test(jsnum): add binary/oct/hex BigInt cases including underscores and uppercase prefixes

[Harsh1925]

This PR adds additional unit tests to cover pseudo BigInt parsing across
non-decimal bases. It addresses the in-file TODO in
internal/jsnum/pseudobigint_test.go ("tests for other bases").

What’s included

• Binary literals (0b/0B), with and without underscores
• Octal literals (0o/0O), with and without underscores
• Hex literals (0x/0X), with and without underscores
• Small number cases (0, 1, 10, 255, etc.), complementing the existing
  large literal tests

Why

• Improves coverage for ParsePseudoBigInt
• Verifies parser correctly handles base prefixes, separators, and case
  sensitivity
• No behavior changes; tests pass with current implementation

[Harsh1925]

@microsoft-github-policy-service agree

[jakebailey]

I do not know why you are modifying this file.

[Harsh1925]

Thanks for the review! I added these cases because of the inline TODO in internal/jsnum/pseudobigint_test.go (“tests for other bases”). The TODO I was responding to is in internal/jsnum/pseudobigint_test.go near the “strip base-10 strings” section (just above the “can parse large literals” test).
The file already covers one large literal per base, but didn’t have small/underscore/uppercase-prefix cases, so I added those to complement the existing “large literals” tests.
If you’d prefer these live in a different package/file, or you don’t want to expand this test, I’m happy to move them (or close). Where would you like these cases to go?

[jakebailey]

I don't know what you're talking about, I'm commenting on the readme

[Harsh1925]

Thanks for the review. The README change slipped in by mistake on this branch.
I opened a clean PR with only the tests: #1678.
Closing this one to keep the diff focused.