l = pure $ Located y x l (f $ take l s)
+lexWith :: (Text -> CoreToken) -> Lexer
+lexWith f (AlexPn _ y x,_,_,s) l = pure $ Located y x l (f $ T.take l s)
 
 -- | The main lexer driver.
-lexCore :: String -> RLPC SrcError [Located CoreToken]
+lexCore :: Text -> RLPC SrcError [Located CoreToken]
 lexCore s = case m of
     Left e   -> addFatal err
         where err = SrcError
@@ -171,9 +179,12 @@ lexCore s = case m of
     where
         m = runAlex s lexStream
 
+lexCoreR :: Text -> RLPC RlpcError [Located CoreToken]
+lexCoreR = liftRlpcErrs . lexCore
+
 -- | @lexCore@, but the tokens are stripped of location info. Useful for
 -- debugging
-lexCore' :: String -> RLPC SrcError [CoreToken]
+lexCore' :: Text -> RLPC SrcError [CoreToken]
 lexCore' s = fmap f <$> lexCore s
     where f (Located _ _ _ t) = t
 
@@ -188,6 +199,14 @@ data ParseError = ParErrLexical String
                 | ParErrParse
                 deriving Show
 
+-- TODO:
+instance IsRlpcError SrcError where
+    liftRlpcErr = RlpcErr . show
+
+-- TODO:
+instance IsRlpcError ParseError where
+    liftRlpcErr = RlpcErr . show
+
 alexEOF :: Alex (Located CoreToken)
 alexEOF = Alex $ \ st@(AlexState { alex_pos = AlexPn _ y x }) ->
     Right (st, Located y x 0 TokenEOF)

src/Core/Parse.y

@@ -3,10 +3,12 @@
 Module : Core.Parse
 Description : Parser for the Core language
 -}
+{-# LANGUAGE OverloadedStrings #-}
 module Core.Parse
     ( parseCore
     , parseCoreExpr
     , parseCoreProg
+    , parseCoreProgR
     , module Core.Lex -- temp convenience
     , parseTmp
     , SrcError
@@ -19,7 +21,12 @@ import Data.Foldable        (foldl')
 import Core.Syntax
 import Core.Lex
 import Compiler.RLPC
+import Lens.Micro
 import Data.Default.Class   (def)
+import Data.Hashable        (Hashable)
+import Data.Text.IO         qualified as TIO
+import Data.Text            qualified as T
+import Data.HashMap.Strict  qualified as H
 }
 
 %name parseCore Module
@@ -55,6 +62,7 @@ import Data.Default.Class   (def)
       '{-#'           { Located _ _ _ TokenLPragma }
       '#-}'           { Located _ _ _ TokenRPragma }
       ';'             { Located _ _ _ TokenSemicolon }
+      '::'            { Located _ _ _ TokenHasType }
       eof             { Located _ _ _ TokenEOF }
 
 %%
@@ -71,7 +79,17 @@ StandaloneProgram :: { Program Name }
 StandaloneProgram : Program eof                 { $1 }
 
 Program         :: { Program Name }
-Program         : ScDefs                        { Program $1 }
+Program         : ScTypeSig ';' Program         { insTypeSig $1 $3 }
+                | ScTypeSig OptSemi             { singletonTypeSig $1 }
+                | ScDef     ';' Program         { insScDef $1 $3 }
+                | ScDef     OptSemi             { singletonScDef $1 }
+
+OptSemi         :: { () }
+OptSemi         : ';'                           { () }
+                | {- epsilon -}                 { () }
+
+ScTypeSig       :: { (Name, Type) }
+ScTypeSig       : Var '::' Type                 { ($1,$3) }
 
 ScDefs          :: { [ScDef Name] }
 ScDefs          : ScDef ';' ScDefs              { $1 : $3 }
@@ -82,6 +100,16 @@ ScDefs          : ScDef ';' ScDefs              { $1 : $3 }
 ScDef           :: { ScDef Name }
 ScDef           : Var ParList '=' Expr          { ScDef $1 $2 $4 }
 
+Type            :: { Type }
+Type            : Type1                         { $1 }
+
+Type1           :: { Type }
+Type1           : '(' Type ')'                  { $2 }
+                | Type1 '->' Type               { $1 :-> $3 }
+                -- do we want to allow symbolic names for tyvars and tycons?
+                | varname                       { TyVar $1 }
+                | conname                       { TyCon $1 }
+
 ParList         :: { [Name] }
 ParList         : Var ParList                   { $1 : $2 }
                 | {- epsilon -}                 { [] }
@@ -123,7 +151,7 @@ Alter           :: { Alter Name }
 Alter           : litint ParList '->' Expr      { Alter (AltData $1) $2 $4 }
 
 Expr1           :: { Expr Name }
-Expr1           : litint                        { LitE $ IntL $1 }
+Expr1           : litint                        { Lit $ IntL $1 }
                 | Id                            { Var $1 }
                 | PackCon                       { $1 }
                 | ExprPragma                    { $1 }
@@ -133,8 +161,8 @@ ExprPragma      :: { Expr Name }
 ExprPragma      : '{-#' Words '#-}'      {% exprPragma $2 }
 
 Words           :: { [String] }
-Words           : word Words                    { $1 : $2 }
-                | word                          { [$1] }
+Words           : word Words                    { T.unpack $1 : $2 }
+                | word                          { [T.unpack $1] }
 
 PackCon         :: { Expr Name }
 PackCon         : pack '{' litint litint '}'    { Con $3 $4 }
@@ -171,7 +199,7 @@ parseError (Located y x l _ : _) = addFatal err
 
 parseTmp :: IO (Module Name)
 parseTmp = do
-    s <- readFile "/tmp/t.hs"
+    s <- TIO.readFile "/tmp/t.hs"
     case parse s of
         Left e -> error (show e)
         Right (ts,_) -> pure ts
@@ -190,5 +218,20 @@ exprPragma _           = addFatal err
 astPragma :: [String] -> RLPC SrcError (Expr Name)
 astPragma = pure . read . unwords
 
+insTypeSig :: (Hashable b) => (b, Type) -> Program b -> Program b
+insTypeSig ts = programTypeSigs %~ uncurry H.insert ts
+
+singletonTypeSig :: (Hashable b) => (b, Type) -> Program b
+singletonTypeSig ts = insTypeSig ts mempty
+
+insScDef :: (Hashable b) => ScDef b -> Program b -> Program b
+insScDef sc = programScDefs %~ (sc:)
+
+singletonScDef :: (Hashable b) => ScDef b -> Program b
+singletonScDef sc = insScDef sc mempty
+
+parseCoreProgR :: [Located CoreToken] -> RLPC RlpcError Program'
+parseCoreProgR = liftRlpcErrs . parseCoreProg
+
 }
 

src/Core/Syntax.hs

@@ -4,11 +4,15 @@ Description : Core ASTs and the like
 -}
 {-# LANGUAGE PatternSynonyms, OverloadedStrings #-}
 {-# LANGUAGE FunctionalDependencies #-}
+{-# LANGUAGE TemplateHaskell #-}
 module Core.Syntax
     ( Expr(..)
     , Type(..)
-    , Literal(..)
+    , pattern TyInt
+    , Lit(..)
     , pattern (:$)
+    , pattern (:@)
+    , pattern (:->)
     , Binding(..)
     , AltCon(..)
     , pattern (:=)
@@ -20,7 +24,9 @@ module Core.Syntax
     , Module(..)
     , Program(..)
     , Program'
+    , unliftScDef
     , programScDefs
+    , programTypeSigs
     , Expr'
     , ScDef'
     , Alter'
@@ -32,40 +38,51 @@ module Core.Syntax
 ----------------------------------------------------------------------------------
 import Data.Coerce
 import Data.Pretty
-import GHC.Generics
 import Data.List                    (intersperse)
 import Data.Function                ((&))
 import Data.String
+import Data.HashMap.Strict          qualified as H
+import Data.Hashable
+import Data.Text                    qualified as T
+import Data.Char
 -- Lift instances for the Core quasiquoters
 import Language.Haskell.TH.Syntax   (Lift)
+import Lens.Micro.TH                (makeLenses)
 import Lens.Micro
 ----------------------------------------------------------------------------------
 
 data Expr b = Var Name
-            | Con Tag Int -- Con Tag Arity
+            | Con Tag Int -- ^ Con Tag Arity
             | Case (Expr b) [Alter b]
             | Lam [b] (Expr b)
             | Let Rec [Binding b] (Expr b)
             | App (Expr b) (Expr b)
-            | LitE Literal
-            | Type Type
+            | Lit Lit
             deriving (Show, Read, Lift)
 
 deriving instance (Eq b) => Eq (Expr b)
 
