action #123745
closedInvestigate failures in welcome.pm on remote_vnc_controller test suite
0%
Description
Observation¶
We have a very strange problem in Build 66.1 in the welcome.pm test module.
The test fails because no needle for 'select-product-sles' matched.
From looking at the needle there should be the "o" radio button selected but it isn't.
Further details¶
The problem happens in this assert_and_click call in welcome.pm
.
Since there are no click areas in the needle defined openQA is supposed to click in the middle of the matching area.
The autoinst-log.txt
shows this:
[2023-01-23T10:27:51.676015+01:00] [debug] [pid:17376] >>> testapi::_handle_found_needle: found welcome-before-select-product-sles-20210217, similarity 0.93 @ 359/409
[2023-01-23T10:27:51.676208+01:00] [debug] [pid:17376] tests/installation/welcome.pm:207 called testapi::assert_and_click
[2023-01-23T10:27:51.676337+01:00] [debug] [pid:17376] <<< testapi::assert_and_click(mustmatch="before-select-product-sles", timeout=30)
[2023-01-23T10:27:51.677084+01:00] [debug] [pid:17376] clicking at 454/414
[2023-01-23T10:27:51.677247+01:00] [debug] [pid:17376] tests/installation/welcome.pm:207 called testapi::assert_and_click
[2023-01-23T10:27:51.677332+01:00] [debug] [pid:17376] <<< testapi::mouse_set(x=454, y=414)
[2023-01-23T10:27:51.677821+01:00] [debug] [pid:17455] mouse_move 454, 414
[2023-01-23T10:27:51.677924+01:00] [debug] [pid:17455] send_pointer_event 0, 454, 414, 1
[2023-01-23T10:27:51.678370+01:00] [debug] [pid:17376] tests/installation/welcome.pm:207 called testapi::assert_and_click
[2023-01-23T10:27:51.678511+01:00] [debug] [pid:17376] <<< testapi::mouse_click(button="left", cursor_down="0.15")
[2023-01-23T10:27:51.678997+01:00] [debug] [pid:17455] pointer_event 1 454, 414
[2023-01-23T10:27:51.679067+01:00] [debug] [pid:17455] send_pointer_event 1, 454, 414, 1
[2023-01-23T10:27:51.829777+01:00] [debug] [pid:17455] pointer_event 0 454, 414
[2023-01-23T10:27:51.829847+01:00] [debug] [pid:17455] send_pointer_event 0, 454, 414, 1
[2023-01-23T10:27:52.830519+01:00] [debug] [pid:17376] tests/installation/welcome.pm:207 called testapi::assert_and_click
[2023-01-23T10:27:52.830796+01:00] [debug] [pid:17376] <<< testapi::mouse_set(x=1023, y=767)
[2023-01-23T10:27:52.831640+01:00] [debug] [pid:17455] mouse_move 1023, 767
[2023-01-23T10:27:52.831739+01:00] [debug] [pid:17455] send_pointer_event 0, 1023, 767, 1
[2023-01-23T10:27:52.832154+01:00] [debug] [pid:17376] tests/installation/welcome.pm:209 called testapi::assert_screen
[2023-01-23T10:27:52.832345+01:00] [debug] [pid:17376] <<< testapi::assert_screen(mustmatch="select-product-sles", timeout=30)
So technically we click at 454/414 which fits to the matching area if we take the screenshot and position the cursor at those absolute coordinates.
But cklicking there does not turn the radio button on.
I filed bsc#1204706 and got that comment from Stefan Hundhammer:
"I just checked: The only way to get a radio button into this state with just mouse operations is to click inside the radio button area (no matter if the click is on the text or on the ( ) indicator) and not also release it, then, while the mouse button is still pressed, move the mouse cursor outside the widget's area and release the mouse button there."
Running an manual installation on a VM verifies this statment, the only way to get the text "green" but the radio button still unselected is by clicking there and then moving the mouse away. Which doesn't happen according to the log.
Acceptance Criteria¶
- AC1: Find out why the radio button is not selected.
- AC2: Fix this problem.
Updated by JERiveraMoya over 1 year ago
- Related to action #121912: Create test suite to configure support server for SLE 15 SP4 added
Updated by JERiveraMoya over 1 year ago
Investigated in the related ticket, probably an old image is causing the issue.
Updated by JERiveraMoya over 1 year ago
- Tags set to qe-yam-refinement
- Status changed from Workable to New
Updated by JERiveraMoya over 1 year ago
- Status changed from New to Rejected
this issue has not been seen for a while. Rejecting it. Please reopen if it reoccurs.
I will check why we don't have there sle 15 sp4, it was supposed to be set in
https://gitlab.suse.de/qsf-y/qa-sle-functional-y/-/merge_requests/475/diffs#14b5739fca83e73d73bda6c083b58b96f0fd3bdc_706_705