@OptimisticLocking(DIRTY) leads to wrong query during delete of circular reference

[mickroll]

Describe the bug

Having a circular reference between two Entities and deleting one of them with cascading enabled, the final delete query fails, if the "root"-Entity uses @OptimisticLocking(DIRTY).

Expected behavior

The following queries are generated and executed:

update `example_entity` set `last_log_entry_id`=null where `id`=?
delete from `log_entry` where `id`=?
delete from `example_entity` where `id`=? and `last_log_entry_id` IS NULL

All query-parameters are filled with non-null, existing values.
Both deletes are successful.

Actual behavior

The following queries are generated and executed:

update `example_entity` set `last_log_entry_id`=null where `id`=?
delete from `log_entry` where `id`=?
delete from `example_entity` where `id`=? and `last_log_entry_id`=?

All query-parameters are filled with non-null, existing values.
The last query is wrong and references old state that was changed in the first query.
This leads to the following error:

javax.persistence.OptimisticLockException: Batch update returned unexpected row count from update [0]; actual row count: 0; expected: 1

How to Reproduce?

see https://github.com/mickroll/quarkus-demonstrator/tree/optimistic-locking-circular-reference-delete
Demonstrator uses Quarkus 2.6.2.Final, but the same Bug also appears on Quarkus 2.8.0.Final.

Output of uname -a or ver

Windows 10

Output of java -version

openjdk version "11.0.12"

GraalVM version (if different from Java)

No response

Quarkus version or git rev

2.6.2.Final / 2.8.0.Final

Build tool (ie. output of mvnw --version or gradlew --version)

Apache Maven 3.6.3

Additional information

No response

[quarkus-bot]

/cc @Sanne, @gsmet, @yrodiere

[Sanne]

hi @mickroll - thanks for the report. May I assume I can ask the Hibernate ORM team to have a look, or is there a specific reason to have reported it to Quarkus?

[mickroll]

Now that you are asking, yes this is hibernate-related. Please feel free to forward.

[famod]

Hey @Sanne, is there a HHH issue already?

[mickroll]

FYI @Sanne i just reported this as a hibernate issue: https://hibernate.atlassian.net/browse/HHH-15218

[famod]

Hi @Sanne @yrodiere, any chance to fix this upstream? 🙂

[Sanne]

sorry for the lack of communications.

I've pinged the ORM team to make sure it gets triaged, seems to have fallen through the cracks. Not sure if we'll have the bandwith to fix it in 5, but let's see what the team thinks.

[Sanne]

this was fixed by @dreab8 in Hibernate ORM, so we'll have a fix when merging #26625

[famod]

Awesome, thanks @dreab8 & @Sanne!

[mickroll]

this was fixed by @dreab8 in Hibernate ORM, so we'll have a fix when merging #26625

Sounds great 🚀