pyATS runtime.synchro.dict not updating nested list during Easypy run

[shubhanshu-mishra]

Hi all,

I am working on a pyATS/Easypy test automation project where I maintain a shared dictionary across multiple test cases using runtime.synchro.dict.

Here is my setup:
# Initialize shared dictionary in main script
shared_dict = runtime.synchro.dict({
"DUT_IP": ["device1", "device2"],
"EXECUTION_FAILURE": False,
"DEVICE_NAME": [],
"TOTAL_PASSED": 0,
"TOTAL_FAILED": 0,
"TOTAL_TEST_CASES": 0,
"TOTALS": { "PASSED": 0, "FAILED": 0, "ERRORED": 0, "TOTAL": 0 },
"TEST_SUITES": {}, # dict of suite_name -> counters
"SUITES": [] # list of executed suites
})

Inside a test case, I try to append the suite name to the shared list:

self.parameters["shared_dict"]["SUITES"].append("test_suite_example")
log.warning(f"Current shared_dict: {self.parameters['shared_dict']}")

Expected behavior:

• "test_suite_example" should appear in shared_dict["SUITES"].

Actual behavior:

• shared_dict["SUITES"] remains empty ([]) even after append().

• Other keys in shared_dict update correctly.

My questions:

1. Is this a limitation of runtime.synchro.dict regarding nested objects like lists or dicts?

2. What is the correct way to modify nested lists inside a runtime.synchro.dict during test execution?

3. Should I switch to using a DB or a JSON dump to persist results across test cases, or can this be safely done in runtime.synchro.dict?

Environment:

• pyATS 24.7

• Python 3.10

• Easypy job with multiple parallel test executions

Any guidance or best practices will be appreciated!

[bigevilbeard · Accepted solution]

I do not think yuou have a limitation here,  because self.parameters["shared_dict"]["SUITES"] is a proxy object that does not automatically synchronize with the underlying dict when you call a method like append. I think the workaround is to treat your nested mutable objects as immutable from the perspective of the synchro dict - get them, modify a copy, and reassign the result.

For your use case here, by using runtime.synchro.dict with the reassignment method is great, so i see no need to switch to a db or a json dump unless you have a requirement to persist the data beyond the single easypy job execution.

Hope this helps.