BigInt bitwise and, or, xor

[thejoshwolfe]

No description provided.

[rwaldron]

This is not a review "blocker", but I'm wondering what you think about using binary integer literal bigints in these tests? For example:

-assert.sameValue(Object(5n) & 3n, 1n, "Object(5n) & 3n === 1n");
-assert.sameValue(3n & Object(5n), 1n, "3n & Object(5n) === 1n");
+assert.sameValue(Object(0b101n) & 0b011n, 0b001n, "Object(0b101n) & 0b011n === 0b001n");
+assert.sameValue(0b011n & Object(0b101n), 0b001n, "0b011n & Object(0b101n) === 0b001n");

I will understand if that's a bunch of work that you don't want to do ;)

[thejoshwolfe]

I like the idea of making the tests more readable. Perhaps using binary literals for the small values? I certainly don't want to write 0xbf2ed51ff75d380fd3be813ec6185780n as a binary literal.

There's not much I can do to make the cases with opposite sign more readable; i still find those pretty hard to understand. 0b11 & -0b11 === 0b01 Writing that with binary literals doesn't really clarify anything. I'd like to leave those cases as they are, since the binary representation might actually be more confusing.

How about I change the cases to use binary notation where 0 <= a <= 7 and 0 <= b <= 7? That would cover all the bigint-non-primitive.js cases, and the first 15 asserts in the bigint.js tests (everything before the first 0xffffffffn case).

[rwaldron]

I certainly don't want to write 0xbf2ed51ff75d380fd3be813ec6185780n as a binary literal.

hahahaha definitely not.

[rwaldron]

How about I change the cases to use binary notation where 0 <= a <= 7 and 0 <= b <= 7? That would cover all the bigint-non-primitive.js cases, and the first 15 asserts in the bigint.js tests (everything before the first 0xffffffffn case).

Sorry, I only meant to suggest using them for bitwise operations :)

[thejoshwolfe]

Sorry, I only meant to suggest using them for bitwise operations :)

Sorry, I don't understand. This PR has 6 files; I am suggesting changes in all 6 files. For the 3 files named bigint-non-primitive.js, converting all the BigInt literals to use 0bxxxn format. For the 3 files named bigint.js converting the BigInt literals in the first 15 asserts to use 0bxxn.

[rwaldron]

Sorry, I don't understand. This PR has 6 files; I am suggesting changes in all 6 files.

The misunderstanding was mine, and I apologize. I just misread what you said.