1. CS 310 – Fall 2006 Pacific University CS310 Lex & Yacc Today’s reference: UNIX Programming Tools: lex & yacc by: Levine, Mason, Brown Chapter 1, 2, 3 November 1, 2006

2. CS 310 – Fall 2006 Pacific University Automatic Parser Generators You supply the tokenizer –Lex/Flex –code to break the string into terminals You supply the grammar Output source code that will produce a parser –Yacc/Bison (produces C code) –NewYacc (produces C code, a bit more powerful than Yacc, keeps more state) –Jacc (produces Java code) –decorate the grammar rules with source code to perform various tasks

3. CS 310 – Fall 2006 Pacific University Lex (Flex*) Lex: –specify the terminals or regular expressions that represent the terminals –produces a C source file as output that will divide input into tokens * Flex and Bison are the GNU equivalents of Lex and Yacc file.l -> lex -> lex.yy.c

4. CS 310 – Fall 2006 Pacific University %{ /* sample demonstration*/ /* identify is and are as verbs */ %} % [\t ]+/* ignore whitespace */ is | are { printf(“%s: is a verb”, yytext); } [a-zA-Z]+ { printf(“%s: not a verb”, yytext);} % main() { yylex(); }

5. CS 310 – Fall 2006 Pacific University %{ /* count verbs and nonverbs */ int verbCount=0, nonVerbCount=0; %} % [\t ]+/* ignore whitespace */ is | are { printf(“%d verbs\n”, ++verbCount); } [a-zA-Z]+ { printf(“%d non-verbs\n”, ++nonVerbCount); } % main() { yylex(); printf(“%d verbs| %d nonverbs\n”, verbCount, nonVerbCount); }

6. CS 310 – Fall 2006 Pacific University Yacc (Bison) Yacc: –take the tokens (terminals) from lex and apply a grammar –produce C file –check syntax –y.tab.c contains the C code to apply the grammar –y.tab.h contains the data structures to be used by lex to pass data to yacc file.y -> yacc -> y.tab.[c|h] ♦ ♦ bison names the file file.tab.[c|h]

7. CS 310 – Fall 2006 Pacific University %{ #include %} %token VERB NOUN % paragraph: sentence { printf(“paragraph\n”); } | paragraph sentence {printf(“paragraph\n”);} sentence: nounphrase verbphrase { printf(“sentence\n”); } nounphrase: NOUN { printf(“noun\n”); } verbphrase: VERB { printf(“verb\n”); } % extern FILE *yyin; extern int lineno; extern char* yytext; main(){ do{ yyparse(); }while(!feof(yyin)); } void yyerror(char *msg){ printf(“ERROR: (%d:%s) %s\n”,lineno, yytext, msg); }

8. CS 310 – Fall 2006 Pacific University What lex file will we need? %{ /* identify is and are as verbs */ /* identify computer, bob, alice as nouns */ #include “y.tab.h” /* shared data with yacc*/ int lineno=1; %} % [\t ]+/* ignore whitespace */ is | are { return(VERB); } computer | bob | alice { return(NOUN);} \n{ lineno++;} %

9. CS 310 – Fall 2006 Pacific University Build the executable yacc –dv yaccfile.y lex lexfile.l gcc –o parser lex.yy.c y.tab.c -lfl Examples will be posted after class on the CS310 schedule! A Makefile will be provided.

10. CS 310 – Fall 2006 Pacific University Shift/Reduce Conflicts stmt: IF expr THEN stmt | IF expr THEN stmt ELSE stmt Solutions: –Rewrite the grammar. –Use precedence to tell the parser how to handle this %nonassoc LOWER_THAN_ELSE %nonassoc ELSE stmt: IF expr THEN stmt %prec LOWER_THAN_ELSE | IF expr THEN stmt ELSE stmt