[project @ 1996-01-08 20:28:12 by partain]

[ghc-hetmet.git] / ghc / lib / hbc / SimpleLex.hs

diff --git a/ghc/lib/hbc/SimpleLex.hs b/ghc/lib/hbc/SimpleLex.hs
new file mode 100644 (file)
index 0000000..b039bfe
--- /dev/null
+++ b/ghc/lib/hbc/SimpleLex.hs
@@ -0,0 +1,26 @@
+-- A very simple, but useful, lexical analyser.
+module SimpleLex(simpleLex) where
+
+oper = "!#$%&*+./<=>?@\\^|:~-"
+-- self-delim ()[]{},;`'"_
+isalunum c = isAlphanum c || c == '_'
+
+simpleLex :: String -> [String]
+simpleLex "" = []
+simpleLex (' ' :cs) = simpleLex cs                     -- ignore white space
+simpleLex ('\t':cs) = simpleLex cs
+simpleLex ('\n':cs) = simpleLex cs
+simpleLex ('-':cs@(c:_)) | isDigit c =                         -- negative numbers
+       let (t:ts) = simpleLex cs 
+       in  ('-':t) : ts
+simpleLex (c:cs) | isDigit c =                                 -- numbers (with optional .)
+       let (nn, cs') = span isDigit cs 
+       in  case cs' of
+           '.':cs'' -> let (d,r) = span isDigit cs'' 
+                       in  (c:nn++'.':d) : simpleLex r
+           _ -> (c:nn) : simpleLex cs'
+simpleLex (c:cs) | isAlpha c =                                 -- identifiers
+       let (nn, cs') = span isalunum cs in (c:nn) : simpleLex cs'
+simpleLex (c:cs) | c `elem` oper =                     -- operator
+       let (nn, cs') = span (`elem` oper) cs in (c:nn) : simpleLex cs'
+simpleLex (c:cs) = [c] : simpleLex cs                  -- self delimiting chars