project-chip
[Bug] [Frontend] TH Test run UI TC mismatch
Describe the bug
The test run UI in the left panel shows a Test Case that is not the same that is running according to log panel in the right
Important: TH UI shows the tests in the left panel according to the order presented at test selection screen, but backend returns a specific test list order in the following endpoint responses: POST /api/v1/test_run_executions/ and POST /api/v1/test_run_executions/{id}/start.
In addition, the endpoint /{id}/applicable_test_cases, is also returning the test cases in a different order than the displayed in Test selection screen.
Steps to reproduce the behavior
- Select project
- Click + Add Test
- Click Start
Expected behavior
Test Run UI in the left panel must reflect the log panel in the right, so it should consider the backend response for POST /api/v1/test_run_executions/{id}/start and also POST /api/v1/test_run_executions/{id}/repeat
Log files
No response
PICS file
No response
Screenshots
No response
Environment
No response
Additional Information
No response
@rquidute
From trying to reproduce the issue I noticed that the scenario in the screenshot can happen if we expand one of the upcoming tests (tests not in execution yet) making it appear like a mismatch.
You can notice from the screenshot that the left hand side isn't displaying a test starting or in execution (there are not icons indicating status or progress).
When there's a test under execution it shows the progress icons to the left.
Might this be a false positive?
If not can you please provide a short video or more details showing the behavior described?
Or might this be a regression? was this not happening on earlier versions of the TH?
Hi @raul-marquez-csa, We are right, the test under execution is not displayed in the screenshot and there is only one progress icons that is present for the current test in progress. I think, since the first test displayed in the left panel is not the first to be executed, might lead user to think it is the wrong test case. The mandatories test cases should be the first ones to be executed.
I'm attaching a video where you can see the whole process. Please also notice the screen right before the test execution (The screen where you are able to select the test cases), the right panel shows the test case in the same order as the execution screen.
https://github.com/user-attachments/assets/7b3b407c-58d5-413f-9dc3-6a10a22a8d0d
Blocked for now
High complexity, many moving parts, and potential major re-design/refactor of both frontend and backend, deeper discussion needed.
I see no easy or straightforward way to fix this, latest progress and attempts are in my previous comment (PR links).
Any other approach or pointers in the correct direction appreciated.
Hi @raul-marquez-csa,
I can try to help you on this issue. I'll take a look on the PRs you've created and if I can think on another approach.
Can you give a summary of complexity you discovered?
I can try to help you on this issue. I'll take a look on the PRs you've created and if I can think on another approach.
Can you give a summary of complexity you discovered?
In general, several places in the backend where the test list is fetched, sorted, re-sorted, then in the frontend as well going through loops and hoops.
What I think is the best case scenario is to have a single source of truth coming from the backend all the way to the UI component in the front end, but as it stands, at least when I looked at it, too many inflection points