Open burnettk opened 6 months ago
The theory here is that with celery the mi child tasks are created and then saved to the db and the pi is queued to run again. This causes the pi to be loaded from the db which loads the task based on guid order which is probably not the order spiff originally had them in. This means we run them out of order and spiff tries to test the condition in a different order which causes this issue.
Is this also fixed now? It worked today.
It seems doubtful it's fixed. We think it's random. It may work sometimes and it may not others.
Yeah, it's not fixed yet.
And this one doesn't seem to work as expected. It should ouput only up to 4 because of the completion condition. But it doesn't work like that.
https://dev.mod.spiff.status.im/process-models/misc:qa:bpmn-model-testing:unit-test-multi-instance-parallel-loop-cardinality https://dev.mod.spiff.status.im/i/36220 This is how the output looks in dev.app for the same model - https://dev.app.spiff.status.im/i/3599
see discussion between jasquant and elizabeth on #1362