README.1

                  --------------------------------------
                  |   CS143 Programming Assignment 1   |
                  | Lexical Analysis - Design Document |
                  --------------------------------------

Author: Frank Chen <frankchn@stanford.edu>

General Overview
================

I've used simple regular expressions for most of the choices I have made in 
writing the lexical analyzer in flex.

I used DoBeforeEachAction and helper functions to count the number of new lines
and characters that is matched and updates a pair of global variables to count
the number of lines parsed and characters parsed in the certain line. 

Identifiers are placed at the top, before whitespace and other catch all regexps
to assist in correctly tie-breaking the 

I also simplified structure by packing everything into the regular expression
instead of carrying state around. For instance, I packed multi-line C-style
comments such as /* */ into one single regular expression. By maximal munch,
the C-style comment will be matched to its maximal extent. As we do not allow
any nested comments, this can be done fairly easily with rudimentary regular
expressions.


Specific Design Details
=======================

I matched all characters in a string except a \n and " using a regular
expression. 

I also handled unclosed multi-line comments and unclosed string literals 
specially For multi-line comments, I matched all characters except the literal
"*/" for unclosed multi-line comments and all characters except " and \n for
unclosed strings and printed out error messages as noted. 

I also handled floats with and without exponents in a separate manner, to
prevent things like 12.34E from being parsed as a float - which would be an 
error. If tehre is an E or e after a float, then the exponents and optionally
a sign must follow.