-data Type = TyInt
-          | TyFun
+data Type = TyFun
           | TyVar Name
           | TyApp Type Type
-          | TyConApp TyCon [Type]
+          | TyCon Name
           deriving (Show, Read, Lift, Eq)
 
-type TyCon = Name
+pattern TyInt :: Type
+pattern TyInt = TyCon "Int#"
 
 infixl 2 :$
-pattern (:$) :: Expr b  -> Expr b  -> Expr b 
+pattern (:$) :: Expr b -> Expr b -> Expr b 
 pattern f :$ x = App f x
 
+infixl 2 :@
+pattern (:@) :: Type -> Type -> Type
+pattern f :@ x = TyApp f x
+
+infixr 1 :->
+pattern (:->) :: Type -> Type -> Type
+pattern a :-> b = TyApp (TyApp TyFun a) b
+
 {-# COMPLETE Binding :: Binding #-}
 {-# COMPLETE (:=) :: Binding #-}
 data Binding b = Binding b (Expr b)
@@ -87,27 +104,33 @@ data Rec = Rec
          deriving (Show, Read, Eq, Lift)
 
 data AltCon = AltData Tag
-            | AltLiteral Literal
+            | AltLit Lit
             | Default
             deriving (Show, Read, Eq, Lift)
 
-data Literal = IntL Int
+data Lit = IntL Int
     deriving (Show, Read, Eq, Lift)
 
-type Name = String
+type Name = T.Text
 type Tag = Int
 
 data ScDef b = ScDef b [b] (Expr b)
     deriving (Show, Lift)
 
+unliftScDef :: ScDef b -> Expr b
+unliftScDef (ScDef _ as e) = Lam as e
+
 data Module b = Module (Maybe (Name, [Name])) (Program b)
     deriving (Show, Lift)
 
-newtype Program b = Program [ScDef b]
+data Program b = Program
+    { _programScDefs   :: [ScDef b]
+    , _programTypeSigs :: H.HashMap b Type
+    }
     deriving (Show, Lift)
 
-programScDefs :: Lens' (Program b) [ScDef b]
-programScDefs = lens coerce (const coerce)
+makeLenses ''Program
+pure []
 
 type Program' = Program Name
 type Expr' = Expr Name
@@ -116,13 +139,20 @@ type Alter' = Alter Name
 type Binding' = Binding Name
 
 instance IsString (Expr b) where
-    fromString = Var
+    fromString = Var . fromString
 
-instance Semigroup (Program b) where
-    (<>) = coerce $ (<>) @[ScDef b] 
+instance IsString Type where
+    fromString "" = error "IsString Type string may not be empty"
+    fromString s
+        | isUpper c     = TyCon . fromString $ s
+        | otherwise     = TyVar . fromString $ s
+        where (c:_) = s
 
-instance Monoid (Program b) where
-    mempty = Program []
+instance (Hashable b) => Semigroup (Program b) where
+    (<>) = undefined
+
+instance (Hashable b) => Monoid (Program b) where
+    mempty = Program mempty mempty
 
 ----------------------------------------------------------------------------------
 

src/Core/TH.hs

@@ -5,20 +5,25 @@ Description : Core quasiquoters
 module Core.TH
     ( coreExpr
     , coreProg
+    , coreProgT
     , core
     )
     where
 ----------------------------------------------------------------------------------
 import Language.Haskell.TH
-import Language.Haskell.TH.Syntax hiding (Module)
+import Language.Haskell.TH.Syntax   hiding (Module)
 import Language.Haskell.TH.Quote
 import Control.Monad                ((>=>))
 import Compiler.RLPC
 import Data.Default.Class           (def)
+import Data.Text                    qualified as T
 import Core.Parse
 import Core.Lex
+import Core.HindleyMilner           (checkCoreProgR)
 ----------------------------------------------------------------------------------
 
+-- TODO: write in terms of a String -> QuasiQuoter
+
 core :: QuasiQuoter
 core = QuasiQuoter
     { quoteExp = qCore
@@ -43,24 +48,40 @@ coreExpr = QuasiQuoter
     , quoteDec = error "core quasiquotes may only be used in expressions"
     }
 
+-- | Type-checked @coreProg@
+coreProgT :: QuasiQuoter
+coreProgT = QuasiQuoter
+    { quoteExp = qCoreProgT
+    , quotePat = error "core quasiquotes may only be used in expressions"
+    , quoteType = error "core quasiquotes may only be used in expressions"
+    , quoteDec = error "core quasiquotes may only be used in expressions"
+    }
+
 qCore :: String -> Q Exp
-qCore s = case parse s of
+qCore s = case parse (T.pack s) of
     Left e       -> error (show e)
     Right (m,ts) -> lift m
     where
         parse = evalRLPC def . (lexCore >=> parseCore)
 
 qCoreExpr :: String -> Q Exp
-qCoreExpr s = case parseExpr s of
+qCoreExpr s = case parseExpr (T.pack s) of
     Left e       -> error (show e)
     Right (m,ts) -> lift m
     where
         parseExpr = evalRLPC def . (lexCore >=> parseCoreExpr)
 
 qCoreProg :: String -> Q Exp
-qCoreProg s = case parseProg s of
+qCoreProg s = case parse (T.pack s) of
     Left e       -> error (show e)
     Right (m,ts) -> lift m
     where
-        parseProg = evalRLPC def . (lexCore >=> parseCoreProg)
+        parse = evalRLPC def . (lexCoreR >=> parseCoreProgR)
+
+qCoreProgT :: String -> Q Exp
+qCoreProgT s = case parse (T.pack s) of
+    Left e      -> error (show e)
+    Right (m,_) -> lift m
+    where
+        parse = evalRLPC def . (lexCoreR >=> parseCoreProgR >=> checkCoreProgR)
 

src/Core/Utils.hs

@@ -7,7 +7,7 @@ module Core.Utils
     ( bindersOf
     , rhssOf
     , isAtomic
-    , insertModule
+    -- , insertModule
     , extractProgram
     , freeVariables
     , ExprF(..)
@@ -19,6 +19,7 @@ import Data.Functor.Foldable
 import Data.Set                     (Set)
 import Data.Set                     qualified as S
 import Core.Syntax
+import Lens.Micro
 import GHC.Exts                     (IsList(..))
 ----------------------------------------------------------------------------------
 
@@ -32,14 +33,14 @@ rhssOf = fromList . fmap f
 
 isAtomic :: Expr b -> Bool
 isAtomic (Var _)  = True
-isAtomic (LitE _) = True
+isAtomic (Lit _)  = True
 isAtomic _        = False
 
