Add logic to open file, lex single character symbols, tests

3 files changed, 68 insertions, 5 deletions

diff --git a/src/lex.lisp b/src/lex.lisp
new file mode 100644
index 0000000..ad386ba
--- /dev/null
+++ b/src/lex.lisp
@@ -0,0 +1,52 @@
+(in-package #:lex)
+
+(defun file->tokens (file)
+  "Opens FILE and parses returns a list of tokens, or
+NIL if the file could not be opened."
+
+  (defun read-tokens (tokens-so-far)
+    "Collects tokens in FILE into TOKENS-SO-FAR."
+    (let ((token (read-token)))
+      (if token
+          (read-tokens (cons token tokens-so-far))
+          (reverse tokens-so-far))))
+
+  (and (probe-file file)
+       (with-open-file (*standard-input* file :direction :input)
+         (read-tokens '()))))
+
+(defun read-token ()
+  "Reads *STANDARD-INPUT* and returns a token, or nil if the end
+of file has been reached.
+Whitespace, commas, colons, and parentheses are token delimiters.
+Comments start with a semi-colon ';' and all tokens after are ignored."
+  (let ((chr (read-char *standard-input* nil)))
+    (cond
+      ((null chr) chr)
+      ((whitespace-char-p chr)
+       (read-token))
+
+      ((char= chr #\;)
+       (progn (read-line *standard-input* nil)
+              (read-token)))
+
+      ((char= chr #\() 'left-paren)
+      ((char= chr #\)) 'right-paren)
+
+      ((digit-char-p chr)
+       (read-immediate chr))
+
+      ((alpha-char-p chr)
+       (read-identifier chr))
+
+      (t (error (format nil "~a is not a valid lexical symbol.~%" chr))))))
+
+(defun read-immediate (chr)
+  'immediate)
+
+(defun read-identifier (chr)
+  'id)
+
+(defun whitespace-char-p (x)
+  (or (char= #\space x)
+      (not (graphic-char-p x))))
diff --git a/src/main.lisp b/src/main.lisp
index c85e392..1f0afdd 100644
--- a/src/main.lisp
+++ b/src/main.lisp
@@ -35,14 +35,19 @@ _/_/                                  _/_/  "
 (defun driver (cmd)
   "Reads in a file and directs lexing, parsing, and binary emission."
   (print-splash)
-  (let ((args (clingon:command-arguments cmd))
-        (parse? (not (clingon:getopt cmd :lex)))
-        (emit? (not (clingon:getopt cmd :parse))))
+  (let* ((args (clingon:command-arguments cmd))
+         (file (car args))
+         (parse? (not (clingon:getopt cmd :lex)))
+         (emit? (not (clingon:getopt cmd :parse))))
     (cond
       ;; complain about num arguments
       ((/= (length args) 1) (error "Wrong number of arguments."))
-      ((not (util:asm-extension? (car args))) (error "The file is not an asm source code file."))))
-  (error-cli "Nitimur in Vetitum"))
+      ((not (util:asm-extension? file))
+       (error "The file is not an asm source code file, or it could not be opened."))
+      (t (let ((tokens (lex:file->tokens file)))
+           (format t "~a" tokens)
+           (format t "Nitimur in Vetitum~%"))))))
+
 
 (defun cli/command ()
   "Returns a clingon command."
diff --git a/src/package.lisp b/src/package.lisp
index 9d21293..44399cb 100644
--- a/src/package.lisp
+++ b/src/package.lisp
@@ -5,3 +5,9 @@
 (defpackage #:util
   (:use #:cl)
   (:export #:asm-extension?))
+
+(defpackage #:lex
+  (:use #:cl)
+  (:export #:file->tokens
+           ;; exported for testing only
+           #:read-token))