Simplify caseup() and casedn() in charsets

After the MDEV-13118 fix there's no code in the server that
wants caseup/casedn to change the argument in place for simple
charsets.  Let's remove this logic and always return the result in a
new string for all charsets, both simple and complex.

1. Removing the optimization that *some* character sets used in casedn()
  and caseup(), which allowed (and required) to change the case in-place,
  overwriting the string passed as the "src" argument.
  Now all CHARSET_INFO's work in the same way:
  non of them change the source string in-place, all of them now convert
  case from the source string to the destination string, leaving
  the source string untouched.

2. Adding "const" qualifier to the "char *src" parameter
   to caseup() and casedn().

3. Removing duplicate implementations in ctype-mb.c.
  Now both caseup() and casedn() implementations for all CJK character sets
  use internally the same function my_casefold_mb()
  (the former my_casefold_mb_varlen()).

4. Removing the "unused" attribute from parameters of some my_case{up|dn}_xxx()
   implementations, as the affected parameters are now *used* in the code.
   Previously these parameters were used only in DBUG_ASSERT().

include/m_ctype.h

@@ -365,7 +365,7 @@ typedef int (*my_charset_conv_mb_wc)(CHARSET_INFO *, my_wc_t *,
 typedef int (*my_charset_conv_wc_mb)(CHARSET_INFO *, my_wc_t,
                                      uchar *, uchar *);
 typedef size_t (*my_charset_conv_case)(CHARSET_INFO *,
-                                       char *, size_t, char *, size_t);
+                                       const char *, size_t, char *, size_t);
 
 
 /* See strings/CHARSET_INFO.txt about information on this structure  */
@@ -565,9 +565,11 @@ extern uint my_instr_simple(CHARSET_INFO *,
 /* Functions for 8bit */
 extern size_t my_caseup_str_8bit(CHARSET_INFO *, char *);
 extern size_t my_casedn_str_8bit(CHARSET_INFO *, char *);
-extern size_t my_caseup_8bit(CHARSET_INFO *, char *src, size_t srclen,
+extern size_t my_caseup_8bit(CHARSET_INFO *,
+                             const char *src, size_t srclen,
                              char *dst, size_t dstlen);
-extern size_t my_casedn_8bit(CHARSET_INFO *, char *src, size_t srclen,
+extern size_t my_casedn_8bit(CHARSET_INFO *,
+                             const char *src, size_t srclen,
                              char *dst, size_t dstlen);
 
 extern int my_strcasecmp_8bit(CHARSET_INFO * cs, const char *, const char *);
@@ -658,17 +660,17 @@ uint my_mbcharlen_8bit(CHARSET_INFO *, uint c);
 /* Functions for multibyte charsets */
 extern size_t my_caseup_str_mb(CHARSET_INFO *, char *);
 extern size_t my_casedn_str_mb(CHARSET_INFO *, char *);
-extern size_t my_caseup_mb(CHARSET_INFO *, char *src, size_t srclen,
-                                         char *dst, size_t dstlen);
-extern size_t my_casedn_mb(CHARSET_INFO *, char *src, size_t srclen,
-                                         char *dst, size_t dstlen);
-extern size_t my_caseup_mb_varlen(CHARSET_INFO *, char *src, size_t srclen,
-                                  char *dst, size_t dstlen);
-extern size_t my_casedn_mb_varlen(CHARSET_INFO *, char *src, size_t srclen,
-                                  char *dst, size_t dstlen);
-extern size_t my_caseup_ujis(CHARSET_INFO *, char *src, size_t srclen,
+extern size_t my_caseup_mb(CHARSET_INFO *,
+                           const char *src, size_t srclen,
+                           char *dst, size_t dstlen);
+extern size_t my_casedn_mb(CHARSET_INFO *,
+                           const char *src, size_t srclen,
+                           char *dst, size_t dstlen);
+extern size_t my_caseup_ujis(CHARSET_INFO *,
+                             const char *src, size_t srclen,
                              char *dst, size_t dstlen);
-extern size_t my_casedn_ujis(CHARSET_INFO *, char *src, size_t srclen,
+extern size_t my_casedn_ujis(CHARSET_INFO *,
+                             const char *src, size_t srclen,
                              char *dst, size_t dstlen);
 extern int my_strcasecmp_mb(CHARSET_INFO * cs,const char *, const char *);
 

sql/item_strfunc.cc

@@ -1572,19 +1572,10 @@ String *Item_str_conv::val_str(String *str)
                     str->alloc((alloced_length= res->length() * multiply)))))
     return 0;
 
