Thread: File lexer works perfectly except for numbers

Hi all -

I'm sorry for dropping in again. (You'll be pleased to know I'm back to work tomorrow so I won't be here anywhere near as often..... )

Anyway - my SQL file lexer now works perfectly except that it prints a weird character instead of a number (when there is a number in the SQL input).

Here's the code -

Code:

 
/* toysql_parser.c  */ 
/* A lexer and parser for a very small subset of SQL.  */
/* This version reads the code from a file. */ 

/* This code is released to the public domain. */ 
/* "Share and enjoy......"  :)     */ 


#include <stdio.h>
#include <ctype.h>
#include <string.h>
#include <stdlib.h>


#define NUMBER_OF_KEYWORDS 10
#define MAXLEN 80    // Maximum length of a token. 


/* Array of our keywords in string form. */ 
char *kw_strings[] = { 
   "select", "as", "from", "where", "and", "or", "not", "in", "is", "null" 
    } ; 
   
    
/*  Search function to search the array of keywords. */ 
int search(char *arr[], int dim, char *str) { 
    
    int i;      
    int found_match;
    
    for (i=0; i<dim; i++) { 
        if ( !strcmp(arr[i] , str ) )  {   
            found_match = 1;        
            break; 
    }   else found_match = 0;    
 }  /* For */     

    return found_match; 
}  /* search */ 


/* Enum of token types. */ 
typedef enum { KEYWORD, IDENT, INTEGER, STRING, 
    PUNCT, _EOF_ } tokentype ; 
     

/* A token. */
typedef struct { 
   tokentype toktype;
   union { 
           char string_value[MAXLEN]; 
           int int_value; 
         }  value; 
 } token ;                



/* Forward declaration. */ 
void parse(token *mytok); 
char *ptr ;   // For strtol use.   


void lex(FILE *fp) {    
   
   token *mytok = malloc(sizeof(token)); 
   tokentype mytype;
    
   int i=0;
   
   char myresult[MAXLEN];
   
   int curr_char = fgetc(fp);    
               
  while (curr_char != '\0') {         
    
  if ( isspace(curr_char) ) { 
        curr_char = fgetc(fp);
       }
                
  else if (isalpha(curr_char) || curr_char=='_')  {      
     while ( (isalnum(curr_char) || curr_char=='_') 
         && i<MAXLEN ) {           
            myresult[i] = curr_char;
            curr_char = fgetc(fp);    
            i++;      
        }       
        
    if (search(kw_strings, NUMBER_OF_KEYWORDS, myresult) == 1 )
           mytype = KEYWORD; 
    else
           mytype = IDENT; 
                
   myresult[i] = '\0' ;   
   
   mytok->toktype = mytype;
   strcpy( mytok->value.string_value, myresult);
   
   parse(mytok);                                        
   memset(&myresult[0], 0, sizeof(myresult));  // Clear token          
   }  // Keyword or identifier 

      
 else if ( curr_char == '"' ) { 
       myresult[i] = curr_char;
       curr_char = fgetc(fp);       
       i++;     
        while ( ( curr_char != '"') && curr_char != '\0' 
          && i<MAXLEN-1) {               
            myresult[i] = curr_char;
            curr_char = fgetc(fp);             
            i++;     
          } 
     myresult[i] = curr_char;  // Append the last quote. 
     myresult[i+1] = '\0' ;    // Append null char. 
     curr_char = fgetc(fp);     // Move on from last quote.           
     mytype = STRING; 
     
     mytok->toktype = mytype; 
     strcpy( mytok->value.string_value, myresult);
           
     parse(mytok); 
     memset(&myresult[0], 0, sizeof(myresult));  // Clear token          
   }  // String              
             
             
   else if ( isdigit(curr_char) )  { 
          while (isdigit(curr_char) && curr_char != '\0' 
             && i<MAXLEN) {     
             myresult[i] = curr_char;                            
            curr_char = fgetc(fp);                         
            i++;     
        }  
   myresult[i] = '\0' ;    // Append null char.        
   mytype = INTEGER ;   
   mytok->toktype = mytype; 
   int j=strtol(myresult, &ptr, 10);
   mytok->value.int_value = j;
      
   parse(mytok); 
   memset(&myresult[0], 0, sizeof(myresult));  // Clear token          
   }  // Number                       
               
               
   else if ( ispunct(curr_char) && curr_char != '_' 
        && curr_char != '"' ) { 
             myresult[i] = curr_char;
             curr_char = fgetc(fp);         
             i++;         
      while (ispunct(curr_char) && curr_char != '\0' 
        && i<MAXLEN) {   
               myresult[i] = curr_char;                        
               curr_char = fgetc(fp);           
               i++;                              
            }         
   mytype = PUNCT;   
   
   mytok->toktype = mytype; 
   strcpy(mytok->value.string_value, myresult);
   
   parse(mytok); 
   memset(&myresult[0], 0, sizeof(myresult));  // Clear token           
   }  // Punct 
   
   else {          
           exit(0); 
        } 
                  
   memset(&myresult[0], 0, sizeof(myresult));  // Clear token   
   i = 0;  // Reset i.                     
   }  // while c != '\0'   
    
   exit(0) ;  
    
}  // lex()  

   
   
/* Not a parser (yet) - just prints the tokens. */ 
void parse(token *mytok) { 
       
  printf("%d  ", mytok->toktype); 
     
  if (mytok->value.string_value) 
       printf("%s\n", mytok->value.string_value);
  else if (mytok->value.int_value) 
       printf("%d\n", mytok->value.int_value);       
  
}        
  
          
          
int main(int argc, char **argv) { 

FILE *myfile; 

myfile = fopen(argv[1], "r"); 

if (myfile == NULL) 
   return -1;  
   
lex(myfile);   

fclose(myfile);

return 0; 

}

The file I've used to test it -

Code:

 
select col1 as country, col2 as city from mytable;
select var23 from mytable where city = "Sydney";
select foo,bar,baz from mytable where foo = 20;

A weird character is printed instead of that "20".

Many thanks in advance for any help!
Cheers -
Andy (latte123)

Originally Posted by Salem

1. Your indentation needs work.
2. Your lexer needs some more sub-functions to clarify the overall function.
3. Have you discovered debuggers yet? For example, a breakpoint on line 133 then single step and examine variables will reveal much.

Hi Salem - thanks for your reply -

Points 1 and 2 - understood. I'll work on those.
Debuggers - I've hardly used them (which probably shows). At least in compiling, I always compile with -Wall and -Wpedantic. I did that with this code - no errors or warnings.
I set the breakpoint as you said and got [ Inferior 1 [ process 3178 ] exited with code 0377. ]
I searched the net for code 0377. Found a few pages but none that said what that meant.
If that's the kind of thing that gdb comes back with, it's not hugely helpful.......

Anyway - thanks again for your reply! I'm aware that you didn't have to reply to my post so I appreciate it.

Cheers -
Andy (latte123)