YACC

YACC

1. Yacc is a parser generator.
2. Basic form of input specification contains three main secitons
   1. Declarations and directives for the parser generator
   2. Rules of grammar
   3. Program text (i.e. action routines, ... )
   separated by lines containing only "%%".
3. Grammar Rules
   • Each rule takes the form:

     non-term : BODY-1  |  . . .  |  BODY-n ;

   • A BODY is just a list of token names, non-terminal names and literals.
   • A literal is a quoted character:

     expr : expr '+' term

     You will probably not use any literals.
   • Token names are used to identify tokens such as the keywords and token classes such as identifier or integer constant.
     • You can't use quoted strings to identify such tokens literally because all the scanner gives to the parser is a token class and token value.
     • So that Yacc can distinguish token names from non-terminal names, all token names must be declared as such.
   • Names are composed of letters, digits, underscores and periods. Each name must start with a letter. Case is significant.
4. Declarations.
   • To declare a token name, type

     %token name

     in the first section of the specification file
   • One can also specify the numeric value that the scanner will return when it sees this token. It is generally easier, however, to let YACC assign token numbers. This is true because YACC produces a file named `y.tab.h' (which we will rename `tn.h') that contains #define's associating token names that can be used in the scanner with the numbers assigned.
5. Standardizing dates: An example.

   %token MONTH
   %token NUMBER
   
   %%
   date : day MONTH year
        | monthnum '/' day '/' year
        | MONTH day ',' year
        ;
   
   monthnum : NUMBER ;
   day  : NUMBER  ;
   year : NUMBER  ;
         

   • The y.tab.h file produced when YACC processes this file looks like:

     # define MONTH 257
     # define NUMBER 258
     

YACC