julianjensen

/**
 * @param {string} url              - image url
 * @param {HTMLImageElement} [img]  - optional pre-existing image element
 * @param {HTMLElement} [parent]    - optional parent element for creation of new image
 * @return {Promise<any>}
 */
const addImage = ( url, img, parent ) => new Promise( resolve => {
    if ( parent != null ) parent.appendChild( img = document.createElement( 'img' ) );
    img.onload = () => { img.onload = null; resolve(); }
    img.src = url;

julianjensen

/**
 * This is a functional component that uses hooks and `redux` and duplicates the functionality
 * of our more traditional component. Without the `redux` `connect` function, we'd have to
 * wrap every component in a `useMemo()` to achieve the same result which I haven't done here.
 */
const StandardComponent = ({ userName, greeting = 'Hello', answeredItAlready, setOkay }) => {
    // Added to show that hooks work with `redux` state. The hook must be in all
    // execution paths and must be in the `render` function. If we want to move it elsewhere
    // then we'd have to make a custom hook that contained other hooks, and call the custom hook
    // from here. Very messy.

julianjensen

/*
 * Scrolls element like absolute until some scollOffset is reached then becomes sticky
 */
#toc {
	position: fixed;
	top: 11em;
}

@supports (top: max(1em, 1px)) {
	#toc {

julianjensen

/** ******************************************************************************************************************
 * @file A very basic HTTP server that supports standard middleware and express-style URLs.
 * @date 07-Jun-2020
 *********************************************************************************************************************/
"use strict";

async function MinimalServer( { PROTOCOL = 'http', host = '127.0.0.1', port = 3000 } )
{
    const protocol = await import( PROTOCOL );
    const protoMethods = protocol.METHODS.map( m => m.toLowerCase() );

julianjensen

Some quick assembly language notes required for the following. You'll need to understand three registers on the Intel CPU.

First, the rip register. On a 16-bit CPU, it's called IP (for Instruction Pointer), and on a 32-bit CPU, it's called eip (for extended instruction pointer), but on a 64-bit modern processor, it's called rip. It holds the index (or offset or number) of the memory location where the CPU is currently executing instructions from. As it reads instructions from memory, it increments to get to the next instruction and so on. Some instructions can cause it to jump to a non-sequential read, such as a branch, jump, call, or return, for example. You can also write to it directly but this is rarely done in practice.

Second, the rsp register (also sp on 16-bit and esp on 32-bit) is the stack pointer. It points to the top of the stack. The stack grows downward. In other words, if the rsp points to memory address 10000 and you push a 32-bit number (4 bytes) onto the stack, then afterward

julianjensen

Permutations

This problem is not that hard if you approach it the easy way, think about it "correctly." If not, then it takes a surprisingly long time to get right. Or maybe I'm just dense. The solution is quite straightforward and after I wrote it, I was surprised that it had taken me the better part of an hour to get completely right and polished. You may be able to crank it out in a few minutes.

You're given an array of whatever values. Some, all, or none of these values can, in turn, be another array, and so on.

julianjensen

// You don't need to a `split("")` here, you can use a string with an index, i.e. `lookupTable[n]`
// Also, while I recognize that the spread operator is slow, I'd be remiss if I didn't mention `[..."ABCD...89+/]`
const lookupTable = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/".split(
    ""  // Why is this on a line by itself? It's confusing. A more readable `.split( "" )` is better
);

// Or `[...lookupTable].reverse()` a bit more succinct and terse, shows your intent by using `.reverse()`
const reverseLookupTable = lookupTable.reduce((acc, char, idx) => {
    acc[char] = idx;
    return acc;

julianjensen

/** ******************************************************************************************************************
 * @file This just counts the number of vowels in some text.
 *
 * To install and play with this, do the following:
 *
 * 1. `mkdir vowel-count`
 * 2. `cd vowel-count`
 * 3. `npm init`
 * 4. Copy this file into the directory, call it anything you like, but `index.js` is normal.
 * 5. `npm i benchmark cli-table2`

julianjensen

Node.js version 7.9, Apr. 2017                                                       Node.js version 9.9, Mar. 2018
  Native .fill() vs fast.fill() (3 items)                                              Native .fill() vs fast.fill() (3 items)
    ✓  Array.prototype.fill() x 26,557,003 ops/sec ±1.88% (90 runs sampled)              ✓  Array.prototype.fill() x 31,067,272 ops/sec ±1.43% (90 runs sampled)
    ✓  Vector.fill() x 83,198,767 ops/sec ±4.40% (84 runs sampled)                       ✓  Vector.fill() x 147,385,123 ops/sec ±1.34% (89 runs sampled)
    ✓  fast.fill() x 83,882,082 ops/sec ±5.03% (84 runs sampled)                         ✓  fast.fill() x 140,299,152 ops/sec ±1.53% (88 runs sampled)

    Result: fast.js is 0.00% faster than Vector.fill().                                  Result: fast.js is 0.00% faster than Array.prototype.fill().

Node.js version 7.9, Apr. 2017                                                       Node.js version 9.9, Mar. 2018
  Native .fill() vs fast.fill() (10 items)      

julianjensen

const
    // Direct
    sqr5 = Math.sqrt( 5 ),
    sqr5l = ( 1 + sqr5 ) / 2,
    sqr5r = ( 1 - sqr5 ) / 2,
    dfib = n => n < 2 ? n : ( ( Math.pow( sqr5l, n ) - Math.pow( sqr5r, n ) ) / sqr5 ),
    // ES6
    fib = ( ( pp = 0, p = 0 ) => () => p = pp + ( ( pp = p ) || 1 ) )(),
    // ES5.1
    ffib = ( function( pp, p ) { pp = p = 0; return function() { return p = pp + ( ( pp = p ) || 1 ); }; } )();