Occasional test fails - Fake XHR

[snapwich]

Type of issue

Bug

Description

This issue is to track a persistent issue with the hooks failing and breaking the tests:

HeadlessChrome 0.0.0 (Mac OS X 10.11.6) Trion adapter tests "after each" hook FAILED
	Error: Uncaught TypeError: Fake XHR onreadystatechange handler threw exception: Cannot read property 'seatbid' of undefined (node_modules/sinon/pkg/sinon.js:4469)
HeadlessChrome 0.0.0 (Mac OS X 10.11.6): Executed 1053 of 1430 (1 FAILED) (skipped 1) (0 secs / 2.904 secs)
HeadlessChrome 0.0.0 (Mac OS X 10.11.6) Trion adapter tests "after each" hook FAILED
	Error: Uncaught TypeError: Fake XHR onreadystatechange handler threw exception: CannotHeadlessChrome 0.0.0 (Mac OS X 10.11.6): Executed 1053 of 1430 (1 FAILED) (skipped 1) (18.494 secs / 2.904 secs)

Steps to reproduce

Run the tests a lot; not easy to reproduce otherwise.

Platform details

OSX, node 6.9.1, npm 3.10.8; but also seems to break occasionally in Travis-CI

Other information

Some similar issues fixed in #1626, but this issue remains and seems a bit more difficult to track down the cause of.

[mkendall07]

Problem might fix itself with prebid 1.0. Not sure it's worth tracking down or not at this point.

[tedrand]

Is there a way to re-run the Travis-CI tests when this happens on a PR update?

[dbemiller]

@tedrand I think one of the project collaborators have to do it.

I think by now everyone who reviews prebid PRs knows about this... so if tests fail, we always check to see if it's a true failure.

I agree with @mkendall07, though... unit tests for adapters are much less complicated in prebid 1.0. There's a good chance these will fix themselves naturally.

[snapwich]

It should be noted that this is still an issue in prebid 1.0. My latest pull-request #1976 failed twice in a row due to these errors before passing on the third try.

The error was in a beforeEach hook one time and afterEach hook the other time, but both related to "unknown function done in handler` or something like that.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.