 ----------------------------------------------------------------------------------
 
 -- TODO: export list awareness
-insertModule :: Module b -> Program b -> Program b
-insertModule (Module _ m) p = p <> m
+-- insertModule :: Module b -> Program b -> Program b
+-- insertModule (Module _ p) = programScDefs %~ (<>m)
 
 extractProgram :: Module b -> Program b
 extractProgram (Module _ p) = p

src/Core2Core.hs

@@ -17,6 +17,7 @@ import Data.List
 import Control.Monad.Writer
 import Control.Monad.State
 import Control.Arrow            ((>>>))
+import Data.Text                qualified as T
 import Numeric                  (showHex)
 import Lens.Micro
 import Core.Syntax
@@ -27,7 +28,7 @@ core2core :: Program' -> Program'
 core2core p = undefined
 
 gmPrep :: Program' -> Program'
-gmPrep p = p' <> Program caseScs
+gmPrep p = p' & programScDefs %~ (<>caseScs)
     where
         rhss :: Applicative f => (Expr z -> f (Expr z)) -> Program z -> f (Program z)
         rhss = programScDefs . each . _rhs
@@ -46,8 +47,7 @@ type Floater = StateT [Name] (Writer [ScDef'])
 runFloater :: Floater a -> (a, [ScDef'])
 runFloater = flip evalStateT ns >>> runWriter
     where
-        -- TODO: safer, uncapturable names
-        ns = [ "nonstrict_case_" ++ showHex n "" | n <- [0..] ]
+        ns = [ T.pack $ "$nonstrict_case_" ++ showHex n "" | n <- [0..] ]
 
 -- TODO: formally define a "strict context" and reference that here
 -- the returned ScDefs are guaranteed to be free of non-strict cases.
@@ -56,7 +56,7 @@ floatNonStrictCases g = goE
     where
         goE :: Expr' -> Floater Expr'
         goE (Var k)             = pure (Var k)
-        goE (LitE l)            = pure (LitE l)
+        goE (Lit l)             = pure (Lit l)
         goE (Case e as)         = pure (Case e as)
         goE (Let Rec bs e)      = Let Rec <$> bs' <*> goE e
             where bs' = travBs goE bs
@@ -78,7 +78,7 @@ floatNonStrictCases g = goE
         goC (f :$ x)            = (:$) <$> goC f <*> goC x
         goC (Let r bs e)        = Let r <$> bs' <*> goE e
             where bs' = travBs goC bs
-        goC (LitE l)            = pure (LitE l)
+        goC (Lit l)             = pure (Lit l)
         goC (Var k)             = pure (Var k)
         goC (Con t as)          = pure (Con t as)
 

src/GM.hs

@@ -22,7 +22,10 @@ import Data.Maybe                   (fromMaybe, mapMaybe)
 import Data.Monoid                  (Endo(..))
 import Data.Tuple                   (swap)
 import Lens.Micro
+import Lens.Micro.Extras            (view)
 import Lens.Micro.TH
+import Lens.Micro.Platform          (packed, unpacked)
+import Lens.Micro.Platform.Internal (IsText(..))
 import Text.Printf
 import Text.PrettyPrint             hiding ((<>))
 import Text.PrettyPrint.HughesPJ    (maybeParens)
@@ -281,7 +284,7 @@ step st = case head (st ^. gmCode) of
                 m = st ^. gmEnv
                 s = st ^. gmStack
                 h = st ^. gmHeap
-                n' = show n
+                n' = show n ^. packed
 
         -- Core Rule 2. (no sharing)
         -- pushIntI :: Int -> GmState
@@ -582,7 +585,7 @@ compiledPrims =
         binop k i = (k, 2, [Push 1, Eval, Push 1, Eval, i, Update 2, Pop 2, Unwind])
 
 buildInitialHeap :: Program' -> (GmHeap, Env)
-buildInitialHeap (Program ss) = mapAccumL allocateSc mempty compiledScs
+buildInitialHeap (view programScDefs -> ss) = mapAccumL allocateSc mempty compiledScs
     where
         compiledScs = fmap compileSc ss <> compiledPrims
 
@@ -612,12 +615,13 @@ buildInitialHeap (Program ss) = mapAccumL allocateSc mempty compiledScs
             | k `elem` domain  = [Push n]
             | otherwise        = [PushGlobal k]
             where
-                n = fromMaybe (error $ "undeclared var: " <> k) $ lookupN k g
+                n = fromMaybe err $ lookupN k g
+                err = error $ "undeclared var: " <> (k ^. unpacked)
                 domain = f `mapMaybe` g
                 f (NameKey n, _) = Just n
                 f _              = Nothing
 
-        compileC _ (LitE l)  = compileCL l
+        compileC _ (Lit l)   = compileCL l
 
         -- >> [ref/compileC]
         compileC g (App f x) = compileC g x
@@ -661,16 +665,16 @@ buildInitialHeap (Program ss) = mapAccumL allocateSc mempty compiledScs
 
         compileC _ _ = error "yet to be implemented!"
 
-        compileCL :: Literal -> Code
+        compileCL :: Lit -> Code
         compileCL (IntL n) = [PushInt n]
 
-        compileEL :: Literal -> Code
+        compileEL :: Lit -> Code
         compileEL (IntL n) = [PushInt n]
 
         -- compile an expression in a strict context such that a pointer to the
         -- expression is left on top of the stack in WHNF
         compileE :: Env -> Expr' -> Code
-        compileE _ (LitE l) = compileEL l
+        compileE _ (Lit l) = compileEL l
         compileE g (Let NonRec bs e) =
                 -- we use compileE instead of compileC
                 mconcat binders <> compileE g' e <> [Slide d]
@@ -738,8 +742,8 @@ buildInitialHeap (Program ss) = mapAccumL allocateSc mempty compiledScs
         argOffset :: Int -> Env -> Env
         argOffset n = each . _2 %~ (+n)
 
-idPack :: Tag -> Int -> String
-idPack t n = printf "Pack{%d %d}" t n
+showCon :: (IsText a) => Tag -> Int -> a
+showCon t n = printf "Pack{%d %d}" t n ^. packed
 
 ----------------------------------------------------------------------------------
 
@@ -855,12 +859,12 @@ showNodeAt = showNodeAtP 0
 showNodeAtP :: Int -> GmState -> Addr -> Doc
 showNodeAtP p st a = case hLookup a h of
     Just (NNum n)        -> int n <> "#"
-    Just (NGlobal _ _)   -> text name
+    Just (NGlobal _ _)   -> textt name
         where
             g = st ^. gmEnv
             name = case lookup a (swap <$> g) of
                 Just (NameKey n)     -> n
-                Just (ConstrKey t n) -> idPack t n
+                Just (ConstrKey t n) -> showCon t n
                 _                    -> errTxtInvalidAddress
     -- TODO: left-associativity
     Just (NAp f x)       -> pprec $ showNodeAtP (p+1) st f
@@ -877,7 +881,7 @@ showNodeAtP p st a = case hLookup a h of
         pprec = maybeParens (p > 0)
 
 showSc :: GmState -> (Name, Addr) -> Doc
-showSc st (k,a) = "Supercomb " <> qquotes (text k) <> colon
+showSc st (k,a) = "Supercomb " <> qquotes (textt k) <> colon
                $$ code
     where
         code = case hLookup a (st ^. gmHeap) of
@@ -900,6 +904,9 @@ showInstr (CaseJump alts) = "CaseJump" $$ nest pprTabstop alternatives
         alternatives = foldr (\a acc -> showAlt a $$ acc) mempty alts
 showInstr i = text $ show i
 
