COMMONWEALTH OF AUSTRALIA Copyright Regulations 1969 WARNING This material has been reproduced and communicated to you by or on behalf of Monash University pursuant to Part VB of the Copyright Act 1968 (the Act). The material in this communication may be subject to copyright under the Act. Any further reproduction or communication of this material by you may be the subject of copyright protection under the Act. Do not remove this notice.

Parsers CSE2303 Formal Methods I Lecture 12

Overview Recursive Descent Parsers LR Parsers Bison

Parsers A parser for a grammar is a program. –Input is a string. –Decides whether the input can be generated by the grammar. Two main types –top-down parsers –bottom-up parsers

Difficulties with Top-down Parsers Left Recursive Grammars i.e A  …  Aw Error Handling Backtracking –Allocating/Deallocating resources –Undoing actions

LR Parser Bottom-up Parser Scan input Left to Right Construct the Rightmost derivation in reverse Implemented using –A Finite Automaton and a Stack.

Pros and Cons Benefits –Can construct a LR parser to recognise most CFGs –The parsers are efficient –Detect syntactical errors as soon as possible Disadvantages –Can’t build LR parser for every CFG –Need a Parser Generator

Bison (yacc) Parser generator –It writes a LR parser Assumption –Grammar is an LALR grammar Needs –set of production rules –An action for each rule Produces –A C program

main() { yyparse(); } int yyparse() { yychar = yylex(); } Parse input int yylex() { return …; } Get next token Return an int representing next token int yyerror() { return 0; } If input cannot be parsed Input

Process Write a bison program and a flex program E.g. example.y and example.l Run them through bison and flex bison -d example.y flex example.l Compile the program with the flag -lfl example.y example.l Contains yyparse() Contains definitions lex.yy.c flex example.tab.c example.tab.h bison Contains yylex()

A Bison Program … definitions … % … rules … % … subroutines … bison does not handle carriage returns

Sections … definition section –Code between %{ … %} copied. –Definitions used to define tokens, types, etc. … rule section –Pairs of production rules and actions. –The productions rules are from a CFG. … subroutine section –Consists of users subroutines. –Copied after the end of the bison generated code –must contain yyerror()

%. {return yytext[0];} \n {return 0;} % simple.l int main() { yyparse(); } simple.c % S: B B {printf(“S -> BB\n”);} ; B: ‘a’ B {printf(“B -> aB\n”);} | ‘b’ {printf(“B -> b\n”);} ; % int yyerror(char* s) { printf(“%s\n”, s); return 0; } simple.y gcc –o simple simple.c lex.yy.c simple.tab.c -lfl

Evaluation of 4+2*3 S  E E  T | T + E T  F | F * T F  INT yylval = 4 S E INT +TE F*T T F F yylval = 2 yylval = 3 =3 =6 =3 =2 =4 =6 =4 =10

%{ #include “plusTimes.tab.h” %} % -?[0-9]+ { yylval = atoi(yytext); return INT; } [ \t]. {return yytext[0];} \n {return 0;} % plusTimes.l %token INT % S: E {printf(“%d\n”, $1);} ; E: T {$$ = $1;} | T ‘+’ E {$$ = $1 + $3;} ; T: F {$$ = $1;} | F ‘*’ T {$$ = $1 * $3;} ; F: INT {$$ = $1;} ; % … plusTimes.y

%{ #include “RealPlusTimes.tab.h” %} Real -?([0-9]+|([0-9]*\.[0-9]+)) % {Real} { yylval.dval = atof(yytext); return REAL; } [ \t]. {return yytext[0];} \n {return 0;} % RealPlusTimes.l %union{double dval;} %token REAL %type E T F % S: E {printf(“%g\n”, $1);} ; E: T {$$ = $1;} | T ‘+’ E {$$ = $1 + $3;} ; T: F {$$ = $1;} | F ‘*’ T {$$ = $1 * $3;} ; F: REAL {$$ = $1;} ; % … RealPlusTimes.y

More Information Check the courseware web site. Man pages –login and type: xman bison Library –“lex & yacc”, by John Levine et al. –“Principles of Compiler Design”, by A.V. Aho and J.D. Ullman –“The Unix Programming Environment”, by Kernighan & Pike. (see chapter 8)

Preparation Read –Chapter in the Text Book.