test: add potential failing test for #110

[Techn1x]

aimed to be a failing test for #110

[vercel]

@Techn1x is attempting to deploy a commit to the universal-ember Team on Vercel.

A member of the Team first needs to authorize it.

[Techn1x]

The secret sauce to get the test to fail was using the value in a modifier before using it for something else.

This test passes when either;

• you comment out the div with the modifier, or
• the value is referenced before the modifier (eg adding {{log this.testCase.endResult}} to the line before the div, or
• you use < ember 5.5ember 5.6

[Techn1x]

This test passes when either;

• or, my production builds seem to behave OK. I can't verify this here in the test though.

[NullVoxPopuli]

so I pulled this down, only to realize that it needs rebased!

can you rebase? thanks!!
(and thanks for the repro!)

[Techn1x]

Whoops! I was operating out of an old fork.

Rebased.

Feel free to modify / undraft this work how you see fit. I called the test "failing case" for lack of a better name, given I'm not sure what the core problem is that it's testing for

[NullVoxPopuli]

5.5 is ok, but 5.6 isn't -- this matches my expectations of ember-source, because 5.6 had the VM update

[NullVoxPopuli]

I'm still investigating, but my worry is that this is related to either of these two PRs:

• Don't infinitely duplicate positional arguments when modifiers update glimmerjs/glimmer-vm#1565
• Refactor on modifier implementation glimmerjs/glimmer-vm#1573

[Techn1x]

Would either of those PRs be expected to exhibit different behaviour between test / prod builds?

[NullVoxPopuli]

I wouldn't expect a difference in behavior, no -- but that doesn't mean something accidentally didn't happen! 🙈