d11wtq/lx

Lx is a lex implementation in ECMAScript (including JavaScript)
---------------------------------------------------------------

Lx was developed as a dependency for another project which needs
to deal with source code and grammar rules.

Lx by itself implements just about the same features as Lex/Flex,
though the naming is consistently different and the routine uses
an object-oriented approach.

Lx is able to recognize the lower level structures in programming
source codes of most languages.  It uses basic string matching
and regular expressions to scan its input, performing user-
specified actions as it matches each structure, usually referred
to as a "terminal symbol" or "token".

Reading the Flex info page will almost certainly help you
understand the API which Lx uses.

A companion library is in development by the same author which
provides the parsing abilities of Yacc, recognising higher
level structures (nonterminal symbols) and interpreting the input
from Lx as it scans using a LALR(1) routine (scanning with a
lookahead symbol).

IMPORTANT: Lx is pre-1.0, it has NOT been optimized in true TDD
           fashion, but it will be optimized once the missing
           features* have been implemented.


Testing
-------

You may wish to run the unit tests provided before you get started.

The unit tests for Lx are written in the ECMAUnit framework and can
be run via an Ant script (if you have a Java Runtime installed,
along with the Ant software).

  ant -f ant-runtests.xml test-all

Alternatively they can be run inside your web browser by opening the
HTML file tests/unit/AllTests.html.


Naming Differences from Flex/Lex
--------------------------------

Though Lx follows the same API as Flex, it does not use exactly the
same names.  Rather than being prefixed "yy", all properties and
routines are members of a global object called Lx. Moreso, with the
exception of lex() and wrap(), all routines and are named with
CamelCase, including an uppercase first letter.  Token names and state
names are traditionally written all in UPPERCASE.

  Equivalents:
  -----------
  
  FILE *yyin              String Lx.In
  FILE *yyout             String Lx.Out
  char *yytext            String Lx.Text
  int yylineno            Integer Lx.LineNo
  int yyleng              Integer Lx.Leng
  int yylval              Object Lx.Lval
  int *START              Integer Lx.START
  int INITIAL             Integer Lx.INITIAL
  
  void ECHO               void Lx.Echo()
  void yymore()           void Lx.More()
  void yyless(n)          void Lx.Less(n)
  void unput(c)           void Lx.Unput(c)
  char input()            String Lx.Input()
  void yy_push_state(s)   void Lx.PushState(s)
  void yy_pop_state()     void Lx.PopState()
  int yy_top_state()      Integer Lx.TopState()
  void BEGIN cond         void Lx.Begin(s)
  void REJECT             void Lx.Reject()
  void terminate()        void Lx.Terminate()

  int yylex()             Integer Lx.lex()
  int yywrap()            Boolean Lx.wrap()
  
The global object Lx always points to the currently scanning analyzer.

* Missing Features (Regardless of the above list):
--------------------------------------------------

 - Lx.Reject()
 - Lx.Restart()
 - Lx.Terminate()
 - Inclusive states
 - Line number tracking (may never be added - easily achieved manually)