JavaScript has two globally available metods for working with Base64: atob() and btoa(). Their names clearly look like they were borrowed from older languages. In C, for instance, the standard library includes functions like atoi and atof:

#include <stdlib.h>

char str[] = "123";
int num = atoi(str); // 123

atoi means ASCII to integer, and atof means ASCII to float (though in reality it returns a double).

So what do you think the atob function does in JavaScript? ASCII to Base64? In other words, converting a regular string into a Base64 string? Nope! It does the exact opposite: it converts a Base64 string into a "regular" string. And btoa, in turn, converts a regular string into Base64!

console.log(btoa('Famabara')); // 'RmFtYWJhcmE='
console.log(atob('RmFtYWJhcmE')); // 'Famabara'

Who thought it was a good idea to swap the names like that? Love JS.

If you try to use Quill Editor with Nuxt 3 when rendering a page in SSR you will get this error:
That means NodeJS doesn't have the global variable 'document'. Because SSR rendering is executed in
a NodeJS environment, not a browser.

The sad fact: <client-only> won't help with this problem. The problem with quill's code is that during import, it assumes it's being executed in the browser. I hope you remember that the code imported from the module is not just imported, but executed, i.e. the authors of Quill wrote it so that the 'document' object is immediately accessed there. Very bad.

One solution is to disable SSR, but its an awful solution. But the second solution is to use dynamic JS imports.

My <script lang="ts" setup> in QuillEditor.vue in Nuxt 3 project:

import 'quill/dist/quill.core.css';
import 'quill/dist/quill.snow.css';
// import Quill from 'quill'; // SSR problem with 'document', need to use dynamic import
import type Quill from 'quill';

const elemForQuillEditor = ref<HTMLDivElement|null>(null);

let editor: null|Quill = null; // Do not store in ref() - it causes bugs!

// ...

async function createQuillAndSetListeners() {
  const Quill = (await import('quill')).default; // This is most important thing - usage of JS dynamic import

  if (!elemForQuillEditor.value) return;
  editor = new Quill(elemForQuillEditor.value, {
    theme: 'snow',
    modules: {
      history: {
        delay: 2000,
        maxStack: 500,
        userOnly: true,
      },
      toolbar: {
        container: [
          ['bold', 'italic', 'underline', 'strike'],
          ['link'],
          [{ list: 'ordered' }, { list: 'bullet' }],
          [{ script: 'sub' }, { script: 'super' }],
          [{ header: [1, 2, 3, 4, 5, 6, false] }],
          [{ color: [] }, { background: [] }],
          ['clean'],
          ['undo', 'redo'],
        ],
        handlers: {
          undo() {
            editor?.history.undo();
          },
          redo() {
            editor?.history.redo();
          },
        },
      },
    },
    placeholder: props.placeholder,
  });
  
  editor.on('text-change', () => {
    if (!editor) return;
    // ... my other code
  });
}

// ...

onMounted(createQuillAndSetListeners);

Now your Quill editor will work in Nuxt 3 even during SSR!

A man has five children - all girls. He dreams of a son, an heir! Now his wife is pregnant again. Let's skip the psychology behind all this and focus on the probabilities.

Emotionally, it feels like after five girls in a row, a boy is due. Surely, the odds of finally having a son are much higher now.

But this is a textbook example of the gambler's fallacy - the false belief that if something happens repeatedly, the opposite is bound to happen soon. That's not how probability works.

According to math, the chance of having exactly five girls in a row followed by another girl is 1 in 64, or about 1.6%. But those five girls are already here. The only thing that matters now is the next birth.

And the odds of having a boy? Still 50/50. Every birth is an independent event (assuming randomness), so the previous outcomes don't change the next one.

So yes, there's still a 50% chance they'll have another girl.