Closed jbphet closed 1 year ago
I can duplicate the issue locally using the query parameter string in the issue, so my local URL is http://localhost:8080/sun/sun_en.html?ea&brand=phet-io&phetioStandalone&phetioCompareAPI&randomSeed=332211.
Running Running "generate-phet-io-api" task
makes the problem go away locally, so I'll commit the resultant changes and keep an eye on CT and see if it clears up.
CT is now clear. Closing.
During today's developer meeting we discussed CT issues related to the phet-io API, and used this issue as an example. I mentioned that I fixed this CT issue without really investigating why the API had changed, and it was recommended that I dig a little deeper and make sure that the API change is okay. I'll reopen the issue.
I've looked over the commit that resulted from the regeneration of the phet-io API file (see above), and the most significant portions seem to be related to availableScreensProperty
, selectedScreenProperty
, and simTitleWithScreenNamePatternStringProperty
, which originate in joist. Since @samreid is the responsible dev for joist, I'm going to assign this issue to him to determine whether the API change that occurred here is expected and reasonable.
Looks like things were added and phetioReadOnly
values were revised. The string changes are to be expected. @jessegreenberg can you comment on the a11y ones?
Sorry, I am not familiar with this change. @zepumph can you comment on whether these Properties related to Voicing should be phetioReadonly
?
@samreid the "a11y ones" are from https://github.com/phetsims/utterance-queue/commit/6a48435a915bd55435dd94cdb7fb2c9853e873e6.
Looks like sun wasn't updated when we added the new screen Properties as well.
The readonly string changes are from https://github.com/phetsims/joist/commit/1adfb13b169f9176343386497b913c3d313d19a7
I can't help but feel like part of the issue is removing many sims from the API checking conversation in https://github.com/phetsims/perennial/commit/6c0647099b364dcc1d9fb862282d858ea780e707.
At least we have pre-commit hooks checking this stuff now. . .
I skimmed the API delta and wasn't worried about any of the a11y or readonly changes. If there is indeed a problem, I would hope it resurfaces during design review or QA testing. Closing.
There is a CT failure on sun, and it has been there since yesterday, so I should probably investigate. Here's the error: