1 files changed, 109 insertions, 2 deletions

diff --git a/sql/strfunc.cc b/sql/strfunc.cc
index 308e6fd3dcd..c03365cfc2b 100644
--- a/sql/strfunc.cc
+++ b/sql/strfunc.cc
@@ -104,7 +104,8 @@ ulonglong find_set(TYPELIB *lib, const char *str, uint length, CHARSET_INFO *cs,
   > 0 position in TYPELIB->type_names +1
 */
 
-uint find_type(TYPELIB *lib, const char *find, uint length, bool part_match)
+uint find_type(const TYPELIB *lib, const char *find, uint length,
+               bool part_match)
 {
   uint found_count=0, found_pos=0;
   const char *end= find+length;
@@ -144,7 +145,8 @@ uint find_type(TYPELIB *lib, const char *find, uint length, bool part_match)
     >0  Offset+1 in typelib for matched string
 */
 
-uint find_type2(TYPELIB *typelib, const char *x, uint length, CHARSET_INFO *cs)
+uint find_type2(const TYPELIB *typelib, const char *x, uint length,
+                CHARSET_INFO *cs)
 {
   int pos;
   const char *j;
@@ -234,3 +236,108 @@ uint check_word(TYPELIB *lib, const char *val, const char *end,
     *end_of_word= ptr;
   return res;
 }
+
+
+/*
+  Converts a string between character sets
+
+  SYNOPSIS
+    strconvert()
+    from_cs       source character set
+    from          source, a null terminated string
+    to            destination buffer
+    to_length     destination buffer length
+
+  NOTES
+    'to' is always terminated with a '\0' character.
+    If there is no enough space to convert whole string,
+    only prefix is converted, and terminated with '\0'.
+
+  RETURN VALUES
+    result string length
+*/
+
+
+uint strconvert(CHARSET_INFO *from_cs, const char *from,
+                CHARSET_INFO *to_cs, char *to, uint to_length, uint *errors)
+{
+  int cnvres;
+  my_wc_t wc;
+  char *to_start= to;
+  uchar *to_end= (uchar*) to + to_length - 1;
+  my_charset_conv_mb_wc mb_wc= from_cs->cset->mb_wc;
+  my_charset_conv_wc_mb wc_mb= to_cs->cset->wc_mb;
+  uint error_count= 0;
+
+  while (1)
+  {
+    /*
+      Using 'from + 10' is safe:
+      - it is enough to scan a single character in any character set.
+      - if remaining string is shorter than 10, then mb_wc will return
+        with error because of unexpected '\0' character.
+    */
+    if ((cnvres= (*mb_wc)(from_cs, &wc,
+                          (uchar*) from, (uchar*) from + 10)) > 0)
+    {
+      if (!wc)
+        break;
+      from+= cnvres;
+    }
+    else if (cnvres == MY_CS_ILSEQ)
+    {
+      error_count++;
+      from++;
+      wc= '?';
+    }
+    else
+      break; // Impossible char.
+
+outp:
+
+    if ((cnvres= (*wc_mb)(to_cs, wc, (uchar*) to, to_end)) > 0)
+      to+= cnvres;
+    else if (cnvres == MY_CS_ILUNI && wc != '?')
+    {
+      error_count++;
+      wc= '?';
+      goto outp;
+    }
+    else
+      break;
+  }
+  *to= '\0';
+  *errors= error_count;
+  return (uint32) (to - to_start);
+
+}
+
+
+/*
+  Searches for a LEX_STRING in an LEX_STRING array.
+
+  SYNOPSIS
+    find_string_in_array()
+      heap    The array
+      needle  The string to search for
+
+  NOTE
+    The last LEX_STRING in the array should have str member set to NULL
+
+  RETURN VALUES
+    -1   Not found
+    >=0  Ordinal position
+*/
+
+int find_string_in_array(LEX_STRING * const haystack, LEX_STRING * const needle,
+                         CHARSET_INFO * const cs)
+{
+  const LEX_STRING *pos;
+  for (pos= haystack; pos->str; pos++)
+    if (!cs->coll->strnncollsp(cs, (uchar *) pos->str, pos->length,
+                               (uchar *) needle->str, needle->length, 0))
+    {
+      return (pos - haystack);
+    }
+  return -1;
+}