*: refactoring the code of batchChecker

[tiancaiamao]

batchChecker is difficult to maintain, we should get rid of it.
In this commit I catch the BatchGet result into the snapshot, in this way we can
achieve the same goal as the batchChecker

What problem does this PR solve?

The batchChecker is difficult to maintain. It's fragile, full of bugs, we should get rid of it.

CREATE TABLE `t` (
  `a` int(11) NOT NULL,
  `b` int(11) NOT NULL,
  `c` int(11) DEFAULT NULL,
  `d` int(11) DEFAULT NULL,
  `e` int(11) DEFAULT NULL,
  PRIMARY KEY (`a`,`b`),
  UNIQUE KEY `b` (`b`,`c`,`d`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_bin
PARTITION BY RANGE ( `b` ) (
  PARTITION p0 VALUES LESS THAN (4),
  PARTITION p1 VALUES LESS THAN (7),
  PARTITION p2 VALUES LESS THAN (11)
) ;

insert into t values (1,2,3,4,5);
insert into t values (1,2,3,4,5),(6,2,3,4,6) on duplicate key update e = e + values(e);

Expect:

mysql> select * from t;
+---+---+------+------+------+
| a | b | c    | d    | e    |
+---+---+------+------+------+
| 1 | 2 |    3 |    4 |   16 |
+---+---+------+------+------+
1 row in set (0.00 sec)

Get:

mysql> select * from t;
+---+---+------+------+------+
| a | b | c    | d    | e    |
+---+---+------+------+------+
| 1 | 2 |    3 |    4 |   11 |
+---+---+------+------+------+
1 row in set (0.00 sec)

What is changed and how it works?

The first step: do not use it.
The second step: clean up the code.

In this commit, I re-implement the insert ... on duplicate key update ... without using the batchChecker.
And I also implement a cache mechanism in the snapshot and the BatchGet API, so the performance should not change.

Check List

Tests

• Unit test (WIP)

Side effects

• Possible performance regression

Related changes

• Need to cherry-pick to the release branch

Release note

• Write release note for bug-fix or new feature.

Fix bugs of insert ... on duplicate update on the partitioned table

[tiancaiamao]

PTAL @jackysp

[tiancaiamao]

I pushed to the wrong branch.
See this one #12108