Thread: mbstowcs issue

Hello, my procedure should take a string and parse all all values separated by comma ',' and store them in objQueue[qid].filtValue array. This array is of wide-char type.

Code:

typedef
struct
{
  char    *name;		
  int     filtValueCount;	
  wchar_t **filtValue;		
} tQueueDef;
tQueueDef *objQueue;

/*
qid - identifier (not important now)
*s - pointer to parsed string
*len - lenght of the parsed string
*cnt - number of commas in the parsed string
*/
void parseFiltValues(int qid, char *s, int *len, int *cnt)
{
  int i, j, k;
  char *ptr, *tmp;

  tmp = (char *)malloc((MSG_LENGTH) * sizeof(char));

  printf("ParseFiltValues qid: %d, len: %d, cnt: %d\n", qid, *len, *cnt);
  if (*len == 0)
    return;

  objQueue[qid].filtValueCount = *cnt;
  objQueue[qid].filtValue = (wchar_t **)malloc(*cnt * sizeof(wchar_t *));

  for (ptr = s, i = j = k = 0; i <= *len; i++, j++)
  {
    switch (s[i])
    {
      case ',':
      case '}':
        strncpy(tmp, ptr, j);
        tmp[j] = '\0';

        objQueue[qid].filtValue[k] = (wchar_t *)malloc(MSG_LENGTH * sizeof(wchar_t));
        mbstowcs(objQueue[qid].filtValue[k], tmp, MSG_LENGTH);

        printf("filt val (%d): %s<%d> <%d>\n", k, tmp, strlen(tmp), wcslen(objQueue[qid].filtValue[k]));

        j = -1;
        ptr = ss + i + 1;  
        k++;
        break;
    } 
  }

  free(tmp);
}

So, for example, for input string {AA,AB,AD,AF}, the output should read:
filt val(0): AA<2> <2>
filt val(1): AB<2> <2>
filt val(2): AD<2> <2>
filt val(3): AF<2> <2>

It works well until an Eastern European character arrives to the input. For input string {AB,AC,AČ,AŇA,AF,AFŠ,AZ} the output reads, which means the mbstowcs does not convert it properly.
filt val(0): AA<2> <2>
filt val(1): AC<2> <2>
filt val(2): AČ<3> <1>
filt val(3): AŇA<4> <1>
filt val(4): AF<2> <2>
filt val(5): AFŠ<4> <2>
filt val(6): AZ<2> <2>

Can anyone help me here please? It's weird because I use similar funtionality elsewhere in the program and there it works just fine... Thanks a lot!