Import section: between %Import and Import%
    This section is copied into the output file before the definition of the lexer class.
Class option: between %ClassOptions and ClassOptions%
    The ClassOptions section allows you to customize the generated lexer class - is it final, what should the name be..
LexerClass: between %LexerClass and LexerClass%
    This should contain code that must be part of the LexerClass - extra members and methods
Java: between %Java and Java%
    Auxilliary classes - the content is copied in after the Lexer class
LexerDef: between %LexerDef and LexerDef%
    The definition of the lexer - regular expressions to actions

Everything outside the sections is ignored. The only required section is the LexerDef section.
You can put comments outside the sections and in the code  ( as java comments - //... and /*...*/ )

In the following I will go through each of these sections.

%Import
    import java.util.*;
Import%

You should at all times avoid using these names as identifiers:
    variables:
         ULEX__index
         ULEX__currentIndex
         ULEX__currentState
         ULEX__theReader
         ULEX__tokenBuf
         ULEX__c
         ULEX__accepted;
         ULEX__start
    methods:
        next_token
        ULEX__next
        getText
        getStartIndex
        getEndIndex
 
 

This section is the most important, and also the only one required. It defines a number of mappings from regular expressions
to actions. Since next_token() is required to return java_cup.runtime.Symbol objects, a mapping is normally just a regular expression to a return new Symbol(..); The syntax is as follows:

    mapping : REGEXP %{ CODE %}

The CODE contains code to be executed when accepting the REGEXP.
If the CODE section is empty ( i.e. no text what so ever ), the string matching REGEXP is ignored. For a description of the supported format of regular expressions see Regular expression format
 
 

I was very much inspired by mosml-lex in the definition of the regular expression format. However, the precedence rules are VERY obscure, so you should use the ( regexp ) format when combining regular expression, e.g. :
('a''a'*) is the same as ('a''a')*. I will probably look into this problem at some time, but for now, you should just use the parenthesis : (('a')('a'*))

Here is the complete syntax for regular expressions:
 

* character constants:
    The following escape sequences are supported: \b, \t, \n, \f, \r, \", \', \\ and \number
    where number is the ascii value (0-128)
    For characters (, ), [ and ] you MUST use the ascii value if the regular expression contains () or [] groupings:
        (    :    40
        )    :    41
        [    :    91
        ]    :    93
 
 

I have provided some examples of lexer file definitions: