From 47c776413ed4e11839ad6838575d0077ddd496a3 Mon Sep 17 00:00:00 2001
From: Josh Rahm <joshuarahm@gmail.com>
Date: Sat, 17 Dec 2022 01:28:06 -0700
Subject: fiddle: have a basic tokenizer working.

---
 src/Language/Fiddle/Tokenizer.hs | 128 +++++++++++++++++++++++++++++----------
 1 file changed, 96 insertions(+), 32 deletions(-)

(limited to 'src/Language/Fiddle/Tokenizer.hs')

diff --git a/src/Language/Fiddle/Tokenizer.hs b/src/Language/Fiddle/Tokenizer.hs
index 9931523..d3239fd 100644
--- a/src/Language/Fiddle/Tokenizer.hs
+++ b/src/Language/Fiddle/Tokenizer.hs
@@ -1,37 +1,101 @@
+{-# LANGUAGE DeriveFunctor #-}
 module Language.Fiddle.Tokenizer where
 
+import Data.Char (isDigit)
+import Data.Text (Text)
 import Language.Fiddle.Types
+import Text.Parsec
+import qualified Text.Parsec
 
 data T
-  = KW_assert_pos
-  | KW_at
-  | KW_bittype
-  | KW_enum
-  | KW_location
-  | KW_object
-  | KW_objtype
-  | KW_option
-  | KW_package
-  | KW_reg
-  | KW_ro
-  | KW_wo
-  | Tok_colon
-  | Tok_comma
-  | Tok_comment
-  | Tok_docComment
-  | Tok_eq
-  | Tok_ident String
-  | Tok_lbrace
-  | Tok_lbracket
-  | Tok_litnum String
-  | Tok_lparen
-  | Tok_package
-  | Tok_rbrace
-  | Tok_rbracket
-  | Tok_rparen
-  | Tok_semi
-
-data Token = Token T SourceSpan
-
-tokenize :: String -> Text -> [Token]
-tokenize srcName txt = undefined
+  = KWAssertPos
+  | Ident !String
+  | KWAt
+  | KWBittype
+  | KWEnum
+  | CommentTok !String
+  | DocCommentTok !String
+  | KWLocation
+  | KWObject
+  | KWObjtype
+  | KWOption
+  | KWPackage
+  | KWReg
+  | KWRo
+  | KWWo
+  | LitNum !String
+  | TokColon
+  | TokComma
+  | TokEq
+  | TokLBrace
+  | TokLBracket
+  | TokLParen
+  | TokRBrace
+  | TokRBracket
+  | TokRParen
+  | TokSemi
+  deriving (Eq, Ord, Show, Read)
+
+data Token a = Token !T a
+  deriving (Eq, Ord, Show, Functor)
+
+parseToken :: (Monad m) => ParsecT Text u m (Token SourceSpan)
+parseToken = spaces *> tok parseToken' <* spaces
+  where
+    tok tp = do
+      p1 <- getPosition
+      t <- tp
+
+      Token t . SourceSpan p1 <$> getPosition
+
+    parseAlNumTok str =
+      case str of
+        "at" -> KWAt
+        "bittype" -> KWBittype
+        "enum" -> KWEnum
+        "location" -> KWLocation
+        "object" -> KWObject
+        "objtype" -> KWObjtype
+        "option" -> KWOption
+        "package" -> KWPackage
+        "reg" -> KWReg
+        "ro" -> KWRo
+        "wo" -> KWWo
+        (h : _) | isDigit h -> LitNum str
+        ident -> Ident ident
+
+    parseComment =
+      try
+        ( do
+            string "//"
+            CommentTok <$> manyTill anyChar (char '\n')
+        )
+        <|> try
+          ( do
+              string "/**"
+              DocCommentTok <$> manyTill anyChar (try $ string "*/")
+          )
+
+    parseSymbol =
+      choice
+        [ char ':' $> TokColon,
+          char ',' $> TokComma,
+          char '=' $> TokEq,
+          char '{' $> TokLBrace,
+          char '[' $> TokLBracket,
+          char '(' $> TokLParen,
+          char '}' $> TokRBrace,
+          char ']' $> TokRBracket,
+          char ')' $> TokRParen,
+          char ';' $> TokSemi
+        ]
+      where
+        a $> b = a >> return b
+
+    parseToken' =
+      fmap parseAlNumTok (many1 (alphaNum <|> char '_'))
+        <|> parseComment
+        <|> parseSymbol
+
+tokenize :: String -> Text -> Either ParseError [Token SourceSpan]
+tokenize = Text.Parsec.runParser (many parseToken <* eof) ()
-- 
cgit