500 Error when creating a test run / adding a test case to a run

[avishnyak]

HINT: try to reproduce bugs and errors at https://public.tenant.kiwitcms.org!
This is always running the latest version and will automatically submit
traceback and debugging information to us!

Description of problem

I get a 500 error whenever I attempt to create a test run or add a test case to an existing test run.

Version or commit hash (if applicable)

8.7-Enterprise (hosted)

Steps to Reproduce

1. Open a test plan (e.g. https://atlaslifttech.tenant.kiwitcms.org/plan/6/dashboard-rerporting-regression-suite#testcases)
2. Click Run -> Create new Test Run
3. Click save (note the test cases are all selected by default)

Actual results

500 Error

Expected results

Navigate to the Run page and be able to execute test cases.

Additional info (Python traceback, logs, etc.)

When attempting to add test cases to the run manually after the fact, I still get a 500 error when updating the Run.
See: https://atlaslifttech.tenant.kiwitcms.org/runs/5/assigncase/

[atodorov]

I get a 500 error whenever I attempt to create a test run or add a test case to an existing test run.

We're seeing the errors reported in Sentry and currently investigating. Will update with more info.

[atodorov]

Steps to Reproduce

1. Open a test plan (e.g. https://atlaslifttech.tenant.kiwitcms.org/plan/6/dashboard-rerporting-regression-suite#testcases)

2. Click Run -> Create new Test Run

3. Click save (note the test cases are all selected by default)

Using these steps to reproduce on our own private tenant I was able to create a new TR.

Actual results

500 Error

When attempting to add test cases to the run manually after the fact, I still get a 500 error when updating the Run.
See: https://atlaslifttech.tenant.kiwitcms.org/runs/5/assigncase/

Note: the tracebacks I'm seeing in Sentry all refer to /runs/<pk>/assigncase/ page which is different from the above steps to reproduce. From what I can see there is:

TP-6 which includes a single test case, TC-1: User is able to log in with valid credentials. There are 0 TestRuns created on your tenant.

If there are 0 test runs then accessing /runs/5/assigncase/ will result in 500 - this is DoesNotExist error on the backend because such object doesn't exist.

There are historical versions for TR-6 and TR-5. Both test runs have been deleted since.

There are historical TestExecution records only for TR-1 and TR-2, both for TC-1 which have been deleted since. So it looks like at some point creating a new test run was possible.

@avishnyak can you provide us with more detailed steps to reproduce? Ideally record a short video so we can see what are you doing and send it over to the support email address.

[atodorov]

Note. I see you've also removed the following TestExecutionStatus objects:

Deleted “ERROR.”
Deleted “IDLE.”
Deleted “RUNNING.”
Deleted “PAUSED.”

Creation of a new TestRun page is looking for the IDLE status, more precisely an ID of 1 which didn't exist and it failed.

I've added back the IDLE status with an ID=1 and will leave this open b/c the underlying code needs to be refactored a bit.

Also by deleting some of the statuses you've removed all statuses with weight=0, that is neutral ones, e.g. not positive, nor negative:
https://kiwitcms.readthedocs.io/en/latest/admin.html?highlight=status#test-execution-statuses

this is also something that the backend should be aware of.

[atodorov]

Implementation notes:

for TestRun.add_case_run:

     94     def add_case_run(self, case, status=1, assignee=None,
     95                      case_text_version=None, build=None,
     96                      sortkey=0):

The status=1 default value is wrong. It should be TestExecutionStatus.objects.filter(weight=0).first() or handled internally withing the method. TestExecutionStatusAdmin should check before allowing deletion that at least 1 positive, negative and neutral statuses exist and show an appropriate warning message if this is going to break if the user proceeds with the deletion.

[avishnyak]

Thank you. This fixed my issue.