From 01dac873baa102c4a8ae9a99b7253d03f9044c90 Mon Sep 17 00:00:00 2001
From: Shen Li <shenli3514@gmail.com>
Date: Sat, 31 Oct 2015 22:37:27 +0800
Subject: [PATCH] *: Address ci problem

---
 ast/expressions.go   |  4 +++-
 ast/parser/parser.y  | 18 ++++++++++++++++++
 ast/parser/scanner.l | 19 ++++++++++++++++++-
 3 files changed, 39 insertions(+), 2 deletions(-)

diff --git a/ast/expressions.go b/ast/expressions.go
index 6c09874750..7c82ed2d2a 100644
--- a/ast/expressions.go
+++ b/ast/expressions.go
@@ -56,7 +56,7 @@ type ValueExpr struct {
 // NewValueExpr creates a ValueExpr with value, and sets default field type.
 func NewValueExpr(value interface{}) *ValueExpr {
 	ve := &ValueExpr{}
-	ve.Data = value
+	ve.Data = types.RawData(value)
 	// TODO: make it more precise.
 	switch value.(type) {
 	case nil:
@@ -82,6 +82,8 @@ func NewValueExpr(value interface{}) *ValueExpr {
 		ve.Type = types.NewFieldType(mysql.TypeVarchar)
 		ve.Type.Charset = "binary"
 		ve.Type.Collate = "binary"
+	case *types.DataItem:
+		ve.Type = value.(*types.DataItem).Type
 	default:
 		panic(fmt.Sprintf("illegal literal value type:%T", value))
 	}
diff --git a/ast/parser/parser.y b/ast/parser/parser.y
index 9c02a8e4c5..127a3b965d 100644
--- a/ast/parser/parser.y
+++ b/ast/parser/parser.y
@@ -239,6 +239,7 @@ import (
 	trim		"TRIM"
 	trueKwd		"true"
 	truncate	"TRUNCATE"
+	underscoreCS	"UNDERSCORE_CHARSET"
 	unknown 	"UNKNOWN"
 	union		"UNION"
 	unique		"UNIQUE"
@@ -1796,6 +1797,23 @@ Literal:
 |	floatLit
 |	intLit
 |	stringLit
+|	"UNDERSCORE_CHARSET" stringLit
+	{
+		// See: https://dev.mysql.com/doc/refman/5.7/en/charset-literal.html
+		tp := types.NewFieldType(mysql.TypeString)
+		tp.Charset = $1.(string)
+		co, err := charset.GetDefaultCollation(tp.Charset)
+		if err != nil {
+			l := yylex.(*lexer)
+			l.errf("Get collation error for charset: %s", tp.Charset)
+			return 1
+		}
+		tp.Collate = co
+		$$ = &types.DataItem{
+			Type: tp,
+			Data: $2.(string),
+		}
+	}
 |	hexLit
 |	bitLit
 
diff --git a/ast/parser/scanner.l b/ast/parser/scanner.l
index 048817d1b9..adaa662651 100644
--- a/ast/parser/scanner.l
+++ b/ast/parser/scanner.l
@@ -27,6 +27,7 @@ import (
 
 	"github.com/pingcap/tidb/ast"
 	"github.com/pingcap/tidb/mysql"
+	"github.com/pingcap/tidb/util/charset"
 	"github.com/pingcap/tidb/util/stringutil"
 )
 
@@ -1010,7 +1011,7 @@ year_month		{y}{e}{a}{r}_{m}{o}{n}{t}{h}
 			return integerType
 
 {ident}			lval.item = string(l.val)
-			return identifier
+			return l.handleIdent(lval)
 
 .			return c0
 
@@ -1101,3 +1102,19 @@ func (l *lexer) bit(lval *yySymType) int {
 	lval.item = b
 	return bitLit
 }
+
+func (l *lexer) handleIdent(lval *yySymType) int {
+	s := lval.item.(string)
+	// A character string literal may have an optional character set introducer and COLLATE clause:
+	// [_charset_name]'string' [COLLATE collation_name]
+	// See: https://dev.mysql.com/doc/refman/5.7/en/charset-literal.html
+	if !strings.HasPrefix(s, "_") {
+		return identifier	
+	}
+	cs, _, err := charset.GetCharsetInfo(s[1:])
+	if err != nil {
+		return identifier	
+	}
+	lval.item = cs
+	return underscoreCS
+}