Added a sleep for onscreen rendered tests

[chaosphere2112]

This should help generalize what @durack1 did in #1341 (adding a sleep to handle the race condition with resized windows). The root of the issue is #1148; I have some tests that resize the window (to make sure everything is well-behaved), and because of #1148 they have to be rendered onscreen. This makes them run slowly, which looks like it adds a race condition– see intermittent failures of the resize manager and background color tests. This makes it so the base class checks if the window is set to render offscreen; if it isn't, it will sleep.

[durack1]

@chaosphere2112 it doesn't seem like you triggered the buildbot tests - you might have to push this PR as a new branch on UV-CDAT/master to get these to fire

[jbeezley]

Yeah, you need to push this to the main repo for the tests to run. Anyway, this change gets a 👍 from me.

[chaosphere2112]

😕 I can't seem to push to this repo. I'm set up with private key/SSH, and it tells me that I can't push, and that I should try using HTTPS instead. That's an issue for me, because I have 2FA set up, and it is a giant pain making that work. Any ideas?

[chaosphere2112]

Never mind, looks like I figured out how to do it over https...

[chaosphere2112]

@jbeezley Looks like the garant buildbot didn't get much of anywhere on this... any ideas?

[jbeezley]

Weird the git clone failed:

error: RPC failed; result=6, HTTP code = 0
fatal: The remote end hung up unexpectedly

I'll restart it.

[chaosphere2112]

@doutriaux1 @jbeezley @aashish24 Anything preventing this from being merged?

[jbeezley]

No, looks good.

[doutriaux1]

@durack1 was telling me yesterday that sleep(1) might be better than sleep(2). But if you're cool with sleep(2) I can approve

[chaosphere2112]

This should only happen for a few things, and I'd rather be over-cautious than under. Go ahead and approve.

[durack1]

@doutriaux1 @chaosphere2112 just FYI, I did pull my sleep(1) back out of #1341 and sure enough the PR failed again due to this race condition so I might have to rebase #1341 to master merging #1393 changes to bring things back to passing..

Just FYI, the single file kludge sleep(1) is now in the master branch - this should probably be reverted as this merged PR #1393 should have fixed that issue..

[durack1]

@doutriaux1 @chaosphere2112 this should be fixed in updated PR #1341 with these edits