json_quote return different result

[xiongjiwei]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. Minimal reproduce step (Required)

create table t (c varchar(20) collate utf8mb4_bin) charset utf8mb4;
insert into t values ('\\');
insert into t values (X'0C');
insert into t values ('"');
insert into t values ('\a');
insert into t values ('\b');
insert into t values ('\t');
insert into t values ('\n');
insert into t values ('\r');
insert into t values (X'10');
select json_quote(group_concat(c order by c, hex(c))) from t;

2. What did you expect to see? (Required)

"\b,\t,\n,\f,\r,\u0010,\",\\,a"

3. What did you see instead (Required)

"\b,\t,\n,\f,\r,\x10,\",\\,a"

4. What is your TiDB version? (Required)

[windtalker]

minimum reproduce:

create table t (c varchar(20) collate utf8mb4_bin) charset utf8mb4;
insert into t values (X'10');
--- in TiDB
mysql> select json_quote(c) from t;
+---------------+
| json_quote(c) |
+---------------+
| "\x10"        |
+---------------+
1 row in set (0.00 sec)
--- in MySQL 8.0.25
mysql> select json_quote(c) from t;
+---------------+
| json_quote(c) |
+---------------+
| "\u0010"      |
+---------------+

[xiongjiwei]

it only occurs if string hex < 0x20, MySQL will convert these characters to \u00xx, tidb is \xxx. I will change to minor

[dveeden]

Another simple test case for this that doesn't reply on a table:

SELECT JSON_QUOTE(CONVERT(0x10 USING utf8mb4));

[dveeden]

Another test:

WITH RECURSIVE nr(n) AS (
    SELECT 1 n
   UNION ALL
    SELECT n+1 FROM nr WHERE n<25
)
SELECT JSON_QUOTE(CONVERT(UNHEX(n) USING utf8mb4)) FROM nr;

mysql-8.0.11-TiDB-v8.2.0-alpha-313-gf35bab8191-dirty> WITH RECURSIVE nr(n) AS (SELECT 1 n UNION ALL SELECT n+1 FROM nr WHERE n<25) SELECT JSON_QUOTE(CONVERT(UNHEX(n) USING utf8mb4)) FROM nr;
+---------------------------------------------+
| JSON_QUOTE(CONVERT(UNHEX(n) USING utf8mb4)) |
+---------------------------------------------+
| "\x01"                                      |
| "\x02"                                      |
| "\x03"                                      |
| "\x04"                                      |
| "\x05"                                      |
| "\x06"                                      |
| "\a"                                        |
| "\b"                                        |
| "\t"                                        |
| "\x10"                                      |
| "\x11"                                      |
| "\x12"                                      |
| "\x13"                                      |
| "\x14"                                      |
| "\x15"                                      |
| "\x16"                                      |
| "\x17"                                      |
| "\x18"                                      |
| "\x19"                                      |
| " "                                         |
| "!"                                         |
| "\""                                        |
| "#"                                         |
| "$"                                         |
| "%"                                         |
+---------------------------------------------+
25 rows in set (0.00 sec)

mysql-8.4.0> WITH RECURSIVE nr(n) AS (SELECT 1 n UNION ALL SELECT n+1 FROM nr WHERE n<25) SELECT JSON_QUOTE(CONVERT(UNHEX(n) USING utf8mb4)) FROM nr;
+---------------------------------------------+
| JSON_QUOTE(CONVERT(UNHEX(n) USING utf8mb4)) |
+---------------------------------------------+
| "\u0001"                                    |
| "\u0002"                                    |
| "\u0003"                                    |
| "\u0004"                                    |
| "\u0005"                                    |
| "\u0006"                                    |
| "\u0007"                                    |
| "\b"                                        |
| "\t"                                        |
| "\u0010"                                    |
| "\u0011"                                    |
| "\u0012"                                    |
| "\u0013"                                    |
| "\u0014"                                    |
| "\u0015"                                    |
| "\u0016"                                    |
| "\u0017"                                    |
| "\u0018"                                    |
| "\u0019"                                    |
| " "                                         |
| "!"                                         |
| "\""                                        |
| "#"                                         |
| "$"                                         |
| "%"                                         |
+---------------------------------------------+
25 rows in set (0.00 sec)

[dveeden]

Are we using strconv.Quote() instead of relying on encoding/json?

package main

import (
	"encoding/json"
	"fmt"
	"strconv"
)

func main() {
	for i := range 38 {
		if i == 0 {
			continue
		}
		j, _ := json.Marshal(string(i))
		fmt.Printf(
			"i=%d\tquote=%s\tj=%s\n",
			i,
			strconv.Quote(string(i)),
			j,
		)
	}
}

i=1	quote="\x01"	j="\u0001"
i=2	quote="\x02"	j="\u0002"
i=3	quote="\x03"	j="\u0003"
i=4	quote="\x04"	j="\u0004"
i=5	quote="\x05"	j="\u0005"
i=6	quote="\x06"	j="\u0006"
i=7	quote="\a"	j="\u0007"
i=8	quote="\b"	j="\b"
i=9	quote="\t"	j="\t"
i=10	quote="\n"	j="\n"
i=11	quote="\v"	j="\u000b"
i=12	quote="\f"	j="\f"
i=13	quote="\r"	j="\r"
i=14	quote="\x0e"	j="\u000e"
i=15	quote="\x0f"	j="\u000f"
i=16	quote="\x10"	j="\u0010"
i=17	quote="\x11"	j="\u0011"
i=18	quote="\x12"	j="\u0012"
i=19	quote="\x13"	j="\u0013"
i=20	quote="\x14"	j="\u0014"
i=21	quote="\x15"	j="\u0015"
i=22	quote="\x16"	j="\u0016"
i=23	quote="\x17"	j="\u0017"
i=24	quote="\x18"	j="\u0018"
i=25	quote="\x19"	j="\u0019"
i=26	quote="\x1a"	j="\u001a"
i=27	quote="\x1b"	j="\u001b"
i=28	quote="\x1c"	j="\u001c"
i=29	quote="\x1d"	j="\u001d"
i=30	quote="\x1e"	j="\u001e"
i=31	quote="\x1f"	j="\u001f"
i=32	quote=" "	j=" "
i=33	quote="!"	j="!"
i=34	quote="\""	j="\""
i=35	quote="#"	j="#"
i=36	quote="$"	j="$"
i=37	quote="%"	j="%"

https://go.dev/play/p/wwO4KFNnOIK