table: check non-BMP characters and return error when the charset is…

… utf8 and sql mode is strict mode (#8738) (#8754)
pingcap · Dec 19, 2018 · 122d26b · 122d26b
1 parent 79180dc
commit 122d26b
Show file tree

Hide file tree

Showing 2 changed files with 41 additions and 7 deletions.
diff --git a/executor/statement_context_test.go b/executor/statement_context_test.go
@@ -81,4 +81,24 @@ func (s *testSuite) TestStatementContext(c *C) {
 	tk.MustExec("set @@tidb_skip_utf8_check = '0'")
 	runeErrStr := string(utf8.RuneError)
 	tk.MustExec(fmt.Sprintf("insert sc2 values ('%s')", runeErrStr))
+
+	// Test non-BMP characters.
+	tk.MustExec(nonStrictModeSQL)
+	tk.MustExec("drop table if exists t1")
+	tk.MustExec("create table t1(a varchar(100) charset utf8);")
+	defer tk.MustExec("drop table if exists t1")
+	tk.MustExec("insert t1 values (unhex('f09f8c80'))")
+	c.Assert(tk.Se.GetSessionVars().StmtCtx.WarningCount(), Greater, uint16(0))
+	tk.MustQuery("select * from t1").Check(testkit.Rows(""))
+	tk.MustExec("insert t1 values (unhex('4040f09f8c80'))")
+	c.Assert(tk.Se.GetSessionVars().StmtCtx.WarningCount(), Greater, uint16(0))
+	tk.MustQuery("select * from t1").Check(testkit.Rows("", "@@"))
+	tk.MustQuery("select length(a) from t1").Check(testkit.Rows("0", "2"))
+	tk.MustExec(strictModeSQL)
+	_, err = tk.Exec("insert t1 values (unhex('f09f8c80'))")
+	c.Assert(err, NotNil)
+	c.Assert(terror.ErrorEqual(err, table.ErrTruncateWrongValue), IsTrue, Commentf("err %v", err))
+	_, err = tk.Exec("insert t1 values (unhex('F0A48BAE'))")
+	c.Assert(err, NotNil)
+	c.Assert(terror.ErrorEqual(err, table.ErrTruncateWrongValue), IsTrue, Commentf("err %v", err))
 }
diff --git a/table/column.go b/table/column.go
@@ -145,6 +145,16 @@ func CastValues(ctx sessionctx.Context, rec []types.Datum, cols []*Column) (err
 	return nil
 }
 
+func handleWrongUtf8Value(ctx sessionctx.Context, col *model.ColumnInfo, casted *types.Datum, str string, i int) (types.Datum, error) {
+	sc := ctx.GetSessionVars().StmtCtx
+	err := ErrTruncateWrongValue.FastGen("incorrect utf8 value %x(%s) for column %s", casted.GetBytes(), str, col.Name)
+	log.Errorf("con:%d %v", ctx.GetSessionVars().ConnectionID, err)
+	// Truncate to valid utf8 string.
+	truncateVal := types.NewStringDatum(str[:i])
+	err = sc.HandleTruncate(err)
+	return truncateVal, err
+}
+
 // CastValue casts a value based on column type.
 func CastValue(ctx sessionctx.Context, val types.Datum, col *model.ColumnInfo) (casted types.Datum, err error) {
 	sc := ctx.GetSessionVars().StmtCtx
@@ -166,18 +176,22 @@ func CastValue(ctx sessionctx.Context, val types.Datum, col *model.ColumnInfo) (
 		return casted, nil
 	}
 	str := casted.GetString()
-	for i, r := range str {
-		if r == utf8.RuneError {
+	utf8Charset := col.Charset == mysql.UTF8Charset
+	for i, w := 0, 0; i < len(str); i += w {
+		runeValue, width := utf8.DecodeRuneInString(str[i:])
+		if runeValue == utf8.RuneError {
 			if strings.HasPrefix(str[i:], string(utf8.RuneError)) {
+				w = width
 				continue
 			}
-			err = ErrTruncateWrongValue.FastGen("incorrect utf8 value %x(%s) for column %s", casted.GetBytes(), str, col.Name)
-			log.Errorf("con:%d %v", ctx.GetSessionVars().ConnectionID, err)
-			// Truncate to valid utf8 string.
-			casted = types.NewStringDatum(str[:i])
-			err = sc.HandleTruncate(err)
+			casted, err = handleWrongUtf8Value(ctx, col, &casted, str, i)
+			break
+		} else if width > 3 && utf8Charset {
+			// Handle non-BMP characters.
+			casted, err = handleWrongUtf8Value(ctx, col, &casted, str, i)
 			break
 		}
+		w = width
 	}
 
 	return casted, errors.Trace(err)