Use sequences of MariaDB, don't treat it as MySQL

[Master-Code-Programmer]

Expected Behavior
For MariaDB its, now for 4 years existing, native sequences get used, not the simulated sequences like MySQL needs them. This is a performance advantage of MariaDB against MySQL, since the code to simulate sequences isn't that simple. It would also be easier to maintain.
Link: https://mariadb.com/kb/en/sequence-overview/
Command to get next value: NEXTVAL(sequence_name).

Current Behavior
Spring Batch uses MySQLs sequence simulation of the MySQLMaxValueIncrementer, whose core method is 80 lines long and all but simple. But it seems it could use something like the far slimmer OracleSequenceMaxValueIncrementer, whose core method is a primitive one-liner.

Context
Ignoring this feature of MariaDB is inefficient.

[fmbenhassine]

We had a discussion about this in #3891, which depends on spring-projects/spring-framework#29306.

The addition of a dedicated incrementer for MariaDB in Spring Framework does not seem to be planned for SF 6, so I suggest we create one in Spring Batch (like we did for SQLServer) to implement this feature in SB 5, and then replace it with the one from SF when it is ready.

I will plan this for the upcoming 5.0 RC2.

[jhoeller]

@fmbenhassine We can pick this up for Spring Framework 6.0 RC3 still, so if you have a candidate implementation for Batch 5, feel free to post it on spring-projects/spring-framework#29306 (or create a separate PR for it).

[fmbenhassine]

@jhoeller That's a great news! Thank you. I will open a PR for it.