Bulk Operations: ensure consistent purged state

[davelopez]

It seems directly purging datasets through the bulk operations endpoint does not ensure the dataset is also flagged as deleted for consistency. It does, checked now with a test.
Also:

• Make sure that trying to undelete a purged dataset does not have any effect.
• Purging already purged datasets will have no effect (and we shouldn't expect a history change).
• Purging a dataset_collection should purge all its contents.

[mvdbeek]

As a note,

does not ensure the dataset is also flagged as deleted for consistency.

that might happen in Celery at any point after the request. I'm not sure how you want to deal with this, but that's probably an important gotcha. Set enable_celery_tasks: true in galaxy.yml and start Galaxy with run.sh to see that behavior.

[davelopez]

that might happen in Celery at any point after the request.

is it a really really bad idea to flag the items as deleted and purged in the database before offloading the actual purge task to celery?
that way we don't even need to wait for the celery task to finish to return from Processing... status.

[mvdbeek]

is it a really really bad idea to flag the items as deleted and purged in the database before offloading the actual purge task to celery?

No, I think that would be fine, although from the backend perspective ideally all of this would happen in Celery (deleting and purging).