+textt :: (IsText a) => a -> Doc
+textt t = t ^. unpacked & text
+
 ----------------------------------------------------------------------------------
 
 lookupN :: Name -> Env -> Maybe Addr
@@ -975,7 +982,8 @@ resultOf p = do
         h = st ^. gmHeap
 
 resultOfExpr :: Expr' -> Maybe Node
-resultOfExpr e = resultOf $ Program
-    [ ScDef "main" [] e
-    ]
+resultOfExpr e = resultOf $
+    mempty & programScDefs .~
+                [ ScDef "main" [] e
+                ]
 

src/RLP/Syntax.hs

@@ -1,59 +0,0 @@
-{-# LANGUAGE OverloadedStrings #-}
-module RLP.Syntax
-    ( RlpExpr
-    )
-    where
-----------------------------------------------------------------------------------
-import Data.Text                    (Text)
-import Lens.Micro
-import Core                         (HasRHS(..), HasLHS(..))
-----------------------------------------------------------------------------------
-
-newtype RlpProgram b = RlpProgram [Decl b]
-
-data Decl b = InfixD  InfixAssoc Int VarId
-            | FunD    VarId [Pat b] (RlpExpr b)
-            | DataD   ConId [ConId] [ConAlt]
-
-data ConAlt = ConAlt ConId [ConId]
-
-data InfixAssoc = Assoc | AssocL | AssocR
-
-data RlpExpr b = LetE  [Bind b] (RlpExpr b)
-               | VarE  VarId
-               | ConE  ConId
-               | LamE  [Pat b] (RlpExpr b)
-               | CaseE (RlpExpr b) [Alt b]
-               | IfE   (RlpExpr b) (RlpExpr b) (RlpExpr b)
-               | AppE  (RlpExpr b) (RlpExpr b)
-               | LitE  (Lit b)
-
--- do we want guards?
-data Alt b = AltA (Pat b) (RlpExpr b)
-
-data Bind b = PatB (Pat b) (RlpExpr b)
-            | FunB VarId [Pat b] (RlpExpr b)
-
-data VarId = NameVar Text
-           | SymVar Text
-
-data ConId = NameCon Text
-           | SymCon Text
-
-data Pat b = VarP VarId
-           | LitP (Lit b)
-           | ConP ConId [Pat b]
-
-data Lit b = IntL Int
-           | CharL Char
-           | ListL [RlpExpr b]
-
--- instance HasLHS Alt Alt Pat Pat where
---     _lhs = lens
---         (\ (AltA p _) -> p)
---         (\ (AltA _ e) p' -> AltA p' e)
-
--- instance HasRHS Alt Alt RlpExpr RlpExpr where
---     _rhs = lens
---         (\ (AltA _ e) -> e)
---         (\ (AltA p _) e' -> AltA p e')

