avoid type_string taking "too long"
Tests like https://openqa.suse.de/tests/976761#step/register_system/18 show that very long type_strings can be problematic
They will always be delicate to problems when they get 'too long' (such as 'over 200 characters'), and we should probably implement a log warning so if tests fail because of missing characters or keyboard buffers being full it the log reader can at least have a suggestion to trim the likely cause of failure and make their tests more resilient.
#1 Updated by riafarov over 2 years ago
In the failed test we actually call assert_script_run function. At first I was thinking about passing wait_screen_change parameter to it and then to underlying type_string call in script_run method in distribution.pm Do you think it's fine if we kill two birds with one stone, so we add warning when string is long and if it's over 200 characters we enable wait_screen_change option. Does it make sense?
#2 Updated by RBrownSUSE over 2 years ago
The issue with the specific test has it's own ticket (https://progress.opensuse.org/issues/19540 ) and I've asked the maintainer involved to look at it :)
One of the suggestions I gave Wei Jiang was to use wait_screen_change :) I think it would be the best way also.
If you really want to do that work for him, I will not object, but please mail him to let him know you are helping because I mailed him directly to make him aware of the ticket.
With the test specific issue being dealt with in it's own ticket, this just leaves the WARN for long type_strings, hence this seperate [tools] ticket. Does that make sense?
#3 Updated by coolo about 2 years ago
- Subject changed from [tools] WARN: when type_string is "too long" to avoid type_string taking "too long"
- Category set to 136
- Priority changed from Normal to Low
- Target version set to Ready
I don't see any point in warning - if we think type_string shouldn't be used for too long, we should just inject something in type_string ourselves.
- Status changed from New to Resolved
- Assignee set to asmorodskyi
I believe this was covered with https://github.com/os-autoinst/os-autoinst/pull/878