fix(aci): set workflow in job when evaluating action filters

[cathteng]

I wrote a test and found that

1. In workflow.evaluate_trigger_conditions(job) we attach the workflow to the job
2. If we have a set of workflows we're evaluating, the last workflow we evaluate remains attached to the job
3. Next in evaluate_workflows_action_filters we end up passing that last workflow to every single filter/IF data condition group we evaluate, even if the filter/IF DCG is not actually attached to that last workflow

We should be correctly setting the workflow for each data condition group we evaluate.

(Wrote a test with a condition that should not be an allowable filter condition for demonstration purposes only)

[cathteng]

maybe i don't need to pop, but not popping caused this problem in the first place

[cathteng]

TIL dictionaries are pass by reference in Python which is why this happens

[iamrajjoshi]

lgtm, good find

[saponifi3d]

pls hold on merging -- i'd like to take a quick look.

[codecov]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
24416	1	24415	299

View the top 1 failed test(s) by shortest run time

tests.sentry.workflow_engine.processors.test_workflow.TestProcessWorkflows::test_error_event__logger

Stack Traces | 4.32s run time

#x1B[1m#x1B[.../workflow_engine/processors/test_workflow.py#x1B[0m:100: in test_error_event__logger
    mock_logger.info.assert_called_with(
#x1B[1m#x1B[.../hostedtoolcache/Python/3.13.1.../x64/lib/python3.13/unittest/mock.py#x1B[0m:977: in assert_called_with
    raise AssertionError(_error_message()) from cause
#x1B[1m#x1B[31mE   AssertionError: expected call not found.#x1B[0m
#x1B[1m#x1B[31mE   Expected: info('workflow_engine.process_workflows.fired_workflow', extra={'workflow_id': 17, 'rule_id': 698, 'payload': {'event': <sentry.eventstore.models.GroupEvent object at 0x7f91419232f0>, 'group_state': {'id': 1, 'is_new': False, 'is_regression': True, 'is_new_group_environment': False}, 'workflow': <Workflow at 0x7f914025f350: id=17, name='error_workflow', organization_id=4555749194989572>}, 'group_id': 931, 'event_id': '550e7bf3e1da42f28b322f8bdecd2936'})#x1B[0m
#x1B[1m#x1B[31mE     Actual: info('workflow_engine.process_workflows.fired_workflow', extra={'workflow_id': 17, 'rule_id': 698, 'payload': {'event': <sentry.eventstore.models.GroupEvent object at 0x7f91419232f0>, 'group_state': {'id': 1, 'is_new': False, 'is_regression': True, 'is_new_group_environment': False}}, 'group_id': 931, 'event_id': '550e7bf3e1da42f28b322f8bdecd2936'})#x1B[0m

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}

[saponifi3d]

we chatted a little in slack, but i'm not 100% sold on this as the fix. it feels like we're fixing a symptom rather than treating the issue.

i'd like to take a closer look at how we're using the job here (maybe we can just grab the environment? 🤔) and think a bit more about why we need to do the pop() -- that is a smell that there's likely a bug higher up in the code and it's causing issues here.

hopefully i can get some keyboard time tomorrow and i might be able to take a look around to see if there's a more holistic fix.

[wedamija]

imo the problem here is that we're mutating WorkflowJob. I think it should actually be a frozen dataclass. If we need to copy it for multiple workflows we can do that so that the original object isn't modified

[cathteng]

i will try this again