src/Rlp/Lex.x

@@ -0,0 +1,343 @@
+{
+{-# LANGUAGE ViewPatterns, LambdaCase #-}
+{-# LANGUAGE GeneralisedNewtypeDeriving #-}
+{-# LANGUAGE OverloadedStrings #-}
+module Rlp.Lex
+    ( P(..)
+    , RlpToken(..)
+    , Located(..)
+    , lexToken
+    , lexDebug
+    , lexCont
+    )
+    where
+import Codec.Binary.UTF8.String (encodeChar)
+import Control.Monad
+import Core.Syntax              (Name)
+import Data.Functor.Identity
+import Data.Char                (digitToInt)
+import Data.Monoid              (First)
+import Data.Maybe
+import Data.Text                (Text)
+import Data.Text                qualified as T
+import Data.Word
+import Data.Default
+import Lens.Micro.Mtl
+import Lens.Micro
+
+import Debug.Trace
+import Rlp.Parse.Types
+}
+
+$whitechar      = [ \t\n\r\f\v]
+
+$nl             = [\n\r]
+$white_no_nl    = $white # $nl
+
+$lower          = [a-z \_]
+$upper          = [A-Z]
+$alpha          = [$lower $upper]
+$digit          = 0-9
+
+$special        = [\(\)\,\;\[\]\{\}]
+$namechar       = [$alpha $digit \' \#]
+$asciisym       = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~\:]
+
+@decimal        = $digit+
+
+@varname        = $lower $namechar*
+@conname        = $upper $namechar*
+@consym         = \: $asciisym*
+@varsym         = $asciisym+
+
+@reservedname = 
+    case|data|do|import|in|let|letrec|module|of|where
+
+@reservedop =
+    "=" | \\ | "->" | "|"
+
+rlp :-
+    
+-- everywhere: skip whitespace
+$white_no_nl+       ;
+
+-- everywhere: skip comments
+-- TODO: don't treat operators like (-->) as comments
+"--".*              ;
+
+-- we are indentation-sensitive! do not skip NLs!. upon encountering a newline,
+-- we check indentation and potentially insert extra tokens. search this file
+-- for the definition of `doBol`
+<0> \n                  { beginPush bol }
+
+-- scan various identifiers and reserved words. order is important here!
+<0>
+{
+    @reservedname       { tokenWith lexReservedName }
+    @conname            { tokenWith TokenConName }
+    @varname            { tokenWith TokenVarName }
+    @reservedop         { tokenWith lexReservedOp }
+    @consym             { tokenWith TokenConSym }
+    @varsym             { tokenWith TokenVarSym }
+}
+
+-- literals -- currently this is just unsigned integer literals
+<0>
+{
+    @decimal            { tokenWith (TokenLitInt . readInt) }
+}
+
+-- control characters
+<0>
+{
+    "{"                 { explicitLBrace }
+    "}"                 { explicitRBrace }
+    ";"                 { constToken TokenSemicolon }
+}
+
+-- consume all whitespace leaving us at the beginning of the next non-empty
+-- line. we then compare the indentation of that line to the enclosing layout
+-- context and proceed accordingly
+<bol>
+{
+    $whitechar          ;
+    \n                  ;
+    ()                  { doBol }
+}
+
+<layout_top>
+{
+    \n                  ;
+    "{"                 { explicitLBrace `thenDo` popLexState }
+    ()                  { doLayout }
+}
+
+{
+
+lexReservedName :: Text -> RlpToken
+lexReservedName = \case
+    "data"      -> TokenData
+    "case"      -> TokenCase
+    "of"        -> TokenOf
+    "let"       -> TokenLet
+    "in"        -> TokenIn
+
+lexReservedOp :: Text -> RlpToken
+lexReservedOp = \case
+    "="     -> TokenEquals
+    "::"    -> TokenHasType
+    "|"     -> TokenPipe
+
+-- | @andBegin@, with the subtle difference that the start code is set
+--   /after/ the action
+thenBegin :: LexerAction a -> Int -> LexerAction a
+thenBegin act c inp l = do
+    a <- act inp l
+    psLexState . _head .= c
+    pure a
+
+andBegin :: LexerAction a -> Int -> LexerAction a
+andBegin act c inp l = do
+    psLexState . _head .= c
+    act inp l
+
+beginPush :: Int -> LexerAction (Located RlpToken)
+beginPush n _ _ = pushLexState n >> lexToken
+
+alexGetByte :: AlexInput -> Maybe (Word8, AlexInput)
+alexGetByte inp = case inp ^. aiBytes of
+    [] -> do
+        (c,t) <- T.uncons (inp ^. aiSource)
+        let (b:bs) = encodeChar c
+                       -- tail the source
+            inp' = inp & aiSource .~ t
+                       -- record the excess bytes for successive calls
+                       & aiBytes .~ bs
+                       -- report the previous char
+                       & aiPrevChar .~ c
+                       -- update the position
+                       & aiPos %~ \ (ln,col) ->
+                                   if c == '\n'
+                                   then (ln+1,1)
+                                   else (ln,col+1)
+        pure (b, inp')
+    
+    _ -> Just (head bs, inp')
+        where
+            (bs, inp') = inp & aiBytes <<%~ drop 1
+
+getInput :: P AlexInput
+getInput = use psInput
+
+getLexState :: P Int
+getLexState = use (psLexState . singular _head)
+
+alexInputPrevChar :: AlexInput -> Char
+alexInputPrevChar = view aiPrevChar
+
+pushLexState :: Int -> P ()
+pushLexState n = psLexState %= (n:)
+
+readInt :: Text -> Int
+readInt = T.foldr f 0 where
+    f c n = digitToInt c + 10*n
+
+constToken :: RlpToken -> LexerAction (Located RlpToken)
+constToken t inp l = do
+    pos <- use (psInput . aiPos)
+    pure (Located (pos,l) t)
+
+tokenWith :: (Text -> RlpToken) -> LexerAction (Located RlpToken)
+tokenWith tf inp l = do
+    pos <- getPos
+    let t = tf (T.take l $ inp ^. aiSource)
+    pure (Located (pos,l) t)
+
+getPos :: P Position
+getPos = use (psInput . aiPos)
+
+alexEOF :: P (Located RlpToken)
+alexEOF = do
+    inp <- getInput
+    pure (Located undefined TokenEOF)
+
+execP :: P a -> ParseState -> Maybe a
+execP p st = runP p st & snd
+
+execP' :: P a -> Text -> Maybe a
+execP' p s = execP p st where
+    st = initParseState s
+
+initParseState :: Text -> ParseState
+initParseState s = ParseState
+    { _psLayoutStack = []
+    -- IMPORTANT: the initial state is `bol` to begin the top-level layout,
+    -- which then returns to state 0 which continues the normal lexing process.
+    , _psLexState = [layout_top,0]
+    , _psInput = initAlexInput s
+    , _psOpTable = mempty
+    }
+
+initAlexInput :: Text -> AlexInput
+initAlexInput s = AlexInput
+    { _aiPrevChar   = '\0'
+    , _aiSource     = s
+    , _aiBytes      = []
+    , _aiPos        = (1,1)
+    }
+
+lexToken :: P (Located RlpToken)
+lexToken = do
+    inp <- getInput
+    c <- getLexState
+    st <- use id
+    -- traceM $ "st: " <> show st
+    case alexScan inp c of
+        AlexEOF -> pure $ Located (inp ^. aiPos, 0) TokenEOF
+        AlexSkip inp' l -> do
+            psInput .= inp'
+            lexToken
+        AlexToken inp' l act -> do
+            psInput .= inp'
+            act inp l
+
+lexCont :: (Located RlpToken -> P a) -> P a
+lexCont = (lexToken >>=)
+
+lexStream :: P [RlpToken]
+lexStream = do
+    t <- lexToken
+    case t of
+        Located _ TokenEOF -> pure [TokenEOF]
+        Located _ t        -> (t:) <$> lexStream
+
+lexDebug :: (Located RlpToken -> P a) -> P a
+lexDebug k = do
+    t <- lexToken
+    traceM $ "token: " <> show t
+    k t
+
+lexTest :: Text -> Maybe [RlpToken]
+lexTest s = execP' lexStream s
+
+indentLevel :: P Int
+indentLevel = do
+    pos <- use (psInput . aiPos)
+    pure (pos ^. _2)
+
+insertToken :: RlpToken -> P (Located RlpToken)
+insertToken t = do
+    pos <- use (psInput . aiPos)
+    pure (Located (pos, 0) t)
+
+popLayout :: P Layout
+popLayout = do
+    -- traceM "pop layout"
+    ctx <- preuse (psLayoutStack . _head)
+    psLayoutStack %= (drop 1)
+    case ctx of
+        Just l      -> pure l
+        Nothing     -> error "uhh"
+
+pushLayout :: Layout -> P ()
+pushLayout l = do
+    -- traceM "push layout"
+    psLayoutStack %= (l:)
+
+popLexState :: P ()
+popLexState = do
+    psLexState %= tail
+
+insertSemicolon, insertLBrace, insertRBrace :: P (Located RlpToken)
+insertSemicolon = {- traceM "inserting semi"   >> -} insertToken TokenSemicolonV
+insertLBrace    = {- traceM "inserting lbrace" >> -} insertToken TokenLBraceV
+insertRBrace    = {- traceM "inserting rbrace" >> -} insertToken TokenRBraceV
+
+cmpLayout :: P Ordering
+cmpLayout = do
+    i <- indentLevel
+    ctx <- preuse (psLayoutStack . _head)
+    case ctx of
+        Just (Implicit n) -> pure (i `compare` n)
+        _                 -> pure GT
+
+doBol :: LexerAction (Located RlpToken)
+doBol inp l = do
+    off <- cmpLayout
+    i <- indentLevel
+    traceM $ "i: " <> show i
+    -- important that we pop the lex state lest we find our lexer diverging
+    popLexState
+    case off of
+        -- the line is aligned with the previous. it therefore belongs to the
+        -- same list
+        EQ -> insertSemicolon
+        -- the line is indented further than the previous, so we assume it is a
+        -- line continuation. ignore it and move on!
+        GT -> lexToken
+        -- the line is indented less than the previous, pop the layout stack and
+        -- insert a closing brace.
+        LT -> popLayout >> insertRBrace
+
+thenDo :: LexerAction a -> P b -> LexerAction a
+thenDo act p inp l = act inp l <* p
+
+explicitLBrace :: LexerAction (Located RlpToken)
+explicitLBrace inp l = do
+    pushLayout Explicit
+    constToken TokenLBrace inp l
+
+explicitRBrace :: LexerAction (Located RlpToken)
+explicitRBrace inp l = do
+    popLayout
+    constToken TokenRBrace inp l
+
+doLayout :: LexerAction (Located RlpToken)
+doLayout _ _ = do
+    i <- indentLevel
+    pushLayout (Implicit i)
+    popLexState
+    insertLBrace
+
+}
+

src/Rlp/Parse.y

@@ -0,0 +1,134 @@
+{
+module Rlp.Parse
+    ( parseRlpProgram
+    )
+    where
+import Rlp.Lex
+import Rlp.Syntax
+import Rlp.Parse.Types
+import Rlp.Parse.Associate
+import Lens.Micro.Mtl
+import Data.List.Extra
+import Data.Fix
+import Data.Functor.Const
+}
+
+%name parseRlpProgram StandaloneProgram
+
+%monad { P }
+%lexer { lexDebug } { Located _ TokenEOF }
+%error { parseError }
+%tokentype { Located RlpToken }
+
+%token
+    varname         { Located _ (TokenVarName $$) }
+    conname         { Located _ (TokenConName $$) }
+    data            { Located _ TokenData }
+    litint          { Located _ (TokenLitInt $$) }
+    '='             { Located _ TokenEquals }
+    '|'             { Located _ TokenPipe }
+    ';'             { Located _ TokenSemicolon }
+    '('             { Located _ TokenLParen }
+    ')'             { Located _ TokenRParen }
+    '->'            { Located _ TokenArrow }
+    vsemi           { Located _ TokenSemicolonV }
+    '{'             { Located _ TokenLBrace }
+    '}'             { Located _ TokenRBrace }
+    vlbrace         { Located _ TokenLBraceV }
+    vrbrace         { Located _ TokenRBraceV }
+
+%right '->'
+
+%%
+
+StandaloneProgram   :: { RlpProgram' }
+StandaloneProgram   : '{' Decls  '}'        {% mkProgram $2 }
+                    | VL  DeclsV VR         {% mkProgram $2 }
+
+VL  :: { () }
+VL  : vlbrace       { () }
+
+VR  :: { () }
+VR  : vrbrace       { () }
+    | error         { () }
+
+Decls               :: { [PartialDecl'] }
+Decls               : Decl ';' Decls        { $1 : $3 }
+                    | Decl ';'              { [$1] }
+                    | Decl                  { [$1] }
+
+DeclsV              :: { [PartialDecl'] }
+DeclsV              : Decl VS Decls         { $1 : $3 }
+                    | Decl VS               { [$1] }
+                    | Decl                  { [$1] }
+
+VS                  :: { Located RlpToken }
+VS                  : ';'                   { $1 }
+                    | vsemi                 { $1 }
+
+Decl        :: { PartialDecl' }
+Decl        : FunDecl                   { $1 }
+            | DataDecl                  { $1 }
+
+DataDecl    :: { PartialDecl' }
+            : data Con TyParams '=' DataCons    { DataD $2 $3 $5 }
+
+TyParams    :: { [Name] }
+            : {- epsilon -}             { [] }
+            | TyParams varname          { $1 `snoc` $2 }
+
+DataCons    :: { [ConAlt] }
+            : DataCons '|' DataCon      { $1 `snoc` $3 }
+            | DataCon                   { [$1] }
+
+DataCon     :: { ConAlt }
+            : Con Type1s                { ConAlt $1 $2 }
+
+Type1s      :: { [Type] }
+            : {- epsilon -}             { [] }
+            | Type1s Type1              { $1 `snoc` $2 }
+
+Type1       :: { Type }
+            : '(' Type ')'              { $2 }
+            | conname                   { TyCon $1 }
+            | varname                   { TyVar $1 }
+
+Type        :: { Type }
+            : Type '->' Type            { $1 :-> $3 }
+            | Type1                     { $1 }
+
+FunDecl     :: { PartialDecl' }
+FunDecl     : Var Params '=' Expr       { FunD $1 $2 (Const $4) Nothing }
+
+Params      :: { [Pat'] }
+Params      : {- epsilon -}             { [] }
+            | Params Pat1               { $1 `snoc` $2 }
+
+Pat1        :: { Pat' }
+            : Var                       { VarP $1 }
+            | Lit                       { LitP $1 }
+
+Expr        :: { PartialExpr' }
+Expr        : Lit                       { Fix . E $ LitEF $1 }
+            | Var                       { Fix . E $ VarEF $1 }
+
+Lit         :: { Lit' }
+Lit         : litint                    { IntL $1 }
+
+Var         :: { VarId }
+Var         : varname                   { NameVar $1 }
+
+Con         :: { ConId }
+            : conname                   { NameCon $1 }
+
+{
+
+mkProgram :: [PartialDecl'] -> P RlpProgram'
+mkProgram ds = do
+    pt <- use psOpTable
+    pure $ RlpProgram (associate pt <$> ds)
+
+parseError :: Located RlpToken -> P a
+parseError = error . show
+
+}

src/Rlp/Parse/Associate.hs

@@ -0,0 +1,99 @@
+{-# LANGUAGE PatternSynonyms, ViewPatterns, ImplicitParams #-}
+module Rlp.Parse.Associate
+    ( associate
+    )
+    where
+--------------------------------------------------------------------------------
+import Data.HashMap.Strict              qualified as H
+import Data.Functor.Foldable
+import Data.Functor.Const
+import Lens.Micro
+import Rlp.Parse.Types
+import Rlp.Syntax
+--------------------------------------------------------------------------------
+
+associate :: OpTable -> PartialDecl' -> Decl' RlpExpr
+associate pt (FunD n as b w)    = FunD n as b' w
+    where b' = let ?pt = pt in completeExpr (getConst b)
+associate pt (TySigD ns t)      = TySigD ns t
+associate pt (DataD n as cs)    = DataD n as cs
+associate pt (InfixD a p n)     = InfixD a p n
+
+completeExpr :: (?pt :: OpTable) => PartialExpr' -> RlpExpr'
+completeExpr = cata completePartial
+
+completePartial :: (?pt :: OpTable) => PartialE -> RlpExpr'
+completePartial (E e)        = completeRlpExpr e
+completePartial p@(B o l r)  = completeB (build p)
+completePartial (Par e)      = completePartial e
+
+completeRlpExpr :: (?pt :: OpTable) => RlpExprF' RlpExpr' -> RlpExpr'
+completeRlpExpr = embed
+
+completeB :: (?pt :: OpTable) => PartialE -> RlpExpr'
+completeB p = case build p of
+    B o l r     -> (o' `AppE` l') `AppE` r'
+        where
+            -- TODO: how do we know it's symbolic?
+            o' = VarE (SymVar o)
+            l' = completeB l
+            r' = completeB r
+    Par e       -> completeB e
+    E e         -> completeRlpExpr e
+
+build :: (?pt :: OpTable) => PartialE -> PartialE
+build e = go id e (rightmost e) where
+    rightmost :: PartialE -> PartialE
+    rightmost (B _ _ r) = rightmost r
+    rightmost p@(E _)   = p
+    rightmost p@(Par _) = p
+
+    go :: (?pt :: OpTable)
+       => (PartialE -> PartialE)
+       -> PartialE -> PartialE -> PartialE
+    go f p@(WithInfo o _ r) = case r of
+            E _     -> mkHole o (f . f')
+            Par _   -> mkHole o (f . f')
+            B _ _ _ -> go (mkHole o (f . f')) r
+        where f' r' = p & pR .~ r'
+    go f _ = id
+
+mkHole :: (?pt :: OpTable)
+       => OpInfo
+       -> (PartialE -> PartialE)
+       -> PartialE
+       -> PartialE
+mkHole _     hole p@(Par _)                 = hole p
+mkHole _     hole p@(E _)                   = hole p
+mkHole (a,d) hole p@(WithInfo (a',d') _ _)
+    | d' < d  = above
+    | d' > d  = below
+    | d == d' = case (a,a') of
+                  -- left-associative operators of equal precedence are
+                  -- associated left
+                  (InfixL,InfixL) -> above
+                  -- right-associative operators are handled similarly
+                  (InfixR,InfixR) -> below
+                  -- non-associative operators of equal precedence, or equal
+                  -- precedence operators of different associativities are
+                  -- invalid
+                  (_,     _)      -> error "invalid expression"
+    where
+        above = p & pL %~ hole
+        below = hole p
+
+examplePrecTable :: OpTable
+examplePrecTable = H.fromList
+    [ ("+",    (InfixL,6))
+    , ("*",    (InfixL,7))
+    , ("^",    (InfixR,8))
+    , (".",    (InfixR,7))
+    , ("~",    (Infix, 9))
+    , ("=",    (Infix, 4))
+    , ("&&",   (Infix, 3))
+    , ("||",   (Infix, 2))
+    , ("$",    (InfixR,0))
+    , ("&",    (InfixL,0))
+    ]
+
+

