test(ast): skip bigint literal value

[hi-ogawa]

Acorn preserves bigint literal 123n as

{
  type: "Literal",
  value: 123n,
  bigint: "123",
  raw: "123n"
  ...
}

but this is serialized into value: "123n" in https://github.com/oxc-project/acorn-test262/blob/ccd386837e02a92036b948239c67abf864cdd785/index.mjs#L56-L60

Since OXC cannot emit non-json 123n and use null instead, this creates many diffs like:

- value: "123n"
+ value: null

For now, I stripped such diff in conformance runner.

[hi-ogawa]

• test(ast): skip bigint literal value #8994 👈 (View in Graphite)
• main

---

How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

• 0-merge - adds this PR to the back of the merge queue
• hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

This stack of pull requests is managed by Graphite. Learn more about stacking.

[overlookmotel]

I suggest we change Oxc's behavior so it matches what's coming out of acorn-test262, instead of skipping it in the conformance runner. Obviously, that's wrong, but anything we do will be wrong!

To create an actual BigInt, we'll need to do a "fixup pass" on JS side after JSON.parse.

We can probably make that quite swift if we also compile and send to JS a list of "paths" to where BigInts are in the AST (e.g. for const x = [123n];, path is ['body', 0, 'declarations', 0, 'init', 'elements', 0, 'value']) - so it doesn't have to visit the whole AST to replace one BigInt. But we'll need a stateful serializer on Rust side to do that.

[overlookmotel]

I changed my mind. Have changed acorn-test262 to output BigInts as null to match Oxc's current behavior - #9021. This removes the need for conformance runner to skip the fields.

I prefer this method, because I'm hoping to complete the custom serializer which omits TS fields itself soon, at which point it'll replace the code for stripping out fields in conformance runner, and I didn't want all these tests to break again at that point!