Open zepumph opened 3 weeks ago
This patch is much more put together, and I think we are close to a commit point. This is the change set meant for main, and not to be MR'd to all hydrogen sims.
This is a nice change because it doesn't involve any PhET-iO API changes to accomplish.
For MR:
Next steps:
supportsDynamicLocale
flag on localeProperty that just checks on availableLocales length> 1
Per status meeting with @jonathanolson @kathy-phet, I will create a new issue for the following specific task listed in https://github.com/phetsims/joist/issues/970#issuecomment-2159459698
I'm not sure if they would be implemented in joist(?) but I will put the issue there.
- Yotta does not correctly track the locale for standard wrapper use. This is because the locale is not provided through a startup query parameter, and phet-io only changes the localeProperty, much later in the process than when GA and yotta process. We can likely fix this by providing the locale to the standard wrapper, knowing that it will flip flop later, but allowing the startup state to be correct for preloads.
QA issue has been created, and RC's deployed. https://github.com/phetsims/qa/issues/1098
Alright. The QA issue showed lots of trouble. The list is in https://github.com/phetsims/qa/issues/1098#issuecomment-2181191980. I'll work on patches here.
I believe that some of these issues in https://github.com/phetsims/qa/issues/1098#issuecomment-2181191980 are still effecting main.
I have confirmed that main is behaving as expected. Removing blocking label.
After solving the issues noted from QA, here are the remaining problems that release-branch-checks are showing. Note the NS 1.5 issue is still lingering from the QA report.
- natural-selection 1.5 phet,phet-io ERROR Error: Error: Assertion failed: required tandems must be supplied [FAIL] en fallback expected for badly formatted locale in studio URL: http://localhost:56286/release-branches/natural-selection-1.5/natural-selection/build/phet-io/wrappers/studio/?&locale= fdsa
Done above.
I believe all problems above have been fixed now. Here are the notes from the patches made since the QA issue was created. I'm running the tests on release branches one more time, and will proceed with deploying rcs if everything passes.
Notes on the errors:
From https://github.com/phetsims/joist/issues/963 and https://github.com/phetsims/chipper/issues/1441, lots of changes occurred to the locale system (like support for localeData in main/babel, locale3, hyphens and case-insensitivity), but we didn't get things working for Standard PhET-iO Wrappers to pass through the locale query parameter. This was reported in https://github.com/phetsims/phet-io/issues/1881#issuecomment-2142674672.
From this @jonathanolson @samreid and I have a new plan for how to support this. We want localeProperty to support setting values that are identical to any value you can provide via the query parameter. This means using the same logic as is in initialize-globals'
checkAndRemapLocale()
.Beginning work can be found here: https://github.com/phetsims/phet-io/issues/1881#issuecomment-2155281836 and below.