diff --git a/src/Main.hs b/src/Main.hs
index 9cd992d..ba73720 100644
--- a/src/Main.hs
+++ b/src/Main.hs
@@ -1,5 +1,14 @@
 module Main where
 
+import qualified Data.ByteString.Lazy as Bsl
+import Data.Foldable
+
+import Presyntax.Parser
+import Presyntax.Tokens
+import Presyntax.Lexer
+
 main :: IO ()
 main = do
-  putStrLn "hello world"
+  t <- Bsl.readFile "test.tt"
+  let Right tks = runAlex t parseProg
+  traverse_ print tks
\ No newline at end of file
diff --git a/src/Presyntax/Lexer.x b/src/Presyntax/Lexer.x
index e7c54cd..42ded0c 100644
--- a/src/Presyntax/Lexer.x
+++ b/src/Presyntax/Lexer.x
@@ -7,31 +7,44 @@ import qualified Data.Text.Encoding as T
 import Presyntax.Tokens
 }
 
-%wrapper "monad-bytestring"
+%wrapper "monadUserState-bytestring"
 
 $alpha = [a-zA-Z]
 $digit = [0-9]
+$white_nol = $white # \n
 
 tokens :-
-  $white+                      ;
-  $alpha [$alpha $digit \_ \']* { yield TokVar }
+  $white_nol+                   ;
   
-  \=                            { always TokEqual }
-  \:                            { always TokColon }
-  \,                            { always TokComma }
-  \*                            { always TokStar }
+-- zero state: normal lexing
+<0> $alpha [$alpha $digit \_ \']* { yield TokVar }
 
-  ".1"                          { always TokPi1 }
-  ".2"                          { always TokPi2 }
-  
-  \\                            { always TokLambda }
-  "->"                          { always TokArrow }
+<0> \=                            { always TokEqual }
+<0> \:                            { always TokColon }
+<0> \,                            { always TokComma }
+<0> \*                            { always TokStar }
+
+<0> ".1"                          { always TokPi1 }
+<0> ".2"                          { always TokPi2 }
+
+<0> \\                            { always TokLambda }
+<0> "->"                          { always TokArrow }
+
+<0> \(                            { always TokOParen }
+<0> \{                            { always TokOBrace }
 
-  \(                            { always TokOParen }
-  \{                            { always TokOBrace }
+<0> \)                            { always TokCParen }
+<0> \}                            { always TokCBrace }
+<0> \;                            { always TokSemi }
+
+<0>       \n                      { begin newline }
+
+-- newline: emit a semicolon when de-denting
+<newline> {
+  \n ;
+  () { offsideRule }
+}
 
-  \)                            { always TokCParen }
-  \}                            { always TokCBrace }
 
 {
 alexEOF :: Alex Token
@@ -41,4 +54,43 @@ alexEOF = do
 
 yield k (AlexPn _ l c, _, s, _) i = pure (Token (k $! (T.decodeUtf8 (Lbs.toStrict (Lbs.take i s)))) l c)
 always k x i = yield (const k) x i
+
+data AlexUserState = AlexUserState { layoutColumns :: [Int], startCodes :: [Int] }
+alexInitUserState = AlexUserState [1] []
+
+just :: Alex a -> AlexAction Token
+just k _ _ = k *> alexMonadScan
+
+getUserState :: Alex AlexUserState
+getUserState = Alex $ \s -> Right (s, alex_ust s)
+
+mapUserState :: (AlexUserState -> AlexUserState) -> Alex ()
+mapUserState k = Alex $ \s -> Right (s { alex_ust = k $! alex_ust s }, ())
+
+pushStartCode :: Int -> Alex ()
+pushStartCode c = do
+  sc <- alexGetStartCode
+  mapUserState $ \s -> s { startCodes = sc:startCodes s }
+  alexSetStartCode c
+  
+popStartCode :: Alex ()
+popStartCode = do
+  sc <- startCodes <$> getUserState
+  case sc of
+    [] -> alexSetStartCode 0
+    (x:xs) -> do
+      mapUserState $ \s -> s { startCodes = xs }
+      alexSetStartCode x
+
+offsideRule :: AlexInput -> Int64 -> Alex Token
+offsideRule i@(AlexPn p line col, _, s, _) l = do
+  ~(col':_) <- layoutColumns <$> getUserState
+  case col `compare` col' of
+    EQ -> do
+      popStartCode
+      pure (Token TokSemi line col)
+    GT -> do
+      popStartCode
+      alexMonadScan
+    LT -> alexError "wrong ass indentation"
 }
\ No newline at end of file
diff --git a/src/Presyntax/Parser.y b/src/Presyntax/Parser.y
index ab90586..90c1e1e 100644
--- a/src/Presyntax/Parser.y
+++ b/src/Presyntax/Parser.y
@@ -10,7 +10,9 @@ import Presyntax.Lexer
 
