Support for the Ruby 2.4 series has ended. See here for reference.

Class Index [+]


Racc Grammar File Reference

Global Structure

Class Block and User Code Block

There are two blocks on the toplevel. One is the 'class' block, the other is the 'user code' block. The 'user code' block MUST be placed after the 'class' block.


You can insert comments about all places. Two styles of comments can be used, Ruby style '#.….' and C style '/*.…..*/'.

Class Block

The class block is formed like this:

  [precedance table]
  [token declarations]
  [expected number of S/R conflicts]
  [semantic value conversion]
  [start rule]

CLASS_NAME is a name of the parser class. This is the name of the generating parser class.

If CLASS_NAME includes '::', Racc outputs the module clause. For example, writing “class M::C” causes the code below to be created:

module M
  class C

Grammar Block

The grammar block describes grammar which is able to be understood by the parser. Syntax is:

(token): (token) (token) (token).... (action)

(token): (token) (token) (token).... (action)
       | (token) (token) (token).... (action)
       | (token) (token) (token).... (action)

(action) is an action which is executed when its (token)s are found. (action) is a ruby code block, which is surrounded by braces:

{ print val[0]
  puts val[1] }

Note that you cannot use '%' string, here document, '%r' regexp in action.

Actions can be omitted. When it is omitted, '' (empty string) is used.

A return value of action is a value of the left side value ($$). It is the value of the result, or the returned value by `return` statement.

Here is an example of the whole grammar block.

  goal: definition rules source { result = val }

  definition: /* none */   { result = [] }
    | definition startdesig  { result[0] = val[1] }
    | definition
             precrule   # this line continues from upper line
        result[1] = val[1]

  startdesig: START TOKEN

You can use the following special local variables in action:

The value of the left-hand side (lhs). A default value is val.

An array of value of the right-hand side (rhs).

A stack of values. DO NOT MODIFY this stack unless you know what you are doing.

Operator Precedence

This function is equal to '%prec' in yacc. To designate this block:

  nonassoc '++'
  left     '*' '/'
  left     '+' '-'
  right    '='

`right` is yacc's %right, `left` is yacc's %left.

`=` + (symbol) means yacc's %prec:

  nonassoc UMINUS
  left '*' '/'
  left '+' '-'

  exp: exp '*' exp
     | exp '-' exp
     | '-' exp       =UMINUS   # equals to "%prec UMINUS"


Racc has bison's “expect” directive.

# Example

class MyParser
  expect 3

This directive declares “expected” number of shift/reduce conflicts. If “expected” number is equal to real number of conflicts, Racc does not print conflict warning message.

Declaring Tokens

By declaring tokens, you can avoid many meaningless bugs. If declared token does not exist or existing token does not decleared, Racc output warnings. Declaration syntax is:



You can write options for Racc command in your Racc file.

options OPTION OPTION ...

Options are:

omits empty action call or not.

uses local variable “result” or not.

You can use 'no_' prefix to invert their meanings.

Converting Token Symbol

Token symbols are, as default,

* naked token string in Racc file (TOK, XFILE, this_is_token, ...)
  --> symbol (:TOK, :XFILE, :this_is_token, ...)
* quoted string (':', '.', '(', ...)
  --> same string (':', '.', '(', ...)

You can change this default by “convert” block. Here is an example:

  PLUS 'PlusClass'      # We use PlusClass for symbol of `PLUS'
  MIN  'MinusClass'     # We use MinusClass for symbol of `MIN'

We can use almost all ruby value can be used by token symbol, except 'false' and 'nil'. These cause unexpected parse error.

If you want to use String as token symbol, special care is required. For example:

  class '"cls"'            # in code, "cls"
  PLUS '"plus\n"'          # in code, "plus\n"
  MIN  "\"minus#{val}\""   # in code, \"minus#{val}\"

Start Rule

'%start' in yacc. This changes start rule.

start real_target

User Code Block

“User Code Block” is a Ruby source code which is copied to output. There are three user code blocks, “header” “inner” and “footer”.

Format of user code is like this:

---- header
  ruby statement
  ruby statement
  ruby statement

---- inner
  ruby statement

If four '-' exist on the line head, Racc treats it as the beginning of the user code block. The name of the user code block must be one word.