JavaCC [tm]: API Routines

For each non-terminal NT in the input grammar file, the following method is generated into the parser class:

• returntype NT(parameters) throws ParseError;

When this method is called, the input stream is parsed to match this non-terminal. On a successful parse, this method returns normally. On detection of a parse error, an error message is displayed and the method returns by throwing the exception ParseError.

Note that all non-terminals in a JavaCC input grammar have equal status; it is possible to parse to any non-terminal by calling the non-terminal's method.

• Token token;
  This variable holds the last token consumed by the parser and can be used in parser actions. This is exactly the same as the token returned by getToken(0).

Typically, the token manager interface is not to be used. Instead all access must be made through the parser interface. However, in certain situations - such as if you are not building a parser and building only the token manager - the token manager interface is useful. The token manager provides the following routine:

• Token getNextToken() throws ParseError;

• TheParser.TheParser(java.io.InputStream stream)
  This creates a new parser object, which in turn creates a new token manager object that reads its tokens from "stream". This constructor is available only when both the options USER_TOKEN_MANAGER and USER_CHAR_STREAM are false. If the option STATIC is true, this constructor (along with other constructors) can be called exactly once to create a single parser object.

• TheParser.TheParser(CharStream stream)
  Similar to the previous constructor, except that this one is available only when the option USER_TOKEN_MANAGER is false and USER_CHAR_STREAM is true.

• void TheParser.ReInit(java.io.InputStream stream)
  This reinitializes an existing parser object. In addition, it also reinitializes the existing token manager object that corresponds to this parser object. The result is a parser object with the exact same functionality as one that was created with the constructor above. The only difference is that new objects are not created. This method is available only when both the options USER_TOKEN_MANAGER and USER_CHAR_STREAM are false. If the option STATIC is true, this (along with the other ReInit methods) is the only way to restart a parse operation for there is only one parser and all one can do is reinitialize it.

• void TheParser.ReInit(CharStream stream)
  Similar to the previous method, except that this one is available only when the option USER_TOKEN_MANAGER is false and USER_CHAR_STREAM is true.

• TheParser(TheParserTokenManager tm) This creates a new parser object which uses an already created token manager object "tm" as its token manager. This constructor is only available if option USER_TOKEN_MANAGER is false. If the option STATIC is true, this constructor (along with other constructors) can be called exactly once to create a single parser object.

• TheParser(TokenManager tm)
  Similar to the previous constructor, except that this one is available only when the option USER_TOKEN_MANAGER is true.

• void TheParser.ReInit(TheParserTokenManager tm)
  This reinitializes an existing parser object with the token manager object "tm" as its new token manager. This method is only available if option USER_TOKEN_MANAGER is false. If the option STATIC is true, this (along with the other ReInit methods) is the only way to restart a parse operation for there is only one parser and all one can do is reinitialize it.

• void TheParser.ReInit(TokenManager tm)
  Similar to the previous method, except that this one is available only when the option USER_TOKEN_MANAGER is true.

• TheParserTokenManager.TheParserTokenManager(CharStream stream)
  Creates a new token manager object initialized to read input from "stream". When the option STATIC is true, this constructor may be called only once. This is available only when USER_TOKEN_MANAGER is false and USER_CHAR_STREAM is true. When USER_TOKEN_MANAGER is false and USER_CHAR_STREAM is false (the default situation), a constructor similar to the one above is available with the type CharStream replaced as follows:
  • When JAVA_UNICODE_ESCAPE is false and UNICODE_INPUT is false, CharStream is replaced by ASCII_CharStream.
  • When JAVA_UNICODE_ESCAPE is false and UNICODE_INPUT is true, CharStream is replaced by UCode_CharStream.
  • When JAVA_UNICODE_ESCAPE is true and UNICODE_INPUT is false, CharStream is replaced by ASCII_UCodeESC_CharStream.
  • When JAVA_UNICODE_ESCAPE is true and UNICODE_INPUT is true, CharStream is replaced by UCode_UCodeESC_CharStream.

• void TheParserTokenManager.ReInit(CharStream stream)
  Reinitializes the current token manager object to read input from "stream". When the option STATIC is true, this is the only way to restart a token manager operation. This is available only when USER_TOKEN_MANAGER is false and USER_CHAR_STREAM is true. When USER_TOKEN_MANAGER is false and USER_CHAR_STREAM is false (the default situation), a constructor similar to the one above is available with the type CharStream replaced as follows:
  • When JAVA_UNICODE_ESCAPE is false and UNICODE_INPUT is false, CharStream is replaced by ASCII_CharStream.
  • When JAVA_UNICODE_ESCAPE is false and UNICODE_INPUT is true, CharStream is replaced by UCode_CharStream.
  • When JAVA_UNICODE_ESCAPE is true and UNICODE_INPUT is false, CharStream is replaced by ASCII_UCodeESC_CharStream.
  • When JAVA_UNICODE_ESCAPE is true and UNICODE_INPUT is true, CharStream is replaced by UCode_UCodeESC_CharStream.

The Token class is the type of token objects that are created by the token manager after a successful scanning of the token stream. These token objects are then passed to the parser and are accessible to the actions in a JavaCC grammar usually by grabbing the return value of a token. The methods getToken and getNextToken described below also give access to objects of this type.

Each Token object has the following fields:

• int kind;
  This is the index for this kind of token in the internal representation scheme of JavaCC. When tokens in the JavaCC input file are given labels, these labels are used to generate "int" constants that can be used in actions. The value 0 is always used to represent the predefined token <EOF>. A constant "EOF" is generated for convenience in the ...Constants file.

• int beginLine, beginColumn, endLine, endColumn;
  These indicate the beginning and ending positions of the token as it appeared in the input stream.

