Fixing compilation warnings introduced in:

> commit e0929951
> Author: Alexander Barkov <bar@mariadb.org>
> Date:   Wed Mar 16 10:55:12 2016 +0400
>
>     MDEV-9665 Remove cs->cset->ismbchar()
>     Using a more powerfull cs->cset->charlen() instead.

strings/ctype-big5.c

@@ -862,7 +862,7 @@ my_strnxfrm_big5(CHARSET_INFO *cs,
   
   for (; dst < de && src < se && nweights; nweights--)
   {
-    if (my_charlen(cs, src, se) > 1)
+    if (my_charlen(cs, (const char *) src, (const char *) se) > 1)
     {
       /*
         Note, it is safe not to check (src < se)

strings/ctype-gbk.c

@@ -3463,7 +3463,7 @@ my_strnxfrm_gbk(CHARSET_INFO *cs,
 
   for (; dst < de && src < se && nweights; nweights--)
   {
-    if (my_charlen(cs, src, se) > 1)
+    if (my_charlen(cs, (const char *) src, (const char *) se) > 1)
     {
       /*
         Note, it is safe not to check (src < se)

strings/ctype-mb.c

@@ -668,7 +668,7 @@ my_strnncollsp_mb_bin(CHARSET_INFO * cs __attribute__((unused)),
 */
 #define my_strnxfrm_mb_non_ascii_char(cs, dst, src, se)                  \
 {                                                                        \
-  switch (my_ismbchar(cs, src, se))                                    { \
+  switch (my_ismbchar(cs, (const char *) src, (const char *) se)) {      \
   case 4:                                                                \
     *dst++= *src++;                                                      \
     /* fall through */                                                   \
@@ -740,7 +740,8 @@ my_strnxfrm_mb(CHARSET_INFO *cs,
   for (; src < se && nweights && dst < de; nweights--)
   {
     int chlen;
-    if (*src < 128 || !(chlen= my_ismbchar(cs, src, se)))
+    if (*src < 128 || !(chlen= my_ismbchar(cs, (const char *) src,
+                                               (const char *) se)))
     {
       /* Single byte character */
       *dst++= sort_order ? sort_order[*src++] : *src++;