-  if (multiply == 1)
-  {
-    str->copy(*res); // Should not fail (it was alloced above)
-    len= converter(collation.collation, (char*) str->ptr(), str->length(),
-                                        (char*) str->ptr(), alloced_length);
-  }
-  else
-  {
-    len= converter(collation.collation, (char*) res->ptr(), res->length(),
-                                        (char*) str->ptr(), alloced_length);
-    str->set_charset(collation.collation);
-  }
+  len= converter(collation.collation, (char*) res->ptr(), res->length(),
+                                      (char*) str->ptr(), alloced_length);
   DBUG_ASSERT(len <= alloced_length);
+  str->set_charset(collation.collation);
   str->length(len);
   return str;
 }

strings/ctype-bin.c

@@ -220,11 +220,11 @@ static size_t my_case_str_bin(CHARSET_INFO *cs __attribute__((unused)),
 
 
 static size_t my_case_bin(CHARSET_INFO *cs __attribute__((unused)),
-                          char *src __attribute__((unused)),
-                          size_t srclen,
-                          char *dst __attribute__((unused)),
-                          size_t dstlen __attribute__((unused)))
+                          const char *src, size_t srclen,
+                          char *dst, size_t dstlen)
 {
+  DBUG_ASSERT(srclen <= dstlen);
+  memcpy(dst, src, srclen);
   return srclen;
 }
 

strings/ctype-euc_kr.c

@@ -9994,8 +9994,8 @@ static MY_CHARSET_HANDLER my_charset_handler=
   my_mb_ctype_mb,
   my_caseup_str_mb,
   my_casedn_str_mb,
-  my_caseup_mb_varlen, /* UPPER() can reduce length: Turkish DOTLESS i -> I */
-  my_casedn_mb,        /* LOWER() does not change length, use simple version*/
+  my_caseup_mb,        /* UPPER() can reduce length: Turkish DOTLESS i -> I */
+  my_casedn_mb,        /* LOWER() does not change length */
   my_snprintf_8bit,
   my_long10_to_str_8bit,
   my_longlong10_to_str_8bit,

strings/ctype-mb.c

@@ -71,95 +71,23 @@ get_case_info_for_ch(const CHARSET_INFO *cs, uint page, uint offs)
 
 
 /*
-  For character sets which don't change octet length in case conversion.
-*/
-size_t my_caseup_mb(CHARSET_INFO * cs, char *src, size_t srclen,
-                    char *dst __attribute__((unused)),
-                    size_t dstlen __attribute__((unused)))
-{
-  register uint32 l;
-  register char *srcend= src + srclen;
-  register const uchar *map= cs->to_upper;
-
-  DBUG_ASSERT(cs->caseup_multiply == 1);
-  DBUG_ASSERT(src == dst && srclen == dstlen);
-  DBUG_ASSERT(cs->mbmaxlen == 2);
-
-  while (src < srcend)
-  {
-    if ((l=my_ismbchar(cs, src, srcend)))
-    {
-      MY_UNICASE_CHARACTER *ch;
-      if ((ch= get_case_info_for_ch(cs, (uchar) src[0], (uchar) src[1])))
-      {
-        *src++= ch->toupper >> 8;
-        *src++= ch->toupper & 0xFF;
-      }
-      else
-        src+= l;
-    }
-    else 
-    {
-      *src=(char) map[(uchar) *src];
-      src++;
-    }
-  }
-  return srclen;
-}
-
-
-size_t my_casedn_mb(CHARSET_INFO * cs, char *src, size_t srclen,
-                    char *dst __attribute__((unused)),
-                    size_t dstlen __attribute__((unused)))
-{
-  register uint32 l;
-  register char *srcend= src + srclen;
-  register const uchar *map=cs->to_lower;
-
-  DBUG_ASSERT(cs->casedn_multiply == 1);
-  DBUG_ASSERT(src == dst && srclen == dstlen);  
-  DBUG_ASSERT(cs->mbmaxlen == 2);
-
-  while (src < srcend)
-  {
-    if ((l= my_ismbchar(cs, src, srcend)))
-    {
-      MY_UNICASE_CHARACTER *ch;
-      if ((ch= get_case_info_for_ch(cs, (uchar) src[0], (uchar) src[1])))
-      {
-        *src++= ch->tolower >> 8;
-        *src++= ch->tolower & 0xFF;
-      }
-      else
-        src+= l;
-    }
-    else
-    {
-      *src= (char) map[(uchar)*src];
-      src++;
-    }
-  }
-  return srclen;
-}
-
-
-/*
-  Case folding functions for character set
-  where case conversion can change string octet length.
+  Case folding functions for CJK character set.
+  Case conversion can optionally reduce string octet length.
   For example, in EUCKR,
     _euckr 0xA9A5 == "LATIN LETTER DOTLESS I" (Turkish letter)
   is upper-cased to to
     _euckr 0x49 "LATIN CAPITAL LETTER I"  ('usual' letter I)
   Length is reduced in this example from two bytes to one byte.
 */
 static size_t
-my_casefold_mb_varlen(CHARSET_INFO *cs,
-                      char *src, size_t srclen,
-                      char *dst, size_t dstlen __attribute__((unused)),
-                      const uchar *map,
-                      size_t is_upper)
+my_casefold_mb(CHARSET_INFO *cs,
+               const char *src, size_t srclen,
+               char *dst, size_t dstlen __attribute__((unused)),
+               const uchar *map,
+               size_t is_upper)
 {
-  char *srcend= src + srclen, *dst0= dst;
+  const char *srcend= src + srclen;
+  char *dst0= dst;
 