• String image;
  This represents the image of the token as it appeared in the input stream.

• Token next;
  A reference to the next regular (non-special) token from the input stream. If this is the last token from the input stream, or if the token manager has not read tokens beyond this one, this field is set to null.

  The description in the above paragraph holds only if this token is also a regular token. Otherwise, see below for a description of the contents of this field.

  Note: There are two kinds of tokens - regular and special. Regular tokens are the normal tokens that are fed to the parser. Special tokens are other useful tokens (like comments) that are not discarded (like white space). For more information on the different kinds of tokens please see the minitutorial on the token manager.

• Token specialToken;
  This field is used to access special tokens that occur prior to this token, but after the immediately preceding regular (non-special) token. If there are no such special tokens, this field is set to null. When there are more than one such special token, this field refers to the last of these special tokens, which in turn refers to the next previous special token through its specialToken field, and so on until the first special token (whose specialToken field is null). The next fields of special tokens refer to other special tokens that immediately follow it (without an intervening regular token). If there is no such token, this field is null.

• static final Token newToken(int ofKind);
  Returns a new token object as its default behavior. If you wish to perform special actions when a token is constructed or create subclasses of class Token and instantiate them instead, you can redefine this method appropriately. The only constraint is that this method returns a new object of type Token (or a subclass of Token).

There are two methods available for this purpose:

• Token TheParser.getNextToken() throws ParseError
  This method returns the next available token in the input stream and moves the token pointer one step in the input stream (i.e., this changes the state of the input stream). If there are no more tokens available in the input stream, the exception ParseError is thrown. Care must be taken when calling this method since it can interfere with the parser's knowledge of the state of the input stream, current token, etc.

• Token TheParser.getToken(int index) throws ParseError
  This method returns the index-th token from the current token ahead in the token stream. If index is 0, it returns the current token (the last token returned by getNextToken or consumed by the parser); if index is 1, it returns the next token (the next token that will be returned by getNextToken of consumed by the parser) and so on. The index parameter cannot be negative. This method does not change the input stream pointer (i.e., it does not change the state of the input stream). If an attempt is made to access a token beyond the last available token, the exception ParseError is thrown. If this method is called from a semantic lookahead specification, which in turn is called during a lookahead determination process, the current token is temporarily adjusted to be the token currently being inspected by the lookahead process. For more details, please see the minitutorial on using lookahead.

When you generate parsers with the options DEBUG_PARSER or DEBUG_LOOKAHEAD, these parsers produce a trace of their activity which is printed to the user console. You can insert calls to the following methods to control this tracing activity:

• void TheParser.enable_tracing()

• void TheParser.disable_tracing()

For convenience, these methods are available even when you build parsers without the debug options. In this case, these methods are no-ops. Hence you can permanently leave these methods in your code and they automatically kick in when you use the debug options.

To help the user in customizing error messages generated by the parser and lexer, the user is offered the facilities described in this section. In the case of the parser, these facilities are only available if the option ERROR_REPORTING is true, while in the case of the lexer, these facilities are always available.

The parser contains the following method definition:

• protected void token_error() { ... }

To customize error reporting by the parser, the parser class must be subclassed and this method redefined in the subclass. To help with creating your error reporting scheme, the following variables are available:

• protected int error_line, error_column;
  The line and column where the error was detected.

• protected String error_string;
  The image of the offending token or set of tokens. When a lookahead of more than 1 is used, more than one token may be present here.

• protected String[] expected_tokens;
  An array of images of legitimate token sequences. Here again, each legitimate token sequence may be more than just one token when a lookahead of more than 1 is used.

The lexer contains the following method definition:

• protected void LexicalError() { ... }

To customize error reporting by the lexer, the lexer class must be subclassed and this method redefined in the subclass. To help with creating your error reporting scheme, the following variables are available:

• protected int error_line, error_column;
  The line and column where the error was detected.

• protected String error_after;
  The partial string that has been read since the last successful token match was performed.

• protected char curChar;
  The offending character.

JavaCC [tm]: JJTree

JJTree has two APIs: it adds some parser methods; and it requires all node objects to implement the Node interface.

JJTree parser methods

JJTree maintains some state in the parser object itself. It encapsulates all this state with an object that can be referred to via the jjtree field.

The parser state implements an open stack where nodes are held until they can be added to their parent node. The jjtree state object provides methods for you to manipulate the contents of the stack in your actions if the basic JJTree mechanisms are not sufficient.

void reset()

Call this to reinitialize the node stack. All nodes currently on the stack are thrown away. Don't call this from within a node scope, or terrible things will surely happen.

Node rootNode();

Returns the root node of the AST. Since JJTree operates bottom-up, the root node is only defined after the parse has finished.

boolean nodeCreated();

Determines whether the current node was actually closed and pushed. Call this in the final action within a conditional node scope.

int arity();

Returns the number of nodes currently pushed on the node stack in the current node scope.

void pushNode(Node n);

Pushes a node on to the stack.

Node popNode();

Returns the node on the top of the stack, and removes it from the stack.

Node peekNode();

Returns the node currently on the top of the stack.

The Node interface

All AST nodes must implement this interface. It provides basic machinery for constructing the parent and child relationships between nodes.

public void jjtOpen();

This method is called after the node has been made the current node. It indicates that child nodes can now be added to it.

public void jjtClose();

This method is called after all the child nodes have been added.

public void jjtSetParent(Node n);
public Node jjtGetParent();

This pair of methods is used to inform the node of its parent.

public void jjtAddChild(Node n, int i);

This method tells the node to add its argument to the node's list of children.

public Node jjtGetChild(int i);

This method returns a child node. The children are numbered from zero, left to right.

int jjtGetNumChildren();

Return the number of children the node has.