BigNumber encoding hex result doesn't match test cases

I haven’t been able to recreate this encoding test case. Is there something I’m doing wrong?

const cbor = require('cbor')
const { BigNumber } = require('bignumber.js')

// This doesn't match the test case here: https://github.com/hildjj/node-cbor/blob/466d01b25959ab1270a409e31453f53b5bba082d/test/cases.js#L695
const num = new BigNumber('9223372036854775807')
console.log(cbor.encode(num).toString('hex')) // c2487fffffffffffffff

// This matches the test case here: https://github.com/hildjj/node-cbor/blob/466d01b25959ab1270a409e31453f53b5bba082d/test/cases.js#L80
const num2 = new BigNumber('18446744073709551616')
console.log(cbor.encode(num2).toString('hex')) // c249010000000000000000

Here’s a repl.it.

[hildjj]

Fixed in version 5.0.2. Use the collapseBigIntegers option.

console.log(cbor.encodeOne(0x7fffffffffffffffn, {
  collapseBigIntegers: true
}).toString('hex'))

generates: 1b7fffffffffffffff

[hildjj]

I guess you’re asking for BigNumber integers and ECMAscript bigints to be encoded as normal integers, if they would fit. I’m open to adding an option for that, but since it’s a breaking change as well as breaking round-tripping, I don’t want to make it the default.