HTML preprocessors can make writing HTML more powerful or convenient. For instance, Markdown is designed to be easier to write and read for text documents and you could write a loop in Pug.
In CodePen, whatever you write in the HTML editor is what goes within the <body>
tags in a basic HTML5 template. So you don't have access to higher-up elements like the <html>
tag. If you want to add classes there that can affect the whole document, this is the place to do it.
In CodePen, whatever you write in the HTML editor is what goes within the <body>
tags in a basic HTML5 template. If you need things in the <head>
of the document, put that code here.
The resource you are linking to is using the 'http' protocol, which may not work when the browser is using https.
CSS preprocessors help make authoring CSS easier. All of them offer things like variables and mixins to provide convenient abstractions.
It's a common practice to apply CSS to a page that styles elements such that they are consistent across all browsers. We offer two of the most popular choices: normalize.css and a reset. Or, choose Neither and nothing will be applied.
To get the best cross-browser support, it is a common practice to apply vendor prefixes to CSS properties and values that require them to work. For instance -webkit-
or -moz-
.
We offer two popular choices: Autoprefixer (which processes your CSS server-side) and -prefix-free (which applies prefixes via a script, client-side).
Any URLs added here will be added as <link>
s in order, and before the CSS in the editor. You can use the CSS from another Pen by using its URL and the proper URL extension.
You can apply CSS to your Pen from any stylesheet on the web. Just put a URL to it here and we'll apply it, in the order you have them, before the CSS in the Pen itself.
You can also link to another Pen here (use the .css
URL Extension) and we'll pull the CSS from that Pen and include it. If it's using a matching preprocessor, use the appropriate URL Extension and we'll combine the code before preprocessing, so you can use the linked Pen as a true dependency.
JavaScript preprocessors can help make authoring JavaScript easier and more convenient.
Babel includes JSX processing.
Any URL's added here will be added as <script>
s in order, and run before the JavaScript in the editor. You can use the URL of any other Pen and it will include the JavaScript from that Pen.
You can apply a script from anywhere on the web to your Pen. Just put a URL to it here and we'll add it, in the order you have them, before the JavaScript in the Pen itself.
If the script you link to has the file extension of a preprocessor, we'll attempt to process it before applying.
You can also link to another Pen here, and we'll pull the JavaScript from that Pen and include it. If it's using a matching preprocessor, we'll combine the code before preprocessing, so you can use the linked Pen as a true dependency.
Search for and use JavaScript packages from npm here. By selecting a package, an import
statement will be added to the top of the JavaScript editor for this package.
Using packages here is powered by esm.sh, which makes packages from npm not only available on a CDN, but prepares them for native JavaScript ESM usage.
All packages are different, so refer to their docs for how they work.
If you're using React / ReactDOM, make sure to turn on Babel for the JSX processing.
If active, Pens will autosave every 30 seconds after being saved once.
If enabled, the preview panel updates automatically as you code. If disabled, use the "Run" button to update.
If enabled, your code will be formatted when you actively save your Pen. Note: your code becomes un-folded during formatting.
Visit your global Editor Settings.
