Persistence improvements using hypersistence optimizer

[pmverma]

Fixes #10650
Fixes #10653

Please note that all are verified using the prod profile only.

• Postgresql

• Mysql

• Mariadb

• Mssql

• Oracle

• Please make sure the below checklist is followed for Pull Requests.

• Travis tests are green

• Tests are added where necessary

• Documentation is added/updated where necessary

• Coding Rules & Commit Guidelines as per our CONTRIBUTING.md document are followed

[Blackdread]

For OneToOne, if MapsId is not used then on one side of the relationship (side without foreign key) will always be eager fetched, even if you put Lazy.

[pmverma]

@Blackdread Yes you are correct. If it is bidirectional then there will be no effect. But for unidirectional, this will work.

[DanielFran]

LGTM

[pmverma]

Thanks @DanielFran!

[cbornet]

There's a problem here : if we set the fetch strategy to be lazy at hibernate level, then it should be done at controller or service level. With the new behavior, if you call the REST endpoint for an entity, you don't see the related entity in the endpoint (it's null) and so it's not displayed in the front-end which is unexpected. It can even be very harmful : if you edit the entity then save without any modification, then the relationship is erased !

[pmverma]

Thanks @cbornet for describing the problem. Actually yes, I generated and tested the PR by skipping the client during generation. I will check it.

[pmverma]

For the moment, I will revert the lazy load. It is better to have a patch release for this.