src/Rlp/Parse/Types.hs

@@ -0,0 +1,160 @@
+{-# LANGUAGE TemplateHaskell #-}
+{-# LANGUAGE ImplicitParams, ViewPatterns, PatternSynonyms #-}
+{-# LANGUAGE LambdaCase #-}
+module Rlp.Parse.Types where
+--------------------------------------------------------------------------------
+import Core.Syntax                  (Name)
+import Control.Monad
+import Control.Monad.State.Class
+import Data.Text                    (Text)
+import Data.Maybe
+import Data.Fix
+import Data.Functor.Foldable
+import Data.Functor.Const
+import Data.Functor.Classes
+import Data.HashMap.Strict          qualified as H
+import Data.Word                    (Word8)
+import Lens.Micro.TH
+import Lens.Micro
+import Rlp.Syntax
+--------------------------------------------------------------------------------
+
+type LexerAction a = AlexInput -> Int -> P a
+
+data AlexInput = AlexInput
+    { _aiPrevChar   :: Char
+    , _aiSource     :: Text
+    , _aiBytes      :: [Word8]
+    , _aiPos        :: Position
+    }
+    deriving Show
+
+type Position =
+    ( Int -- line
+    , Int -- column
+    )
+
+data RlpToken
+    -- literals
+    = TokenLitInt Int
+    -- identifiers
+    | TokenVarName Name
+    | TokenConName Name
+    | TokenVarSym Name
+    | TokenConSym Name
+    -- reserved words
+    | TokenData
+    | TokenCase
+    | TokenOf
+    | TokenLet
+    | TokenIn
+    -- reserved ops
+    | TokenArrow
+    | TokenPipe
+    | TokenHasType
+    | TokenLambda
+    | TokenEquals
+    -- control symbols
+    | TokenSemicolon
+    | TokenLBrace
+    | TokenRBrace
+    | TokenLParen
+    | TokenRParen
+    -- 'virtual' control symbols, inserted by the lexer without any correlation
+    -- to a specific symbol
+    | TokenSemicolonV
+    | TokenLBraceV
+    | TokenRBraceV
+    | TokenEOF
+    deriving (Show)
+
+newtype P a = P { runP :: ParseState -> (ParseState, Maybe a) }
+    deriving (Functor)
+
+instance Applicative P where
+    pure a = P $ \st -> (st,Just a)
+    liftA2 = liftM2
+
+instance Monad P where
+    p >>= k = P $ \st ->
+        let (st',a) = runP p st
+        in case a of
+            Just x  -> runP (k x) st'
+            Nothing -> (st', Nothing)
+
+instance MonadState ParseState P where
+    state f = P $ \st ->
+        let (a,st') = f st
+        in (st', Just a)
+
+data ParseState = ParseState
+    { _psLayoutStack        :: [Layout]
+    , _psLexState           :: [Int]
+    , _psInput              :: AlexInput
+    , _psOpTable            :: OpTable
+    }
+    deriving Show
+
+data Layout = Explicit
+            | Implicit Int
+            deriving (Show, Eq)
+
+data Located a = Located (Position, Int) a
+    deriving (Show)
+
+type OpTable = H.HashMap Name OpInfo
+type OpInfo = (Assoc, Int)
+
+-- data WithLocation a = WithLocation [String] a
+
+data RlpParseError = RlpParErrOutOfBoundsPrecedence Int
+                   | RlpParErrDuplicateInfixD
+    deriving (Eq, Ord, Show)
+
+----------------------------------------------------------------------------------
+
+-- absolute psycho shit (partial ASTs)
+
+type PartialDecl' = Decl (Const PartialExpr') Name
+
+data Partial a = E (RlpExprF Name a)
+               | B Name (Partial a) (Partial a)
+               | Par (Partial a)
+               deriving (Show, Functor)
+
+pL :: Traversal' (Partial a) (Partial a)
+pL k (B o l r) = (\l' -> B o l' r) <$> k l
+pL _ x         = pure x
+
+pR :: Traversal' (Partial a) (Partial a)
+pR k (B o l r) = (\r' -> B o l r') <$> k r
+pR _ x         = pure x
+
+type PartialE = Partial RlpExpr'
+
+-- i love you haskell
+pattern WithInfo :: (?pt :: OpTable) => OpInfo -> PartialE -> PartialE -> PartialE
+pattern WithInfo p l r <- B (opInfoOrDef -> p) l r
+
+opInfoOrDef :: (?pt :: OpTable) => Name -> OpInfo
+opInfoOrDef c = fromMaybe (InfixL,9) $ H.lookup c ?pt
+
+-- required to satisfy constraint on Fix's show instance
+instance Show1 Partial where
+    liftShowsPrec :: forall a. (Int -> a -> ShowS)
+                  -> ([a] -> ShowS)
+                  -> Int -> Partial a -> ShowS
+
+    liftShowsPrec sp sl p m = case m of
+        (E e)       -> showsUnaryWith lshow "E" p e
+        (B f a b)   -> showsTernaryWith showsPrec lshow lshow "B" p f a b
+        (Par e)     -> showsUnaryWith lshow "Par" p e
+        where
+            lshow :: forall f. (Show1 f) => Int -> f a -> ShowS
+            lshow = liftShowsPrec sp sl
+
+type PartialExpr' = Fix Partial
+
+makeLenses ''AlexInput
+makeLenses ''ParseState
+

