Changes due to stage 3 reviews to present in July 2023

[nicolo-ribaudo]

The bar for changes to the proposal is now high, but there are some changes that are either already part of the proposal spec or are planned. I will bring these up for consensus at the July TC39 meeting.

• Collect all the attributes in dynamic import before validating them (Stage 3 reviews #137 (comment)). The difference between the old and the new behavior is currently not tested neither by test262 not by V8's tests, so regardless of what we decide we should add a test for this.
• Remove the [no LineTerminator here] restriction before with. This was necessary with assert to avoid ASI ambiguities, but given that with is a reserved word we don't have this problem anymore. I noticed this wile implementing the proposal in Babel, and it has also been reported at No need for NLTH with with keyword #136.
• Allow any LiteralPropertyName in import attributes keys, making with { 0: "a", 1n: "b" } syntactically valid. This is for consistency with objects, and thus with the dynamic import form. (Normative: Add Import Attributes ecma262#3057 (review))

[nicolo-ribaudo]

Slides: https://docs.google.com/presentation/d/1XKSeyxhCiSrzqJRqZ6ioYeqHh72oBHkd9izufPiRktY

[nicolo-ribaudo]

Consensus on all these changes. However, after plenary I received this feedback from SpiderMonkey: #145