Project

General

Profile

action #55703

send_keys() doesn't guarantee that the application will handle modifiers+normal_key properly

Added by SLindoMansilla 10 months ago. Updated 6 months ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Feature requests
Target version:
Start date:
2019-04-15
Due date:
% Done:

75%

Estimated time:
(Total: 126.00 h)
Difficulty:
Duration:

Description

Observation

After several issues with "missing keys" when modifiers+normal_keys are involved, we realized that we can reproduced the same issues that openQA was showing by manually pressing the two keys simultaneosly. Those experiments, while far from being a scientific prove, help us understand what could be going wrong.

Different applications handles those keys different (eg. gedit, firefox).

Suggestion

  • Be sure that when using send_keys(), they are sent in the positional order. (eg. if alt-a is sent, alt should be sent first and then 'a', with enough time to let the application handles them.
  • Check with coolo about implementation of the VNC wrapper.

Subtasks

action #54422: [functional][u] LO recent documents: alt-f not working on wayland (TW)Resolved

openQA Tests - action #50399: [functional][y] Installer doesn't react to key pressNew

openQA Tests - action #54425: [functional][u] hexchat: ctrl-a mis-interpreted as 'a' since we switched to QEMUVGA "std"Resolveddimstar

action #57077: [functional][u][sporadic] test fails in vlc - open network protocol setting page is unreliable with ctrl-nResolvedcoolo


Related issues

Related to openQA Tests - action #39926: [functional][u][sporadic][virtio] test fails in vlc - vlc is started twice, needle target_match_vlc seems wrongRejected2018-08-17

Related to openQA Tests - action #43889: [functional][u][virtio][wayland] test fails in ooffice - openQA makes spelling mistakesBlocked2018-11-16

Related to openQA Tests - action #54056: [functional][u] test fails in virtman_viewResolved2019-07-10

History

#1 Updated by SLindoMansilla 10 months ago

  • Subject changed from [epic] send_keys() doesn't guarantee that the application will handle modifiers+normal_key properly to [epic][functional][u] send_keys() doesn't guarantee that the application will handle modifiers+normal_key properly

#2 Updated by SLindoMansilla 10 months ago

  • Status changed from New to Workable

Sub-Tasks in place

#3 Updated by SLindoMansilla 10 months ago

  • Related to action #39926: [functional][u][sporadic][virtio] test fails in vlc - vlc is started twice, needle target_match_vlc seems wrong added

#4 Updated by SLindoMansilla 10 months ago

  • Related to action #43889: [functional][u][virtio][wayland] test fails in ooffice - openQA makes spelling mistakes added

#5 Updated by SLindoMansilla 10 months ago

  • Assignee changed from SLindoMansilla to mgriessmeier

#6 Updated by zluo 9 months ago

  • Blocks action #54056: [functional][u] test fails in virtman_view added

#8 Updated by zluo 7 months ago

it seems that the PR doesn't solve our issue: #54056: [functional][u] test fails in virtman_view

#9 Updated by coolo 7 months ago

  • Blocks deleted (action #54056: [functional][u] test fails in virtman_view)

#10 Updated by coolo 7 months ago

  • Project changed from openQA Tests to openQA Project
  • Subject changed from [epic][functional][u] send_keys() doesn't guarantee that the application will handle modifiers+normal_key properly to send_keys() doesn't guarantee that the application will handle modifiers+normal_key properly
  • Category changed from Bugs in existing tests to Feature requests
  • Status changed from Workable to New
  • Assignee deleted (mgriessmeier)
  • Target version set to Ready

https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/8862 shows that send_key('alt-v') is too fast in cases. We should either make combinations of keys generally a lot slower or provide an option to do so.

#11 Updated by SLindoMansilla 7 months ago

  • Related to action #54056: [functional][u] test fails in virtman_view added

#12 Updated by SLindoMansilla 7 months ago

The relation helps us in our queries to know that someone is working on that so that we don't duplicate tickets/efforts

#13 Updated by coolo 7 months ago

I just removed the 'blocked by' as it's obviously not blocked by this - as seen by the PR

#14 Updated by ggardet_arm 7 months ago

send_key does not care about VNC_TYPING_LIMIT contrary to type_string.

We could try to update send_key to respect VNC_TYPING_LIMIT and if it is not enough, we could implement a slow_send_key or add an arg to send_key to adjust it, as suggested by @coolo.

#15 Updated by okurz 7 months ago

that sounds like a very good idea.

#16 Updated by coolo 7 months ago

I would make the default slower in general - the typing limit should be respected nonetheless

#18 Updated by okurz 7 months ago

  • Status changed from New to Feedback
  • Assignee set to coolo
  • Target version changed from Ready to Current Sprint

coolo created a spin-off in https://github.com/os-autoinst/os-autoinst/pull/1260 . coolo can you track this with it's effect on production?

#19 Updated by coolo 7 months ago

  • Status changed from Feedback to Resolved

we're good. So far no complains about typing too slow - and actually fixing some cases

Also available in: Atom PDF