Changing clip area does not work with renderer: canvas

[jessegreenberg]

From in phetsims/resistance-in-a-wire#164. In resistance-in-a-wire, I tried setting WireNode's dotsNode to use renderer: canvas, but I discovered that this broke the clip area that is further down in the file.

dotsNode.clipArea = wireBody.shape.ellipticalArc( -width / 2, 0, PERSPECTIVE_FACTOR * height / 2, height / 2, 0, 3 * Math.PI / 2, Math.PI / 2, true );

@jonathanolson is this indicative of something wrong with clip area, or is setting renderer: canvas the wrong thing to do here?

[jessegreenberg]

On the other hand, over slack @pixelzoom said

I tried something similar with ACID_BASE_SOLUTIONS/MagnifierNode. Line 74, viewportNode, added renderer: 'canvas'. Clipping continues to work fine.

[jessegreenberg]

Ah, the clip area is working but doesn't update once set initially with renderer: canvas. Here it is in RIAW:

Confirmed in acid-base-solutions, I modified MagnifierNode to shrink the clipArea over time, here it is with default renderer:

And when viewportNode is given renderer: canvas, the clip area no longer updates.

[jessegreenberg]

In the above test in acid-base-solutions, I am updating the clipArea in step. With renderer: canvas, there is no observable update in the clip area in the animation frame, but as I cycle through the radio buttons I do see the clip area shrink:

[jessegreenberg]

Adding block-publication label so this is on the radar before someone goes into an RC. But if someone requires a changing clip area with canvas the issue should be obvious right away before an RC is created.

[zepumph]

This is also holding up the next rc release of RIAW, because of phetsims/resistance-in-a-wire#164.

[jonathanolson]

It looks like the listeners for clip area (and Canvas rendering) were not being added to all of the nodes that would potentially affect it. That should be fixed with the above commit.

@jessegreenberg can you verify that this fixes the issue? (snapshot comparison doesn't show any major regressions).

[jessegreenberg]

Thanks @jonathanolson! I just verified that this fixes this issue in resistance-in-a-wire, so this can be closed.

[zepumph]

This will need to be cherry picked into the RIAW RC branch, reopening and assigning to myself, see phetsims/resistance-in-a-wire#168 for more details.

[zepumph]

RIAW has been covered, see phetsims/resistance-in-a-wire#164 (comment). Closing