The Lexiparse class may be used to instantiate a combined lexical analyzer / parser generator for making programming language interpreters and/or compilters.

First you must: * Code your language's grammar definition (notation described later herein) * Include and instantiate an instance of the Lexiparse class, passing your grammar definition and any non-default options into the constructor.

The sample file "burp.js" illustrates with implementation of the "burp" language.

Grammar Definition

The grammar definition is a JavaScript object, such as the following illustrative example:

grammar = {
	'stmt':[
		['input',':var', getInput],
		['output',':expr', doOutput],
		['var','=',':expr', doAssignment]
	],
	'expr':[
		':numlit',
		':var',
		[':expr','+',':expr', doAddition],
		[':expr','+',':expr', doSubtraction],
		['(',':expr',')']
	],
	'var':[/^[A-Za-z][A-Za-z0-9]+/,getValue],
	'numlit':[/^[+-]?\d+(\.\d+)?/],
}

The 'stmt', 'expr', 'var', and 'numlit' elements are named grammar

segments. Each holds an array of options, any of which will evaluate to it. These options may be strings, regular expressions, or sub-arrays (representing sequences). Each works like this:

String
	A string not starting with a colon is interpreted as a literal to match

in the program source code. However, if the string begins with a colon (and not a double-colon than the rest of the string is interpreted as the name of a grammar segment. For example, ':expr' means as defined under segment 'expr'. However, '::expr' would mean the text literal '::expr'.

Regular Expression
	Regular expressions must always begin with a karat character ('^') but

are otherwise normal regular expressions to match against program source code. It will extract exactly what the JavaScript Rexexp.exec() produces, so you may make use of parenthesis, etc., as desired.

Sequence
	If a segment option is an array itself, the elements of that array are

interpreted as required items in sequence. Furthermore, the last element may optionally be a function to preprocess the option's returned values. For example, the 'stmt' segment has 3 sequences under it. The 'expr' segment has 5 options, the last 3 of which are sequences.

Constructor Options

caseful
	This may be true (default) for case sensitivity or false for case 
	insensitivity.

ignore
	This should hold an array of characters to ignore.  Usually, this will
	be the whitespace characters of your language.
	NOTE: In the future, I might want to add a way to specify ignored 
	characters for each named segment, individually.

sequenceBlind (TODO: not yet implemeneted)
	If true, sequences will match if all elements exist in place, in order,
	but will ignore any extraneous elements before, after, or between.

literalBlind (TODO: not yet implemented)
	If true, tolerate keyword mispellings where still recognizable.