feat: allow duplicate registry values

[maribethb]

The basics

• I validated my changes

The details

Resolves

Fixes #7924

Proposed Changes

Does not throw an error if you attempt to register the same object twice (under the same type and case-insensitive name)

Reason for Changes

This is similar to #7983. The difference in behavior is subtle:

In this PR: If you try to register the same TestObject twice, first under the name test and second under the name Test, then we'll allow you to register the object without throwing an error and we'll also update the cased name to be Test, which only matters if you later call getAllItems(type, true) where we give you the most-recently used cased version of the name. This is consistent with how it works if you pass a different object, same name, and pass the opt_allowOverrides parameter: we'll update the case of the stored name.

In #7983: If you try to register the same TestObject twice, first under the name test and second under the name Test, then we'll allow you to register the object without throwing an error. But we won't update the most-recently-cased version of the name. This is inconsistent with how it works if you pass a different object, same name, and pass the opt_allowOverrides. It's also a behavior change where previously, you could register the same object and intentionally change the case by passing the opt_allowOverrides flag.

I think maintaining consistency with how the overrides flag works is better, and it makes it easier to explain the "last used case" part of the registry that keeps track of the case-sensitive name you registered an item under. That's why I think this PR should be merged instead of #7983 (or #7982)

Test Coverage

Added unit test for this case.

Documentation

Maybe? We don't really explain the registry anywhere. The most likely place this gets explained is in the documentation about how installing fields/blocks plugins works.

Additional Information

See #7924 and #7987 for related discussion.

[rachel-fenichel]

I agree with your logic in choosing to be consistent with the behaviour for new names with opt_allowOverrides