Fix various parser idempotency issues and parsing errors

[raskad]

This PR fixes a bunch of parser idempotency issues, but also some parsing errors where our parser was behaving non spec compliant. I found all of the issues via the parser-idempotency fuzzer.

There are a lot of changes here regarding the names of function expressions and their handling via has_binding_identifier. Since there are some more issues with that, I plan to alter our current handling to fix those and align the handling more with the spec in a follow up to this.

It changes the following:

• Fix cases where destructuring assignments would accept invalid assignment operators
• Fix ArrayLiteral formatting
• Prevent Literal::Undefined from being constructed by Arbitrary as it is only used for optimizations
• Fix Literal::BigInt formatting
• Remove tracking of excluded keys in object literals from ast and get them at compile time
• Fix formatting for object literal methods
• Fix Arbitrary implementation for TemplateLiteral
• Fix formatting for unary expressions to always include a space between operator and target
• Fix formatting of function expressions that do not have a binding identifier
• Fix formatting of classes
• Fix formatting of computed property names
• Move handling of object literal function names to the compiler
• Fix required check of semicolon at the end of a throw statement
• Fix formatting of TaggedTemplate

[github-actions]

Test262 conformance changes

Test result	main count	PR count	difference
Total	48,212	48,212	0
Passed	43,264	43,264	0
Ignored	1,413	1,413	0
Failed	3,535	3,535	0
Panics	0	0	0
Conformance	89.74%	89.74%	0.00%

[HalidOdat]

Nice work! Just needs a rebase :)