URL:
  <http://savannah.nongnu.org/bugs/?19079>

                 Summary: sscanf %s eats 1 char too much
                 Project: AVR C Runtime Library
            Submitted by: None
            Submitted on: Sunday 02/18/2007 at 15:13 UTC
                Category: Library
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: libc code
                  Status: None
                 Privacy: Private
        Percent Complete: 0%
             Assigned to: None
        Originator Email: [EMAIL PROTECTED]
             Open/Closed: Open
         Discussion Lock: Any

    _______________________________________________________

Details:

Hello, I think we have a bug here.

example code:

char wsp_N[10];
char wsp_E[11];
int  fixva  = 0;
int  nr_sat = 0;

int x = "$GP,194000.00,5001.4934,N,01954.5390,E,0,02,2.9";
int nrconv = sscanf(x, "$GP,%*u.%*u,%9s,N,%10s,E,%i,%i,%*s",
                                wsp_N, wsp_E, &fixva, &nr_sat);

nrconv will be equal 1, wsp_N equals 5001.4934, wsp_E wont be event touched!
It is because sscanf eats 1 char too much. This occurs only in situation
where pattern %9s will terminate on max string length (9 here), no on
whitespace. Then will eat ',' character above, before N from. If we remove
',' from pattern, the second %s will match, and nrconv will be equal 2. And
so on.

There is a fix, which works for me:
[EMAIL PROTECTED]:~/work/avrlibc/avr-libc-1.4.5/libc/stdio$ diff vfscanf.c
vfscanf.c.org
290,291d289
<
<                               ungetc(i,stream);       /* we eat too much
*/
[EMAIL PROTECTED]:~/work/avrlibc/avr-libc-1.4.5/libc/stdio$

Regards,
Jerzy




    _______________________________________________________

Reply to this item at:

  <http://savannah.nongnu.org/bugs/?19079>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.nongnu.org/



_______________________________________________
AVR-libc-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/avr-libc-dev

Reply via email to