MDEV-9665 Remove cs->cset->ismbchar()

Using a more powerfull cs->cset->charlen() instead.

1 files changed, 36 insertions, 3 deletions

diff --git a/include/m_ctype.h b/include/m_ctype.h
index ee49e94ef99..b891e89d00d 100644
--- a/include/m_ctype.h
+++ b/include/m_ctype.h
@@ -400,7 +400,6 @@ struct my_charset_handler_st
 {
   my_bool (*init)(struct charset_info_st *, MY_CHARSET_LOADER *loader);
   /* Multibyte routines */
-  uint    (*ismbchar)(CHARSET_INFO *, const char *, const char *);
   uint    (*mbcharlen)(CHARSET_INFO *, uint c);
   size_t  (*numchars)(CHARSET_INFO *, const char *b, const char *e);
   size_t  (*charpos)(CHARSET_INFO *, const char *b, const char *e,
@@ -972,8 +971,42 @@ size_t my_convert_fix(CHARSET_INFO *dstcs, char *dst, size_t dst_length,
 #define my_strcasecmp(s, a, b)        ((s)->coll->strcasecmp((s), (a), (b)))
 #define my_charpos(cs, b, e, num)     (cs)->cset->charpos((cs), (const char*) (b), (const char *)(e), (num))
 
-#define use_mb(s)                     ((s)->cset->ismbchar != NULL)
-#define my_ismbchar(s, a, b)          ((s)->cset->ismbchar((s), (a), (b)))
+#define use_mb(s)                     ((s)->mbmaxlen > 1)
+/**
+  Detect if the leftmost character in a string is a valid multi-byte character
+  and return its length, or return 0 otherwise.
+  @param cs  - character set
+  @param str - the beginning of the string
+  @param end - the string end (the next byte after the string)
+  @return    >0, for a multi-byte character
+  @rerurn    0,  for a single byte character, broken sequence, empty string.
+*/
+static inline
+uint my_ismbchar(CHARSET_INFO *cs, const char *str, const char *end)
+{
+  int char_length= (cs->cset->charlen)(cs, (const uchar *) str,
+                                           (const uchar *) end);
+  return char_length > 1 ? (uint) char_length : 0U;
+}
+
+
+/**
+  Return length of the leftmost character in a string.
+  @param cs  - character set
+  @param str - the beginning of the string
+  @param end - the string end (the next byte after the string)
+  @return  <=0 on errors (EOL, wrong byte sequence)
+  @return    1 on a single byte character
+  @return   >1 on a multi-byte character
+
+  Note, inlike my_ismbchar(), 1 is returned for a single byte character.
+*/
+static inline
+uint my_charlen(CHARSET_INFO *cs, const char *str, const char *end)
+{
+  return (cs->cset->charlen)(cs, (const uchar *) str,
+                                 (const uchar *) end);
+}
 #ifdef USE_MB
 #define my_mbcharlen(s, a)            ((s)->cset->mbcharlen((s),(a)))
 #else