Unblock LSC - Patch 2 - Fix initializer of instance members that reference identifiers declared in the constructor

[rileyajones]

Motivation for features / changes

There is an ongoing LSC cl/652971829 which changes how instance methods are set. The change is blocked because of our split codebase. This PR is intended as a replacement for go/tbpr/6883

[arcra]

I don't see this one in the before snapshot. Was this change intended?

[rileyajones]

This is a strange case where the code is actually duplicated. I removed the other copy.

[arcra]

About the failures, I can see some error messages related to being "[unable to] bind to 'cardObserver' since it isn't a known property of [some components]", which (I'm speculating, because I'm not super familiar with FE/Angular) causes other is not a known element error messages that are somewhat misleading.

That cardObserver input was changed in #6900, so (again, speculating) this issue might only be related to some of these observables that were recently changed to be initialized in the constructor, and which are used as inputs for other components. It's possible that adding a type annotation could fix it (?).

Then again... CI passed for that previous change, and I'm not sure if something in the changes on this PR make it so it only breaks with these new changes, so all of this might be a red herring (sorry if that's the case)... but just adding some thoughts here, for later.

[arcra]

About the failures, I can see some error messages related to being "[unable to] bind to 'cardObserver' since it isn't a known property of [some components]", which (I'm speculating, because I'm not super familiar with FE/Angular) causes other is not a known element error messages that are somewhat misleading.

That cardObserver input was changed in #6900, so (again, speculating) this issue might only be related to some of these observables that were recently changed to be initialized in the constructor, and which are used as inputs for other components. It's possible that adding a type annotation could fix it (?).

Then again... CI passed for that previous change, and I'm not sure if something in the changes on this PR make it so it only breaks with these new changes, so all of this might be a red herring (sorry if that's the case)... but just adding some thoughts here, for later.

Riley mentioned this might be just a typing issue, but unrelated to the test breakages, which would make sense since it was not a blocker for the previous PR to be merged.