[parser] *: change default charset from utf8 to utf8mb4 (#13)

pingcap · Oct 9, 2021 · 6f2c711 · 6f2c711
1 parent 2095376
commit 6f2c711
Show file tree

Hide file tree

Showing 5 changed files with 23 additions and 16 deletions.
diff --git a/parser/charset/charset.go b/parser/charset/charset.go
@@ -115,6 +115,11 @@ func GetDefaultCollation(charset string) (string, error) {
 	return c.DefaultCollation, nil
 }
 
+// GetDefaultCharsetAndCollate returns the default charset and collation.
+func GetDefaultCharsetAndCollate() (string, string) {
+	return mysql.DefaultCharset, mysql.DefaultCollationName
+}
+
 // GetCharsetInfo returns charset and collation for cs as name.
 func GetCharsetInfo(cs string) (string, string, error) {
 	c, ok := charsets[strings.ToLower(cs)]

diff --git a/parser/mysql/charset.go b/parser/mysql/charset.go
@@ -551,13 +551,15 @@ var CollationNames = map[string]uint8{
 
 // MySQL collation information.
 const (
-	UTF8Charset          = "utf8"
-	UTF8MB4Charset       = "utf8mb4"
-	DefaultCharset       = UTF8Charset
-	DefaultCollationID   = 83
-	BinaryCollationID    = 63
-	UTF8DefaultCollation = "utf8_bin"
-	DefaultCollationName = UTF8DefaultCollation
+	UTF8Charset    = "utf8"
+	UTF8MB4Charset = "utf8mb4"
+	DefaultCharset = UTF8MB4Charset
+	// DefaultCollationID is utf8mb4_bin(46)
+	DefaultCollationID      = 46
+	BinaryCollationID       = 63
+	UTF8DefaultCollation    = "utf8_bin"
+	UTF8MB4DefaultCollation = "utf8mb4_bin"
+	DefaultCollationName    = UTF8MB4DefaultCollation
 
 	// MaxBytesOfCharacter, is the max bytes length of a character,
 	// refer to RFC3629, in UTF-8, characters from the U+0000..U+10FFFF range

diff --git a/parser/parser.go b/parser/parser.go
diff --git a/parser/parser.y b/parser/parser.y
@@ -4072,8 +4072,8 @@ CastType:
 			x.Flag |= mysql.BinaryFlag
 		}
 		if x.Charset == "" {
-			x.Charset = charset.CharsetUTF8
-			x.Collate = charset.CollationUTF8
+			x.Charset = mysql.DefaultCharset
+			x.Collate = mysql.DefaultCollationName
 		}
 		$$ = x
 	}
@@ -4142,8 +4142,8 @@ CastType:
 	{
 		x := types.NewFieldType(mysql.TypeJSON)
 		x.Flag |= mysql.BinaryFlag | (mysql.ParseToJSONFlag)
-		x.Charset = charset.CharsetUTF8
-		x.Collate = charset.CollationUTF8
+		x.Charset = mysql.DefaultCharset
+		x.Collate = mysql.DefaultCollationName
 		$$ = x
 	}
 

diff --git a/parser/types/field_type.go b/parser/types/field_type.go
@@ -204,7 +204,7 @@ func (ft *FieldType) FormatAsCastType(w io.Writer) {
 		if ft.Flag&mysql.BinaryFlag != 0 {
 			fmt.Fprint(w, " BINARY")
 		}
-		if ft.Charset != charset.CharsetBin && ft.Charset != charset.CharsetUTF8 {
+		if ft.Charset != charset.CharsetBin && ft.Charset != mysql.DefaultCharset {
 			fmt.Fprintf(w, " %s", ft.Charset)
 		}
 	case mysql.TypeDate: