Can we add isBigInteger? As a work-around, what do you recommend?

See pointbiz/bitaddress.org#132

The same issue exists in this repository.

There was some discussion about changing this to be backed by a Buffer. I think that'll actually be slower since operations will then be preformed on a 8 bit numbers as opposed to what it uses now.

By analyzing the code, it does appear to in fact take advantage of JavaScript's capability of handling integers up to 53 bits. So porting to JavaScript native typed arrays may actually introduce a performance penalty.

Maybe there may be a way to leverage some of the concepts of asm.js? Not sure. It's worth looking at potential changes and benchmarking them.

Could be improved: http://cryptocoinjs.com/modules/misc/bigi/
File on Github: https://github.com/cryptocoinjs/cryptocoinjs.com/blob/gh-pages/markdown/modules/misc/bigi.md

Having a big integer, this is the string representation (10 radix): "61112663125671126633722436054102056730607837166180410453588415302234407842306"

The method toDERInteger() which convert to the bigendian bytes would return:
[0, -121, 28, -118, -93, -127, -101, 97, 127, 69, -4, -15, 127, -126, 42, -104, -89, 8, 44, 66, 30, 73, -84, -5, -54, -93, -87, 83, -28, 113, 30, 50, 2]

But the same method from pointycastle returns:
https://github.com/PointyCastle/pointycastle/blob/master/lib/src/utils.dart#L21

unsigned:
[135, 28, 138, 163, 129, 155, 97, 127, 69, 252, 241, 127, 130, 42, 152, 167, 8, 44, 66, 30, 73, 172, 251, 202, 163, 169, 83, 228, 113, 30, 50, 2]

convert to signed:
[-121, 28, -118, -93, -127, -101, 97, 127, 69, -4, -15, 127, -126, 42, -104, -89, 8, 44, 66, 30, 73, -84, -5, -54, -93, -87, 83, -28, 113, 30, 50, 2]

Wondering what is the leading 0...

I have added bigi flow definitions here

flow-typed/flow-typed#205

(not really an issue, just wanting to let you know)

Seems like there are issues in converting buffers to BigInteger instances.
Below you can see the sample code example and the output of it in node.js environment.

'use strict';
const BigInteger = require('bigi');
const bigNegativeNumberString = '-12345678900';
const bigIRepresentation = new BigInteger(bigNegativeNumberString);

console.log([
    bigIRepresentation.toString(),
    BigInteger.fromBuffer(bigIRepresentation.toBuffer()).toString(),
    BigInteger.fromHex(bigIRepresentation.toHex()).toString()
]);

Output:
[ '-12345678900', '1087165948876', '1087165948876' ]

The toByteArrayUnsigned() seems broken:

b=BigInteger('1234567890',10)
b.toByteArrayUnsigned()

results in: [73, -106, 2, -46] while there should be no negative numbers.

In fact, it gives the same result than toByteArray()

e3271bf#commitcomment-6146920

• https://code.google.com/p/octane-benchmark/source/browse/latest/trunk/crypto.js?r=28

Borrow benchmarks from here? https://github.com/indutny/bn.js/tree/master/benchmarks

I have a private key of type buffer.

In my local env, when running gatsby build the line BigInteger.fromBuffer(privateKey) console logs this:

BigInteger {0: 5218218, 1: 6291014, 2: 34656281, 3: 1776116, 4: 53518743, 5: 10808093, 6: 62725491, 7: 63913863, 8: 50275223, 9: 1979384, t: 10, s: 0}

When I am in production, run with gatsby build, the same exact line outputs this:

n {0: 5218218, 1: 6291014, 2: 34656281, 3: 1776116, 4: 53518743, 5: 10808093, 6: 62725491, 7: 63913863, 8: 50275223, 9: 1979384, t: 10, s: 0}

so the ecsda signature fails, saying it expects a BigInteger and is receiving type n

Until we have a constructor: #18
... or a fromInt method, what is the best way to create a BigInteger from a number (even one with a decimal place)?

Here is my work-around, but it produced a number that was much too large in the second case:

BigInteger("11").toString() == "11"
BigInteger("1.1").toString() == "11"

Both lines return true

     TypeError: Cannot read property 'nextBytes' of undefined
      at BigInteger.bnpFromNumber [as fromNumber] (node_modules/bigi/lib/bigi.js:775:6)
      at new BigInteger (node_modules/bigi/lib/bigi.js:7:36)

Is it meant to be supported?

It appears the Miller-Rabin test implemented in bigi always chooses the base from the set of primes < 1000, however it is possible to construct a composite that appears prime relative to any fixed set of witnesses, and the paper https://math.dartmouth.edu/~carlp/PDF/reliable.pdf has a proof that there exist integers with no small witnesses at all.

The best fix is probably to choose the witnesses randomly from the range [2,n) which avoids the problem entirely.

I see byteCount method is listed in the document. But when I tried to use it, it throws error:
Object has no method 'byteCount'
Thanks :)

It would be nice if you could initialize bigi with an intereger like so

var bi = require('bigi'); 
var aNumber = bi(9);

Currently if you try this you get

typeError: Cannot call method 'nextBytes' of undefined
    at BigInteger.bnpFromNumber [as fromNumber] (/home/null/code/bigi/lib/bigi.js:775:7)
    at new BigInteger (/home/null/code/bigi/lib/bigi.js:7:36)
    at BigInteger (/home/null/code/bigi/lib/bigi.js:4:12)

For reference: cryptocoinjs/ecurve#25

@dcousens thoughts?

Code is like this below

const num1 = 726
const BigIntegerOne = bigi.fromHex(toHex(num1))

function toHex(num) {
  let hex = num.toString(16)

  return hex
}

When I used the method 'fromHex', I found that it thow an error, "Incomplete hex".
So, I don't know what does it mean.
Could you please tell me the reason why you handle this error message.
Hope see you response in few time.
Thank you.