Timeout error in test cases

Bug #1203010 reported by Javier Collado
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
webbrowser-app
New
Undecided
Unassigned

Bug Description

Reviewing the test case results for the webbrowser application in maguro:
http://10.97.0.1:8080/job/saucy-touch-maguro-smoke-webbrowser-app-autopilot/13/console

I see that all the failures there are related to the autopilot's Eventually matcher.

This means one of:
- there's a race condition (unlikely because 10 seconds seems a reasonable timeout value)
- the application is misbehaving and not really performing the actions that autopilot asked it to do
- there's a communication problem between the application and the autopilot processes and the updates in the widgets aren't correctly communicated.

Tags: qa-touch
tags: added: qa-touch
description: updated
Revision history for this message
Olivier Tilloy (osomon) wrote :

Most (if not all) of those failures are due to a bug in autopilot whereby it’s unable to type colons when run on a device.
As a result, the URLs entered for the tests are incorrect, and the pages don’t load properly. This is already tracked by bug #1181216, I don’t think there is a need for this bug report (I would mark it as duplicate).

Revision history for this message
Javier Collado (javier.collado) wrote :

If the reason for those test failures is known. Please mark them as skipped in the code to avoid any confusion in the future since the autopilot error doesn't provide much information by itself. Once that bug is fixed, then the changes to skip them can be reverted. Please let me know what do you think about this.

Revision history for this message
Olivier Tilloy (osomon) wrote :

No, I don’t think we should skip those tests. They all pass on a desktop or in the VMs used by jenkins. They fail only when run on a device.

This bug in autopilot is critical (and has been marked as such for more than a month), and should be addressed asap (I’ve been discussing it with concerned parties today, hopefully we’ll get a solution soon). Skipping tests is not a solution, it only hides the problem.

Revision history for this message
Javier Collado (javier.collado) wrote :

Can the tests be skipped only when run in a device? If that's the case, I'd still be in favor of doing that. I agree on that skipping test cases won't fix the problem; but, as I said, it will be less confusing when reviewing results.

Revision history for this message
Olivier Tilloy (osomon) wrote :

If that was only one test or two, I’d tend to agree, but we’re talking about half the tests failing here (~ 20 tests), the bug has to be fixed urgently, there’s no point in pretending it’s not there by skipping tests. Besides, I don’t know if we can skip tests conditionally.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.