src/Rlp/Syntax.hs

@@ -0,0 +1,177 @@
+-- recursion-schemes
+{-# LANGUAGE DeriveFunctor, DeriveFoldable, DeriveTraversable #-}
+-- recursion-schemes
+{-# LANGUAGE TemplateHaskell, TypeFamilies #-}
+{-# LANGUAGE OverloadedStrings, PatternSynonyms #-}
+module Rlp.Syntax
+    ( RlpModule(..)
+    , RlpProgram(..)
+    , RlpProgram'
+    , rlpmodName
+    , rlpmodProgram
+    , RlpExpr(..)
+    , RlpExpr'
+    , RlpExprF(..)
+    , RlpExprF'
+    , Decl(..)
+    , Decl'
+    , Bind(..)
+    , Where
+    , Where'
+    , ConAlt(..)
+    , Type(..)
+    , pattern (:->)
+    , Assoc(..)
+    , VarId(..)
+    , ConId(..)
+    , Pat(..)
+    , Pat'
+    , Lit(..)
+    , Lit'
+    , Name
+
+    -- TODO: ugh move this somewhere else later
+    , showsTernaryWith
+
+    -- * Convenience re-exports
+    , Text
+    )
+    where
+----------------------------------------------------------------------------------
+import Data.Text                    (Text)
+import Data.Text                    qualified as T
+import Data.String                  (IsString(..))
+import Data.Functor.Foldable.TH     (makeBaseFunctor)
+import Data.Functor.Classes
+import Lens.Micro
+import Lens.Micro.TH
+import Core.Syntax                  hiding (Lit)
+import Core                         (HasRHS(..), HasLHS(..))
+----------------------------------------------------------------------------------
+
+data RlpModule b = RlpModule
+    { _rlpmodName       :: Text
+    , _rlpmodProgram    :: RlpProgram b
+    }
+
+newtype RlpProgram b = RlpProgram [Decl RlpExpr b]
+    deriving Show
+
+type RlpProgram' = RlpProgram Name
+
+-- | The @e@ parameter is used for partial results. When parsing an input, we
+-- first parse all top-level declarations in order to extract infix[lr]
+-- declarations. This process yields a @[Decl (Const Text) Name]@, where @Const
+-- Text@ stores the remaining unparsed function bodies. Once infixities are
+-- accounted for, we may complete the parsing task and get a proper @[Decl
+-- RlpExpr Name]@.
+
+data Decl e b = FunD    VarId [Pat b] (e b) (Maybe (Where b))
+              | TySigD  [VarId] Type
+              | DataD   ConId [Name] [ConAlt]
+              | InfixD  Assoc Int Name
+              deriving Show
+
+type Decl' e = Decl e Name
+
+data Assoc = InfixL
+           | InfixR
+           | Infix
+           deriving Show
+
+data ConAlt = ConAlt ConId [Type]
+            deriving Show
+
+data RlpExpr b = LetE  [Bind b] (RlpExpr b)
+               | VarE  VarId
+               | ConE  ConId
+               | LamE  [Pat b] (RlpExpr b)
+               | CaseE (RlpExpr b) [(Alt b, Where b)]
+               | IfE   (RlpExpr b) (RlpExpr b) (RlpExpr b)
+               | AppE  (RlpExpr b) (RlpExpr b)
+               | LitE  (Lit b)
+               deriving Show
+
+type RlpExpr' = RlpExpr Name
+
+type Where b = [Bind b]
+type Where' = [Bind Name]
+
+-- do we want guards?
+data Alt b = AltA (Pat b) (RlpExpr b)
+           deriving Show
+
+data Bind b = PatB (Pat b) (RlpExpr b)
+            | FunB VarId [Pat b] (RlpExpr b)
+            deriving Show
+
+data VarId = NameVar Text
+           | SymVar Text
+           deriving Show
+
+instance IsString VarId where
+    -- TODO: use symvar if it's an operator
+    fromString = NameVar . T.pack
+
+data ConId = NameCon Text
+           | SymCon Text
+           deriving Show
+
+data Pat b = VarP VarId
+           | LitP (Lit b)
+           | ConP ConId [Pat b]
+           deriving Show
+
+type Pat' = Pat Name
+
+data Lit b = IntL Int
+           | CharL Char
+           | ListL [RlpExpr b]
+           deriving Show
+
+type Lit' = Lit Name
+
+-- instance HasLHS Alt Alt Pat Pat where
+--     _lhs = lens
+--         (\ (AltA p _) -> p)
+--         (\ (AltA _ e) p' -> AltA p' e)
+
+-- instance HasRHS Alt Alt RlpExpr RlpExpr where
+--     _rhs = lens
+--         (\ (AltA _ e) -> e)
+--         (\ (AltA p _) e' -> AltA p e')
+
+makeBaseFunctor ''RlpExpr
+
+deriving instance (Show b, Show a) => Show (RlpExprF b a)
+
+type RlpExprF' = RlpExprF Name
+
+-- society if derivable Show1
+instance (Show b) => Show1 (RlpExprF b) where
+    liftShowsPrec sp _ p m = case m of
+        (LetEF bs e)            -> showsBinaryWith showsPrec sp "LetEF" p bs e
+        (VarEF n)               -> showsUnaryWith showsPrec "VarEF" p n
+        (ConEF n)               -> showsUnaryWith showsPrec "ConEF" p n
+        (LamEF bs e)            -> showsBinaryWith showsPrec sp "LamEF" p bs e
+        (CaseEF e as)           -> showsBinaryWith sp showsPrec "CaseEF" p e as
+        (IfEF a b c)            -> showsTernaryWith sp sp sp "IfEF" p a b c
+        (AppEF f x)             -> showsBinaryWith sp sp "AppEF" p f x
+        (LitEF l)               -> showsUnaryWith showsPrec "LitEF" p l
+
+showsTernaryWith :: (Int -> x -> ShowS)
+                 -> (Int -> y -> ShowS)
+                 -> (Int -> z -> ShowS)
+                 -> String -> Int
+                 -> x -> y -> z
+                 -> ShowS
+showsTernaryWith sa sb sc name p a b c = showParen (p > 10)
+    $ showString name
+    . showChar ' ' . sa 11 a 
+    . showChar ' ' . sb 11 b
+    . showChar ' ' . sc 11 c
+
+--------------------------------------------------------------------------------
+
+makeLenses ''RlpModule
+

tst/Arith.hs

@@ -6,6 +6,7 @@ module Arith
     ) where
 ----------------------------------------------------------------------------------
 import Data.Functor.Classes             (eq1)
+import Lens.Micro
 import Core.Syntax
 import GM
 import Test.QuickCheck
@@ -48,7 +49,7 @@ instance Arbitrary ArithExpr where
                     -- i don't feel like dealing with division at the moment
                     [ IntA <$> int
                     , NegateA <$> arbitrary
-                    -- , IdA <$> arbitrary
+                    , IdA <$> arbitrary
                     , b (:+)
                     , b (:-)
                     , b (:*)
@@ -70,13 +71,13 @@ instance Arbitrary ArithExpr where
 --         coreResult  = evalCore (toCore e)
 
 toCore :: ArithExpr -> Program'
-toCore expr = Program
+toCore expr = mempty & programScDefs .~
         [ ScDef "id" ["x"] $ Var "x"
         , ScDef "main" [] $ go expr
         ]
     where
         go :: ArithExpr -> Expr'
-        go (IntA n)    = LitE (IntL n)
+        go (IntA n)    = Lit (IntL n)
         go (NegateA e) = "negate#" :$ go e
         go (IdA e)     = "id" :$ go e
         go (a :+ b)    = f "+#" a b

tst/Core/HindleyMilnerSpec.hs

@@ -0,0 +1,46 @@
+{-# LANGUAGE QuasiQuotes, OverloadedStrings #-}
+module Core.HindleyMilnerSpec
+    ( spec
+    )
+    where
+----------------------------------------------------------------------------------
+import Core.Syntax
+import Core.TH                  (coreExpr)
+import Core.HindleyMilner
+import Control.Monad.Errorful
+import Data.Either              (isLeft)
+import Test.Hspec
+----------------------------------------------------------------------------------
+
+-- TODO: more tests. preferrably property-based. lol.
+spec :: Spec
+spec = do
+    it "should infer `id 3` :: Int" $
+        let g = [ ("id", "a" :-> "a") ]
+        in infer' g [coreExpr|id 3|] `shouldBe` Right TyInt
+
+    it "should not infer `id 3` when `id` is specialised to `a -> a`" $
+        let g = [ ("id", ("a" :-> "a") :-> "a" :-> "a") ]
+        in infer' g [coreExpr|id 3|] `shouldSatisfy` isLeft
+
+    -- TODO: property-based tests for let
+    it "should infer `let x = 3 in id x` :: Int" $
+        let g = [ ("id", "a" :-> "a") ]
+            e = [coreExpr|let {x = 3} in id x|]
+        in infer' g e `shouldBe` Right TyInt
+
+    it "should infer `let x = 3; y = 2 in (+#) x y` :: Int" $
+        let g = [ ("+#", TyInt :-> TyInt :-> TyInt) ]
+            e = [coreExpr|let {x=3;y=2} in (+#) x y|]
+        in infer' g e `shouldBe` Right TyInt
+    
+    it "should find `3 :: Bool` contradictory" $
+        let e = [coreExpr|3|]
+        in check' [] (TyCon "Bool") e `shouldSatisfy` isLeft
+
+infer' :: Context' -> Expr' -> Either TypeError Type
+infer' g e = fmap fst . runErrorful $ infer g e
+
+check' :: Context' -> Type -> Expr' -> Either TypeError ()
+check' g t e = fmap fst . runErrorful $ check g t e
+

tst/GMSpec.hs

@@ -36,3 +36,6 @@ spec = do
         it "k 3 ((/#) 1 0)" $ do
             resultOf Ex.constDivZero `shouldBe` Just (NNum 3)
 
+        it "id (case ... of { ... })" $ do
+            resultOf Ex.idCase `shouldBe` Just (NNum 5)
+