Editorial: Add an assertion for PrivateEnvironment cannot be null

[kimjg1119]

In RelationalExpression, evaluation of RelationalExpression : PrivateIdentifier in ShiftExpression makes an alarm in ESMeta since it is not trivial to know that running execution context's PrivateEnvironment never becomes null.

It can be resolved in the same way with step 2 of MakePrivateReference; add an assertion to annotate it.

ESMeta alarm:

[ParamTypeMismatch] Call[9993] argument assignment to first parameter _privateEnv_ when Call[9993] function call from RelationalExpression[7,0].Evaluation (step 6, 7:33-92) to ResolvePrivateIdentifier
- expected: Record[PrivateEnvironmentRecord]
- actual  : Record[PrivateEnvironmentRecord] | Null

[michaelficarra]

I don't see why it's not possible for the running execution context's PrivateEnvironment to be null here. It seems like a spec bug to me.

[bakkot]

I don't see why it's not possible for the running execution context's PrivateEnvironment to be null here. It seems like a spec bug to me.

PrivateEnvironment can only be *null* if there is no containing class, in which case early errors statically prevent using the a.#b syntax.

[michaelficarra]

Oh okay, I didn't see the early errors, but I figured that was it.

[bakkot]

It's the ones which refer to AllPrivateIdentifiersValid. Mechanically it says that you can't refer to a.#name unless #name is declared in a containing class, but since you can't declare a private name without a class, that ends up meaning you can't use the syntax at all outside of a class.