Remove SqlQueryScheduler#abort method

[arhimondr]

Description

SqlQueryScheduler registers a query state listener internally to
terminate query scheduling when the query is cancelled, failed or
finished early

Is this change a fix, improvement, new feature, refactoring, or other?

Refactor

Is this a change to the core query engine, a connector, client library, or the SPI interfaces? (be specific)

Core engine

How would you describe this change to a non-technical end user or system administrator?

Non user visible

Related issues, pull requests, and links

-

Documentation

(X) No documentation is needed.
( ) Sufficient documentation is included in this PR.
( ) Documentation PR is available with #prnumber.
( ) Documentation issue #issuenumber is filed, and can be handled later.

Release notes

(X) No release notes entries required.
( ) Release notes entries required with the following suggested text:

# Section
* Fix some things. ({issue}`issuenumber`)

[sopel39]

Shouldn't queryScheduler.set(null); be called anyway (e.g. to free resources)?

[arhimondr]

The queryScheduler variable is nullified in buildQueryInfo after building a final query info. It is currently not nullified in case of a failure or completion anywhere else, thus it seemed to be weird to nullify it here.

[sopel39]

It is currently not nullified in case of a failure or completion anywhere else, thus it seemed to be weird to nullify it here.

So was this a race between buildQueryInfo and nullification here?

[arhimondr]

I doubt there any. It looked very weird. Later on there's another check where the scheduler is not getting nullified: https://github.com/trinodb/trino/blob/master/core/trino-main/src/main/java/io/trino/execution/SqlQueryExecution.java#L423.

[sopel39]

Do we have test for aborting? Can it be that this listener covers some flow that is not covered by SqlQueryScheduler?

[arhimondr]

Today the SqlQueryScheduler is responsible of attaching the listeners that terminate the scheduling process upon query completion (due to a failure / cancellation or early finish): https://github.com/trinodb/trino/blob/master/core/trino-main/src/main/java/io/trino/execution/scheduler/SqlQueryScheduler.java#L303. SqlQueryScheduler attaches the listener before the scheduling process starts and uses an internal lock to prevent race conditions.

These listeners do exactly what the abort method does. Having an external listener calling the abort method is not necessary and might be confusing for the readers.

[sopel39]

I would like another set of eyes to look at it, e.g: @losipiuk or @martint

[sopel39]

It is currently not nullified in case of a failure or completion anywhere else, thus it seemed to be weird to nullify it here.

So was this a race between buildQueryInfo and nullification here?