re-enable all assertions in PhetioObject

[samreid]

This will help us identify the following issues:
(a) cases where an event triggers the an event on the same instance
(b) cases where an object is disposed while triggering an event.

From #46

[samreid]

Bayes is not currently operational, so we will have to rely on local testing to help catch issues.

[samreid]

Please see phetsims/equality-explorer#25 for an issue that was created in Equality Explorer about this assertion.

[samreid]

Local Aqua testing for phet brand (with the 3 assertions enabled) indicated issues in these sims:

• build-a-molecule
• expression-exchange
• molecule-shapes
• molecule-shapes-basics
• states-of-matter
• states-of-matter-basics
• sugar-and-salt-solutions problems noted in Namespaces are missing sugar-and-salt-solutions#29
• sun

I included all red sims (even those with seemingly unrelated issues), in case the unrelated issues may be masking this issue.

[samreid]

The problem in build-a-molecule is exactly what we have been talking about in phetsims/scenery#725 and phetsims/scenery#726. The drag event is calling endDrag. This is causing trouble for PhetioObject's assertions because PhetioObject's new assumption is that an event cannot trigger an event on the same object.

Putting it in a setTimeout solves the problem, but may not work well for input event playback.

[samreid]

I wonder if the problems in #48 (comment) were fixed by commits in #50 ?

[samreid]

I added the workaround to use PHET_CORE/Timer to invoke the end drag in the next animation frame.

[samreid]

Next round of testing

• build-an-atom: cannot dispose while event is in progress

[samreid]

After proposing changes to fix build-an-atom style cases, I added another assertion to make sure objects are not disposed twice, and it is triggering in Circuit Construction Kit.

[samreid]

Fixed a latent bug in CCK CircuitElement disposal.

[samreid]

After fixing that, Next round of testing had a "cannot start event while event is in progress" in black box study.

• circuit-construction-kit-black-box-study

[samreid]

I suspect it could be too restrictive to forbid objects from triggering double events on themselves (even though it seems rare).

[samreid]

On the other hand, in a previous case and this case catching double triggering revealed an underlying problem in the code, so maybe it is good to keep (until we find a case we cannot circumvent). Committing fix for MomentaryButtonModel shortly.

[samreid]

On second thought, I'm not convinced the change above is a good move--it works around the problem at hand but when the button becomes disabled and it does not change the value it may be incorrect for the data stream.

[samreid]

Better solution committed above.

[samreid]

Everything is passing aqua fuzz (except for sugar and salt solutions which has namespace problems). I'll test phet-io aqua before committing.

[samreid]

Everything is green on phet-io aqua. grunt lint-everything is clear, one commit coming up.

[samreid]

The changes in PhetioObject warrant another review by @pixelzoom. I'll create a new issue to make sure it is set for review.