*: set latin1 to be an alias for utf8mb4 in TiDB

pingcap · May 5, 2022 · 55c4527 · 55c4527
1 parent 4e21680
commit 55c4527
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 31 deletions.
diff --git a/ddl/ddl_api.go b/ddl/ddl_api.go
@@ -4177,8 +4177,10 @@ func checkModifyCharsetAndCollation(toCharset, toCollate, origCharset, origColla
 
  if (origCharset == charset.CharsetUTF8 && toCharset == charset.CharsetUTF8MB4) ||
  (origCharset == charset.CharsetUTF8 && toCharset == charset.CharsetUTF8) ||
- (origCharset == charset.CharsetUTF8MB4 && toCharset == charset.CharsetUTF8MB4) {
- // TiDB only allow utf8 to be changed to utf8mb4, or changing the collation when the charset is utf8/utf8mb4.
+ (origCharset == charset.CharsetUTF8MB4 && toCharset == charset.CharsetUTF8MB4) ||
+ (origCharset == charset.CharsetLatin1 && toCharset == charset.CharsetUTF8) ||
+ (origCharset == charset.CharsetLatin1 && toCharset == charset.CharsetUTF8MB4) {
+ // TiDB only allow utf8/latin1 to be changed to utf8mb4, or changing the collation when the charset is utf8/utf8mb4/latin1.
  return nil
  }
 

diff --git a/parser/charset/encoding_latin1.go b/parser/charset/encoding_latin1.go
@@ -14,20 +14,18 @@
 package charset
 
 import (
- "bytes"
-
  "golang.org/x/text/encoding"
 )
 
 // EncodingLatin1Impl is the instance of encodingLatin1.
-// TiDB uses utf8 implementation for latin1 charset because of the backward compatibility.
+// In TiDB, latin1 is an alias for utf8, so uses utf8 implementation for latin1
 var EncodingLatin1Impl = &encodingLatin1{encodingUTF8{encodingBase{enc: encoding.Nop}}}
 
 func init() {
  EncodingLatin1Impl.self = EncodingLatin1Impl
 }
 
-// encodingLatin1 compatibles with latin1 in old version TiDB.
+// encodingLatin1 compatibles with latin1 in old version TiDB
 type encodingLatin1 struct {
  encodingUTF8
 }
@@ -36,25 +34,3 @@ type encodingLatin1 struct {
 func (e *encodingLatin1) Name() string {
  return CharsetLatin1
 }
-
-// Peek implements Encoding interface.
-func (e *encodingLatin1) Peek(src []byte) []byte {
- if len(src) == 0 {
- return src
- }
- return src[:1]
-}
-
-// IsValid implements Encoding interface.
-func (e *encodingLatin1) IsValid(src []byte) bool {
- return true
-}
-
-// Tp implements Encoding interface.
-func (e *encodingLatin1) Tp() EncodingTp {
- return EncodingTpLatin1
-}
-
-func (e *encodingLatin1) Transform(dest *bytes.Buffer, src []byte, op Op) ([]byte, error) {
- return src, nil
-}
diff --git a/parser/mysql/charset.go b/parser/mysql/charset.go
@@ -574,6 +574,7 @@ var CollationNames = map[string]uint8{
 const (
  UTF8Charset = "utf8"
  UTF8MB4Charset = "utf8mb4"
+ Latin1Charset = "latin1"
  DefaultCharset = UTF8MB4Charset
  // DefaultCollationID is utf8mb4_bin(46)
  DefaultCollationID = 46
@@ -592,9 +593,9 @@ const (
  MaxBytesOfCharacter = 4
 )
 
-// IsUTF8Charset checks if charset is utf8 or utf8mb4
+// IsUTF8Charset checks if charset is utf8, utf8mb4 or latin1
 func IsUTF8Charset(charset string) bool {
- return charset == UTF8Charset || charset == UTF8MB4Charset
+ return charset == UTF8Charset || charset == UTF8MB4Charset || charset == Latin1Charset
 }
 
 // RangeGraph defines valid unicode characters to use in column names. It strictly follows MySQL's definition.

diff --git a/util/collate/collate.go b/util/collate/collate.go
@@ -98,7 +98,7 @@ func NewCollationEnabled() bool {
 func CompatibleCollate(collate1, collate2 string) bool {
  if (collate1 == "utf8mb4_general_ci" || collate1 == "utf8_general_ci") && (collate2 == "utf8mb4_general_ci" || collate2 == "utf8_general_ci") {
  return true
- } else if (collate1 == "utf8mb4_bin" || collate1 == "utf8_bin") && (collate2 == "utf8mb4_bin" || collate2 == "utf8_bin") {
+ } else if (collate1 == "utf8mb4_bin" || collate1 == "utf8_bin" || collate1 == "latin1_bin") && (collate2 == "utf8mb4_bin" || collate2 == "utf8_bin") {
  return true
  } else if (collate1 == "utf8mb4_unicode_ci" || collate1 == "utf8_unicode_ci") && (collate2 == "utf8mb4_unicode_ci" || collate2 == "utf8_unicode_ci") {
  return true