Oh, be ware that your Allzpark preference will be cleaned when running test.
Update: fixed in 24ba9b9

Good initiative. :) Happy for you to take the lead and determine whether and when this is ready to merge.

So I was trying to run tests in Azure pipielines, and bumped into a few problems. Leaving notes here.

Ubuntu 16.04 can install pyside with apt-get install python-pyside, but no pyside2 available. Ubuntu 20.04 can install pyside2, but no pyside. Using Xvfb + pyside2 passed the test, seems doesn't works with pyside (tests still runnning after an hour). Qt5 has an environment var QT_QPA_PLATFORM, set to "offscreen" seems can be an alternative of Xvfb.

Building PySide on MacOS takes long long time.

Windows Python 2.7 somehow passed all the tests but still error out at the end (powershell return code 1, cmd exited with code -1073741819).

bleeding-rez doesn't handled memory-based package repository's address properly, should be an easy fix (see log).
(update: the package path was being assumed as filesystem path and normalized here)

Although time.clock has been deprecated but not yet removed before Python 3.8 accroding to doc, but the Python 3.7 in Azure pipelines' Unbuntu doesn't have that attribute. (fixed in 99c7de7)

Vandorized transitions module isn't using vendorized six (see log). (fixed in ec45019)

Here's my ci history of all the testing commits. 😓

Setting this PR as "ready for review" for running tests.

Ubuntu 16.04 can install pyside with apt-get install python-pyside, but no pyside2 available. Ubuntu 20.04 can install pyside2, but no pyside

Sorry I meant to reply to this sooner! Yes, haha. I know exactly what you mean. The Qt.py project struggles with this too.

I see you just dropped 2.7 support in testing, I think that's the way to go. There's no need to test this with every version of Python and Qt binding. Keep it simple; some tests in some form are better than none in any form.

Keep it simple; some tests in some form are better than none in any form.

Totally !
One last thing I am trying to solve is this: "....QThread: Destroyed while thread is still running", this error pops up randomly, but shouldn't be hard to resolve.

Is it an error though? I suspect it's showing up just as Python is about to shutdown; from another error.

I suspect it's showing up just as Python is about to shutdown; from another error.

Not sure.., but I see there are 4 dots before the line "QThread: Destroyed while thread is still running", and there are 5 tests, which means previous 4 all passed but the last one. So I guess it's because the fast exiting. 🤔

There's a verbose mode to tests (if it's using nose?) (which I really think should be default, but alas) to turn those dots into lines, with the name of each test on it an a FAIL or OK. That could help.

Yap.. should have more info :(
For future good, let's use nose !

Hooray ! Refactored and nose adopted, all tests passed !
I think we are ready to merge this !

Woho! Nice work :)

Merging this !