coolaj86

function escapeHtml(str) {
  // TODO more
  return str
    .replace(/&/g, '&')
    .replace(/</g, '&lt;')
    .replace(/>/g, '&lt;')
    .replace(/'/g, '&#39;')
    .replace(/"/g, '&quot;')
  ;
}

coolaj86

  // XPub to Public Parts
  wallet.addrFromXPubKey = async function (xpubKey, index) {
    let derivedRoot = HdKey.fromExtendedKey(xpubKey);
    return await wallet.addrFromXPubRoot(derivedRoot, index);
  };
  wallet.addrFromXPubRoot = async function (derivedRoot, index) {
    let derivedChild = derivedRoot.deriveChild(index);
    let addrFromXPubKey = await b58c.encode({
      version: myCoin.pubKeyHashVersion,
      pubKeyHash: derivedChild.pubKeyHash.toString("hex"),

coolaj86

BogoSort Wins!

The for loop outperforms the binary search.
(well, the forEach did, until we changed the caching in an unrelated function)

And the optimizations have (almost) nothing to do with the algorithms.
(changing any code can massively change the performance of any benchmark)

See the video: https://www.youtube.com/watch?v=0mmi44ZB2C0

coolaj86

K&R Style

• Required (but not enforced) for JavaScript
• Required AND enforced for Go

function kr() {
  // foo
}

coolaj86

Why is this bad?

const something = (params) => value;

TL;DR Because this ISN'T bad

coolaj86

Pyramids of Doom in JavaScript

• callbacks
• promises
• try/catch

solution: await.catch

async function doLotsOfStuff(params) {

coolaj86

{
  "hash": "716dac449e7f7c02a69cb0ec37ea3b905f3f01aac707c1db2dd38bfa9deedb20",
  "version": 1,
  "inputs": [
    {
      "prevTxId": "ea5ff9b5bdcce85e3bbdf7c0d84a63681265cf630481bc48e6009d7b5dbad069",
      "outputIndex": 1,
      "sequenceNumber": 4294967295,
      "script": "47304402207a88c7eb79b5307288e73bce404963cb3d8f76db7a3fa3924ab633cfbaf9f49d02204de2d008ce5b38383ad9dbfccaa99bed880fe949166ce2d7ce0c6034b52177070121038081fa1a40e680a1e7ea5be15895669b1a6fb5ea884dabbfa688e7097d0d9011",
      "scriptString": "71 0x304402207a88c7eb79b5307288e73bce404963cb3d8f76db7a3fa3924ab633cfbaf9f49d02204de2d008ce5b38383ad9dbfccaa99bed880fe949166ce2d7ce0c6034b521770701 33 0x038081fa1a40e680a1e7ea5be15895669b1a6fb5ea884dabbfa688e7097d0d9011"

coolaj86

(Ab)uses of Reduce

Watch: https://youtu.be/X85QdHe1tDw?t=409 (Starts at 6:50)

reduce is one of the least understood Array methods, and probably the most abused.

Use #0 Actually Reducing

Reducing, as the good functional programmer intended:

coolaj86

Can you convert a Value Object (i.e. JSON) to YAML?

My Expectation

My expectation is that a mid-senior level developer should be able to accomplish this task in under 4 hours:

• write a basic solution in 30-60 minutes
• debug / test / fine-tune for another 60-90 minutes
• document, package, and publish within the remaining time

I could see it reasonable even if a developer takes 2 days to solve this for the first time - possibly because they needed to take a break but didn't, such as feeling pressure and getting brain freeze, taking an initial wrong approach, over-thinking the problem, and then just keeping beating the dead horse, etc - especially if they've never had to work independently before.

coolaj86

The Developer Litmus Test

I used to think that fizzbuzz was a stupid way to test if someone is a developer.

Now I think that it's a good litmus test.

However, I think I've found a better one...