CT: perennial is failing scenerystack-test

[pixelzoom]

perennial has been failing CT for at least a couple of weeks with the same error(s) in "scenerystack-test". The output is huge (209998 lines!), so I can't copy-paste it from CT to this issue. The first part of the output is shown below in "Details" below.

Since perennial is one of PhET's core repos for build tools, and since this is related to SceneryStack, it seems like some priority should be given to fixing this.

perennial : scenerystack-test
npm run failed with status code 1:

perennial@0.0.0 scenerystack-test
bash bin/sage run js/scripts/scenerystack-test.ts

cloning demo-sim
npm update demo-sim
info: npm update in ./.scenerystack/demo-sim
cloning scenery-template-parcel
npm update scenery-template-parcel
info: npm update in ./.scenerystack/scenery-template-parcel
cloning scenery-template-vite
npm update scenery-template-vite
info: npm update in ./.scenerystack/scenery-template-vite
cloning scenerystack-tests
npm update scenerystack-tests
info: npm update in ./.scenerystack/scenerystack-tests
cloning sim-template-parcel
npm update sim-template-parcel
info: npm update in ./.scenerystack/sim-template-parcel
cloning sim-template-vite
npm update sim-template-vite
info: npm update in ./.scenerystack/sim-template-vite
tsc demo-sim
tsc scenery-template-parcel
tsc scenery-template-vite
tsc scenerystack-tests
tsc sim-template-parcel
tsc sim-template-vite
demo-sim vite serve

debug: Running npx scenerystack checkout from ./.scenerystack
debug: stdout: [scenerystack]: checkout
...

[jonathanolson]

Above fix will patch the current failure. However since there isn't the expectation (currently) that other devs should adjust scenerystack to be compatible (and the many hours of trying to get tests working nicely on sparky), I'm considering whether it is best to have this removed from CT. @zepumph thoughts?

I'd rather have a improved local setup for when I'm building scenerystack.

[zepumph]

If this test is failing because it is more challenging than we thought to test scenerystack on CT, and it will require time to get the test to pass because of challenges with the kind of test we are doing, then we should have a conversation about investing in support of the scenerystack ecosystem in this way. If the conversation is instead that CT is failing because scenerystack's build is actually failing this often because of changes to our codebase, then isn't the test doing what it is supposed to? If the latter, then probably we need a process change so that the scenerystack build can be maintained with our codebase.

[zepumph]

This is passing since this commit made yesterday. 5e68fe8. Closing