Added by mkravec over 7 years ago. Updated over 7 years ago.
0%
Description
Command send_key $cmd{next} is ignored in user_settings or pressed too fast. Test then writes root passwords on user screen.
Next button works on second try (from user_settings_root), so I think this is openqa issue.
I am not sure this is entirely correct. The corresponding log excerpt is:
16:01:59.8032 17414 >>> testapi::_check_backend_response: found inst-userinfostyped-20160504, similarity 1.00 @ 523/222
16:01:59.8033 Debug: /var/lib/openqa/share/tests/sle/tests/installation/user_settings.pm:30 called testapi::check_screen
16:01:59.8034 17414 <<< testapi::check_screen(mustmatch='autologindisabled', timeout=30)
…
16:02:00.2425 17414 >>> testapi::_check_backend_response: found autologindisabled-20160504, similarity 1.00 @ 567/525
16:02:00.2426 Debug: /var/lib/openqa/share/tests/sle/tests/installation/user_settings.pm:40 called testapi::send_key
16:02:00.2427 17414 <<< testapi::send_key(key='alt-n')
16:02:00.4440 Debug: /var/lib/openqa/share/tests/sle/tests/installation/user_settings.pm:44 called testapi::check_screen
16:02:00.4442 17414 <<< testapi::check_screen(mustmatch='inst-userpasswdtoosimple', timeout=13)
…
16:02:13.4411 17421 no change 0
16:02:14.6922 17421 MATCH(inst-userpasswdtoosimple-20160504:0.00)
16:02:14.6933 17414 >>> testapi::_check_backend_response: match=inst-userpasswdtoosimple timed out after 13
16:02:14.7583 Debug: /var/lib/openqa/share/tests/sle/tests/installation/user_settings.pm:48 called testapi::record_soft_failure
16:02:14.7584 17414 <<< testapi::record_soft_failure(reason='bsc#937012')
…
16:02:14.7603 Debug: /var/lib/openqa/share/tests/sle/tests/installation/user_settings_root.pm:19 called testapi::assert_screen
16:02:14.7604 17414 <<< testapi::assert_screen(mustmatch='inst-rootpassword', timeout=30)
…
16:02:45.6280 17414 >>> testapi::_check_backend_response: match=inst-rootpassword timed out after 30
So it finds the 'autologindisabled' which is the last correct screenshot shown. Then it presses 'alt-n' and checks for the dialog 'inst-userpasswdtoosimple' which doesn't show up in 13 seconds. The next action after recording soft failure is the assert_screen for 'inst-rootpassword' which also times out so there is no typing after the password except for the 'alt-n' itself.
From the y2logs on https://openqa.suse.de/tests/518168/file/user_settings_root-y2logs.tar.bz2 in YaST2/macro_inst_initial.ycp I can see what the installer got as input. The last entry block there is
{
//
// 2016-08-16 12:02:00
UI::ChangeWidget( `id (`new_user), `Value, true ); // YRadioButton "Create New User"
UI::ChangeWidget( `id (`full_name), `Value, "Bernhard M. Wiedemann" ); // YInputField "User's Full Name"
UI::ChangeWidget( `id (`username), `Value, "bernhard" ); // YInputField "Username"
UI::ChangeWidget( `id (`root_pw), `Value, false ); // YCheckBox "Use this password for system administrator"
UI::ChangeWidget( `id (`autologin), `Value, false ); // YCheckBox "Automatic Login"
// UI::MakeScreenShot( "/tmp/yast2-0045" );
UI::FakeUserInput( `next );
return;
}
See the last UI::FakeUserInput( next ); ? I read this that the installer actually got the 'alt-n' and interprets it as 'next' but then still fails. The file y2log does not show an obvious error to me so I suggest you make this a proper bugzilla bug.
Test has passed in last 2 builds, I think we can close this poo.
Let's try the following: Find a useful hypothesis what could be the cause for each failure and close a ticket only with at least one accepted hypothesis. I hope you won't be mad at me but I guess this could be the best way for us to learn what could break :-)
https://openqa.suse.de/tests/540716 looks similiar