   DBUG_ASSERT(cs->mbmaxlen == 2);
 
@@ -193,22 +121,22 @@ my_casefold_mb_varlen(CHARSET_INFO *cs,
 
 
 size_t
-my_casedn_mb_varlen(CHARSET_INFO * cs, char *src, size_t srclen,
+my_casedn_mb(CHARSET_INFO * cs, const char *src, size_t srclen,
                     char *dst, size_t dstlen)
 {
   DBUG_ASSERT(dstlen >= srclen * cs->casedn_multiply); 
   DBUG_ASSERT(src != dst || cs->casedn_multiply == 1);
-  return my_casefold_mb_varlen(cs, src, srclen, dst, dstlen, cs->to_lower, 0);
+  return my_casefold_mb(cs, src, srclen, dst, dstlen, cs->to_lower, 0);
 }
 
 
 size_t
-my_caseup_mb_varlen(CHARSET_INFO * cs, char *src, size_t srclen,
-                    char *dst, size_t dstlen)
+my_caseup_mb(CHARSET_INFO * cs, const char *src, size_t srclen,
+             char *dst, size_t dstlen)
 {
   DBUG_ASSERT(dstlen >= srclen * cs->caseup_multiply);
   DBUG_ASSERT(src != dst || cs->caseup_multiply == 1);
-  return my_casefold_mb_varlen(cs, src, srclen, dst, dstlen, cs->to_upper, 1);
+  return my_casefold_mb(cs, src, srclen, dst, dstlen, cs->to_upper, 1);
 }
 
 

strings/ctype-simple.c

@@ -214,28 +214,26 @@ size_t my_casedn_str_8bit(CHARSET_INFO * cs,char *str)
 }
 
 
-size_t my_caseup_8bit(CHARSET_INFO * cs, char *src, size_t srclen,
-                      char *dst __attribute__((unused)),
-                      size_t dstlen __attribute__((unused)))
+size_t my_caseup_8bit(CHARSET_INFO * cs, const char *src, size_t srclen,
+                      char *dst, size_t dstlen)
 {
-  char *end= src + srclen;
+  const char *end= src + srclen;
   register const uchar *map= cs->to_upper;
-  DBUG_ASSERT(src == dst && srclen == dstlen);
+  DBUG_ASSERT(srclen <= dstlen);
   for ( ; src != end ; src++)
-    *src= (char) map[(uchar) *src];
+    *dst++= (char) map[(uchar) *src];
   return srclen;
 }
 
 
-size_t my_casedn_8bit(CHARSET_INFO * cs, char *src, size_t srclen,
-                      char *dst __attribute__((unused)),
-                      size_t dstlen __attribute__((unused)))
+size_t my_casedn_8bit(CHARSET_INFO * cs, const char *src, size_t srclen,
+                      char *dst, size_t dstlen)
 {
-  char *end= src + srclen;
+  const char *end= src + srclen;
   register const uchar *map=cs->to_lower;
-  DBUG_ASSERT(src == dst && srclen == dstlen);
+  DBUG_ASSERT(srclen <= dstlen);
   for ( ; src != end ; src++)
-    *src= (char) map[(uchar) *src];
+    *dst++= (char) map[(uchar) *src];
   return srclen;
 }