Job session lingers around in state after job is finished

Issue description

This is a minor technical issue, which potentially could cause harm, but has not in the last years.

Developer comments

This issue surfaced when FastDelete was moved from static to instance recently. This caused assert 510 in TestTopincsJob->test_chore to fail. Assert 510 checks whether the job file is no longer in the file system after the chore was deleted. The deletion of the job file is handled by a trigger and triggers were not enabled in the test, because jobs use a dedicated session where triggers are always enabled. Since the test passed up to now, there must be a problem in TopincsState, since the chore is deleted in the test not in the job. So the trigger execution is initiated outside the job, therefore it is necessary to enable triggers in the session of the test.

Here more formally: * Test T1 is executed with session S1. S1 has triggers disabled. * In T1 a job J1 for chore C1 is run in R1. * In R1 a session S2 is created. S2 has triggers always enabled. * R1 finished. * T1 deletes the chore. * Up to the change: assert 510 passes. Thus it must have used S2, which is the bug at hand. * After the change: assert 510 fails. Since it correctly uses S1.

Reporting date

2026-10-28

Reported by

Robert Cerny

Resolving date

2026-10-28

Resolved by

Robert Cerny

Resolved in version

Topincs 10.1.0 – next release

Work sessions

Start	2026-10-28T08:30:18
End	2026-10-28T09:39:51
Participant	Robert Cerny