Re: encoding

[Paul DuBois]

At 6:51 PM +0000 11/19/00, Gustavo Vieira Goncalves Coelho Rios wrote:
>Dear gentleman,
>
>i am writing my own version of mysql_escape_string (the reason i dont
>want to use mysql version is i dont want to encode the NULL carac.)

Why is that? If your string contains null bytes and you insert it into
a query that you send to the server, the query string will be incorrect.

>
>Here it is:
>
>unsigned int
>enc2mysql(char *to, char *from, unsigned int len)
>{
>         unsigned int to_len;
>
>         for (to_len = 0; len; ++from, len--)
>                 switch (*from) {
>                 case '\n' :
>                         *to++ = '\\';
>                         *to++ = 'n';
>                         to_len += 2;
>                 break;
>                 case '\r' :
>                         *to++ = '\\';
>                         *to++ = 'r';
>                         to_len += 2;
>                 break;
>                 case '\\' : case '\'' : case '"' :
>                         *to++ = '\\';
>                         *to++ = *from;
>                         to_len += 2;
>                 break;
>                 default :
>                 case '\\' : case '\'' : case '"' :
>                         *to++ = '\\';
>                         *to++ = *from;
>                         to_len += 2;
>                 break;
>                 default :
>                         *to++ = *from;
>                         to_len++;
>                 }

Why do you have multiple cases for \, ', ", and default?


>         *to = *from;
>
>         return to_len;
>}
>
>
>My questions is: what should i encode CTRL Z to?

The easiest way to find out is to examine the source for mysql_escape_string,
which is in libmysql/libmysql.c.  This is what it does:

     case '\032':            /* This gives problems on Win32 */
       *to++= '\\';
       *to++= 'Z';
       break;

-- 
Paul DuBois, paul@stripped