Blue background overlapping with other parts parts of screen with wire movement

[etwa4650]

Test Device

JORDAN

Operating System

MAC OS 10.14.4

Browser

Firefox

Problem Description

for phetsims/qa#311

While moving one end of a wire, the blue background will appear over parts the black banner at the bottom and the toolbox on the right hand side.

Seems like the effects might be more dramatic with more wires connected/added.
Was able to duplicate in both the "Intro" and "Lab" simulations.

This also occurs in the published version.

Screen recording in the link below.

Steps to Reproduce

Add a wire vertically on the left hand side of screen.
Add a wire horizontally at the bottom of the screen, close to the bottom edge.
Drag and stretch the right end of the bottom wire around.

Note: I have been able to replicate this with just a single horizontal wire at the bottom, but it is more easily reproduced with the vertical wire.

Visuals

https://drive.google.com/file/d/116UdA4XaiKHMGhf4_u8MpAs8Y6SAU3xz/view?usp=sharing

[samreid]

Great discovery @etwa4650! @jonathanolson have you seen anything like this or have any ideas about it?

[samreid]

I am unable to reproduce this problem on Mac OS 10.14.4 running Firefox 66.0.3 (64-bit). What firefox version was this seen on? Is there another QA machine that can be used to try to try to reproduce this behavior?

[KatieWoe]

@samreid I'm seeing it on 66.0.3, though it seems less pronounced for me than in the video. @amanda-phet checked on her machine and we couldn't reproduce it there.

[KatieWoe]

Looks like it may be related to phetsims/area-model-common#175, if that helps

[etwa4650]

Also seen on Windows 10 in phetsims/qa#369

Although it seems much harder to replicate.

[samreid]

@ariel-phet would you like investigation into this issue before publication of CCK DC and DC Virtual Lab with the fuse component? It was the only issue pointed out during RC.1. If you would like investigation into this, I believe @jonathanolson may be best versed in problems with Firefox repaint regions.

[kathy-phet]

Does this exist in the published version? If yes, then I would vote for publishing as is, if not, then investigate what change to scenery may be causing this issue? ... but just defer, not close.

[jonathanolson]

I'm also curious about whether WebGL was used for that error (as it seems similar to some instances in WebGL with invalid pixel data)?

[samreid]

Does this exist in the published version? If yes, then I would vote for publishing as is, if not, then investigate what change to scenery may be causing this issue? ... but just defer, not close.

It was stated in the initial issue description that:

This also occurs in the published version.

@etwa4650 or @KatieWoe can you additionally test the sim with ?webgl=false to see whether that suppresses the problem?

[ariel-phet]

@samreid a few points:

1. We are not supporting firefox and Mac anymore, so that is not a concern
2. I can reproduce on FF and Win 10 (in both dev and published version)
3. ?webgl=false does NOT suppress the problem
4. Pretty easy to trigger with even a single wire

[ariel-phet]

Since this exists in the published version, will defer for now.