yahoo-eng-team team mailing list archive

[Richard Jones <r1chardj0n3s@xxxxxxxxx>]

Public bug reported:

The attached is an example of a test run that failed, I believe because
it is testing for text in a cell that is updated as the test is run.
Therefore the StaleElementReferenceException race condition

1. element is fetched,
2. element is removed from DOM
3. element is tested (specifically text content) -> StaleElementReferenceException

is triggered. The code in integration_tests/basewebobject.py
_wait_until() could be made more robust by handling the
StaleElementReferenceException and returning False (ie. no match).

An alternative is to have the test in question wait until the data being
updated is stable, but the specific test failure I see
(integration_tests/tests/test_instances:TestInstances.test_filter_instances)
will not be able to do this reliably, there is still a race condition
between the cell indicating instability and the cell indicating status.
I believe the most reliable solution here is to add robustness against
the StaleElementReferenceException race condition in the _wait_until
method.

** Affects: horizon
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Dashboard (Horizon).
https://bugs.launchpad.net/bugs/1626830

Title:
  Integration tests do not handle StaleElementReferenceException well

Status in OpenStack Dashboard (Horizon):
  New

Bug description:
  The attached is an example of a test run that failed, I believe
  because it is testing for text in a cell that is updated as the test
  is run. Therefore the StaleElementReferenceException race condition

  1. element is fetched,
  2. element is removed from DOM
  3. element is tested (specifically text content) -> StaleElementReferenceException

  is triggered. The code in integration_tests/basewebobject.py
  _wait_until() could be made more robust by handling the
  StaleElementReferenceException and returning False (ie. no match).

  An alternative is to have the test in question wait until the data
  being updated is stable, but the specific test failure I see
  (integration_tests/tests/test_instances:TestInstances.test_filter_instances)
  will not be able to do this reliably, there is still a race condition
  between the cell indicating instability and the cell indicating
  status. I believe the most reliable solution here is to add robustness
  against the StaleElementReferenceException race condition in the
  _wait_until method.

To manage notifications about this bug go to:
https://bugs.launchpad.net/horizon/+bug/1626830/+subscriptions