What is the issue with the Infra Standard?

See: #230, #643, #660, whatwg/html#10924, whatwg/html#10315

I think the following would work:

• Define "equals" for all of Infra's primitive data types. Most are "obvious". Special notes:
  • Infra doesn't really define numbers, but we should call out that we intend mathematical equality by default
  • We already have a definition of "is" for strings
• Disallow using "equals" on lists and maps, since it is too confusing.
• Not 100% sure on the right path for structs, but my initial proposal:
  • Say that it works by default by comparing itemwise
  • Specifications can override the definition of "equals" for their particular struct (e.g. as URL does)
  • Note this means that structs containing lists do not work by default, hmm.
• Say that every other type we use in specs defaults to "reference equality", but can define their custom equals overload if they want.
  • Examples of such types that are not just structs in disguise: algorithm steps, IDL values, JS values
• With "equals" now defined for all types except lists and maps and things containing them, define "deep equals" for lists and maps, which recursively uses "deep equals" on keys/values/items
  • This can replace https://infra.spec.whatwg.org/#set-equal
  • If we think it's OK, per the above, for "equals" to not work on structs containing lists and maps, then we should define "deep equals" for structs.
    • It kind of sucks that you have to check if your struct recursively contains list and maps, before determining which operator to use? But maybe it's good for clarity?
• Optionally define "shallow equals" for sets and maps, although I'm not sure we have any use cases
• Go through existing places in Infra that use ambiguous notions of equality, like our definition of "list contains", and replace them to use "equals".