 }
 
-%name parseExp Exp
+%name parseExp  Exp
+%name parseStmt Statement
+%name parseProg Program
 
 %tokentype { Token }
 
@@ -32,6 +34,7 @@ import Presyntax.Lexer
   
   '->'    { Token TokArrow _ _ }
   ':'     { Token TokColon _ _ }
+  ';'     { Token TokSemi  _ _ }
   '='     { Token TokEqual _ _ }
   ','     { Token TokComma _ _ }
   '*'     { Token TokStar _ _ }
@@ -43,12 +46,12 @@ import Presyntax.Lexer
 
 Exp :: { Expr }
 Exp
-  : ExpProj Exp                       { App Ex $1 $2 }
-  | ExpProj '{' Exp '}'               { App Im $1 $3 }
+  : Exp ExpProj                       { App Ex $1 $2 }
+  | Exp '{' Exp '}'                   { App Im $1 $3 }
 
   | '\\' LambdaList '->' Exp          { makeLams $2 $4 }
-  | '(' VarList ':' Exp ')' '->' Exp  { makePis Ex $2 $4 $7 }
-  | '{' VarList ':' Exp '}' '->' Exp  { makePis Im $2 $4 $7 }
+  | '(' VarList ':' Exp ')' ProdTail  { makePis Ex $2 $4 $6 }
+  | '{' VarList ':' Exp '}' ProdTail  { makePis Im $2 $4 $6 }
   | ExpProj '->' Exp                  { Pi Ex (T.singleton '_') $1 $3 }
 
   | '(' VarList ':' Exp ')' '*'  Exp  { makeSigmas $2 $4 $7 }
@@ -56,6 +59,11 @@ Exp
 
   | ExpProj                           { $1 }
 
+ProdTail :: { Expr }
+  : '(' VarList ':' Exp ')' ProdTail  { makePis Ex $2 $4 $6 }
+  | '{' VarList ':' Exp '}' ProdTail  { makePis Im $2 $4 $6 }
+  | '->' Exp                          { $2 }
+
 LambdaList :: { [(Plicity, Text)] }
   : var            { [(Ex, $1)] }
   | var LambdaList { (Ex, $1):$2 }
@@ -80,6 +88,14 @@ Tuple :: { Expr }
   : Exp           { $1 }
   | Exp ',' Tuple { Pair $1 $3 }
 
+Statement :: { Statement }
+  : var ':' Exp { Decl $1 $3 }
+  | var '=' Exp { Defn $1 $3 }
+
+Program :: { [Statement] }
+  : Statement             { [$1] }
+  | Statement ';' Program { $1:$3 }
+
 {
 lexer cont = alexMonadScan >>= cont
 
diff --git a/src/Presyntax/Tokens.hs b/src/Presyntax/Tokens.hs
index 46c4f23..93c5fd9 100644
--- a/src/Presyntax/Tokens.hs
+++ b/src/Presyntax/Tokens.hs
@@ -21,6 +21,8 @@ data TokenClass
 
   | TokPi1
   | TokPi2
+
+  | TokSemi
   deriving (Eq, Show, Ord)
 
 data Token