From d42203dde80502f1f952dc739624158b8a05a94a Mon Sep 17 00:00:00 2001
From: "endy.li" <25311962+heroicNeZha@users.noreply.github.com>
Date: Tue, 7 Dec 2021 11:05:26 +0800
Subject: [PATCH] feat - schema support chinese (#3380)

* feat - support chinese

* fix - remove debug anno

* support `chinese`

* add test cases

* fix encode bug

Co-authored-by: cpw <13495049+CPWstatic@users.noreply.github.com>
Co-authored-by: Yee <2520865+yixinglu@users.noreply.github.com>
Co-authored-by: Sophie <84560950+Sophie-Xie@users.noreply.github.com>
---
 src/parser/parser.yy                     |  3 +-
 src/parser/scanner.lex                   | 17 ++++++-
 tests/tck/features/schema/Schema.feature | 63 ++++++++++++++++++++++++
 3 files changed, 81 insertions(+), 2 deletions(-)

diff --git a/src/parser/parser.yy b/src/parser/parser.yy
index 04f1bf1699d..0ea8ede2672 100644
--- a/src/parser/parser.yy
+++ b/src/parser/parser.yy
@@ -213,7 +213,7 @@ static constexpr size_t kCommentLengthLimit = 256;
 %token <boolval> BOOL
 %token <intval> INTEGER
 %token <doubleval> DOUBLE
-%token <strval> STRING VARIABLE LABEL IPV4
+%token <strval> STRING VARIABLE LABEL IPV4 CHINESE_LABEL
 
 %type <strval> name_label unreserved_keyword predicate_name
 %type <expr> expression
@@ -406,6 +406,7 @@ static constexpr size_t kCommentLengthLimit = 256;
 
 name_label
     : LABEL { $$ = $1; }
+    | CHINESE_LABEL { $$ = $1; }
     | unreserved_keyword { $$ = $1; }
     ;
 
diff --git a/src/parser/scanner.lex b/src/parser/scanner.lex
index 007a1f93153..4fe773332d2 100644
--- a/src/parser/scanner.lex
+++ b/src/parser/scanner.lex
@@ -43,7 +43,11 @@ HEX                         ([0-9a-fA-F])
 OCT                         ([0-7])
 IP_OCTET                    ([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])
 
-
+U                           [\x80-\xbf]
+U2                          [\xc2-\xdf]
+U3                          [\xe0-\xef]
+U4                          [\xf0-\xf4]
+CHINESE_LABEL               ({U2}{U}|{U3}{U}{U}|{U4}{U}{U}{U})+
 
 %%
 
@@ -467,6 +471,17 @@ IP_OCTET                    ([0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])
                                 // Must match /* */
                                 throw GraphParser::syntax_error(*yylloc, "unterminated comment");
                             }
+\`{CHINESE_LABEL}\`         {
+                                yylval->strval = new std::string(yytext + 1, yyleng - 2);
+                                if (yylval->strval->size() > MAX_STRING) {
+                                    auto error = "Out of range of the LABEL length, "
+                                                  "the  max length of LABEL is " +
+                                                  std::to_string(MAX_STRING) + ":";
+                                    delete yylval->strval;
+                                    throw GraphParser::syntax_error(*yylloc, error);
+                                }
+                                return TokenType::CHINESE_LABEL;
+                            }
 .                           {
                                 /**
                                  * Any other unmatched byte sequences will get us here,
diff --git a/tests/tck/features/schema/Schema.feature b/tests/tck/features/schema/Schema.feature
index b203ea87b62..f30f79a6d27 100644
--- a/tests/tck/features/schema/Schema.feature
+++ b/tests/tck/features/schema/Schema.feature
@@ -769,6 +769,69 @@ Feature: Insert string vid of vertex and edge
       ALTER EDGE edge_not_null_default1 CHANGE (name FIXED_STRING(10) DEFAULT 10)
       """
     Then a ExecutionError should be raised at runtime: Invalid param!
+    # chinese tag without quote mark
+    When executing query:
+      """
+      CREATE TAG 队伍( 名字 string);
+      """
+    Then a SyntaxError should be raised at runtime:
+    # chinese tag and chinese prop
+    When executing query:
+      """
+      CREATE TAG `队伍`(`名字` string);
+      """
+    Then the execution should be successful
+    # show chinese tags
+    When executing query:
+      """
+      SHOW TAGS
+      """
+    Then the result should contain:
+      | Name   |
+      | "队伍" |
+    # alter chinese tag
+    When executing query:
+      """
+      ALTER TAG `队伍` ADD (`类别` string);
+      """
+    Then the execution should be successful
+    # desc chinese tag
+    When executing query:
+      """
+      DESCRIBE TAG `队伍`
+      """
+    Then the result should be, in any order:
+      | Field  | Type     | Null  | Default | Comment |
+      | "名字" | "string" | "YES" | EMPTY   | EMPTY   |
+      | "类别" | "string" | "YES" | EMPTY   | EMPTY   |
+    # chinese edge and chinese prop
+    When executing query:
+      """
+      CREATE EDGE `服役`();
+      """
+    Then the execution should be successful
+    # show chinese edge
+    When executing query:
+      """
+      SHOW EDGES;
+      """
+    Then the result should contain:
+      | Name   |
+      | "服役" |
+    # alter chinese edge
+    When executing query:
+      """
+      ALTER EDGE `服役` ADD (`时间` timestamp);
+      """
+    Then the execution should be successful
+    # desc chinese edge
+    When executing query:
+      """
+      DESCRIBE EDGE `服役`
+      """
+    Then the result should be, in any order:
+      | Field  | Type        | Null  | Default | Comment |
+      | "时间" | "timestamp" | "YES" | EMPTY   | EMPTY   |
     When executing query:
       """
       DROP SPACE issue2009;