<!DOCTYPE html>
<html>
<meta name="charset" content="utf-8" />
<body>
<style>
body {
font-family: "Helvetica Neue", sans-serif;
}
ul {
list-style-type: none;
padding-left : 0;
user-select: none;
-webkit-user-select: none;
}
ul li {
margin-bottom: 1px;
background: rgb(240,240,240);
padding: 3px;
display: block;
color: black;
}
ul li.selected {
background: rgb(150,150,255);
color: white;
}
tt {
display: inline-block;
font-size: 1.05rem;
color: red;
background: rgb(250,230,230);
padding: 1px 2px;
border: 1px solid;
border-radius: 2px;
}
.test-good, .test-bad {
font-family: monospace;
border: 1px solid;
border-left: 5px solid;
padding: 2px 5px;
margin: 1px 0;
}
.test-good {
color: green;
background: rgb(220,255,220);
}
.test-bad {
color: red;
background: rgb(255,220,220);
}
</style>
<h1>Let's talk list selections.</h1>
<p>Rules by which the Finder seems to do it seem rather convoluted. They revolve around two items of the list:</p>
<ol>
<li><tt>target</tt> - the item getting clicked now</li>
<li><tt>anchor</tt> - the item that was clicked during the previous action</li>
</ol>
<p>Funny but my made-up <tt>anchor</tt> terminology seems to match the Webkit source - <a href="https://github.com/WebKit/WebKit/blob/5d2e5ba9bf14947a90b4ecfe9ef2249508968e20/Source/WebCore/html/HTMLSelectElement.cpp#L645">see here.</a></p>
<p>Effectively this means that the list selection actions are <b>stateful</b> - it is not possible to implement them just using the list of currently selected indices and the held shortcut keys, you have to carry around the index of the anchor item as well.</p>
<script id="entire">
const MODE_SINGLE = Symbol('single');
const MODE_CONTINUOUS = Symbol('cont');
const MODE_DISCONTINUOUS = Symbol('discont');
function makeSetWithRange(fromNumber, toNumber) {
if (toNumber < fromNumber) {
[fromNumber, toNumber] = [toNumber, fromNumber];
}
const numElements = toNumber - fromNumber + 1;
const increments = Array.from(Array(numElements).keys());
const itemsArr = increments.map((n) => fromNumber + n);
return new Set(itemsArr);
}
function setUnionOf(iterator1, iterator2) {
return new Set([...iterator1, ...iterator2]);
}
function listReselectModeFromEvent(event) {
return event.metaKey ? MODE_DISCONTINUOUS : (event.shiftKey ? MODE_CONTINUOUS : MODE_SINGLE);
}
// If mode is "single", select only that item and reset the anchor to
// that item
function listReselectSingle(targetIndex) {
return [targetIndex, new Set([targetIndex])];
}
// If mode is "command", change the selection state of the given item.
// The command-selected item becomes the new anchor for the subsequent selection
function listReselectDiscontinuous(selectedIndicesSet, targetIndex) {
// Toggle the membership in the set and set the anchor index
// to the clicked element
if (selectedIndicesSet.has(targetIndex)) {
selectedIndicesSet.delete(targetIndex);
} else {
selectedIndicesSet.add(targetIndex);
}
return [targetIndex, selectedIndicesSet];
}
// If the mode is "shift", there is some work to be done.
function listReselectContinuous(selectedIndicesSet, anchorIndex, targetIndex) {
// This is a special case. When nothing is selected shift+select
// starts from 0 and resets the anchor to the target index.
if (selectedIndicesSet.size === 0) {
return [targetIndex, makeSetWithRange(0, targetIndex)]
}
// Unmark selected elements adjacent to the anchor in both directions,
// and stop at the first element which is not selected.
// Worst performance is a single linear scan of
// the entire set.
let [left, right] = [anchorIndex - 1, anchorIndex + 1];
while (selectedIndicesSet.delete(left)) {
left--;
}
while (selectedIndicesSet.delete(right)) {
right++;
}
// Reselect the elements between the anchor and the target,
// and add the remaining elements which already were selected.
const freshlySelected = makeSetWithRange(anchorIndex, targetIndex);
const union = setUnionOf(freshlySelected, selectedIndicesSet);
return [anchorIndex, union];
}
// Returns the anchorIndex for the next listReselect call,
// and a fresh Set of selected indices.
function listReselect({selectedIndices, mode, anchorIndex, targetIndex}) {
// Make sure we never mutate the set/Array given to us
const selectedIndicesSet = new Set([...selectedIndices]);
switch (mode) {
case MODE_SINGLE:
return listReselectSingle(targetIndex);
case MODE_CONTINUOUS:
return listReselectContinuous(selectedIndicesSet, parseInt(anchorIndex) || 0, targetIndex);
case MODE_DISCONTINUOUS:
return listReselectDiscontinuous(selectedIndicesSet, targetIndex);
default:
throw new Error(`Unknown mode ${mode}`);
}
}
</script>
<h2>Try it out</h2>
<p>Try Command+ and Shift+ selecting them.</p>
<ul id='groceries'>
<li>Apples</li>
<li>Oranges</li>
<li>Cucumbers</li>
<li>Kiwis</li>
<li>Peaches</li>
<li>Plums</li>
<li>Grapes</li>
<li>Lettuce</li>
<li>Cabbages</li>
<li>Aubergines</li>
<li>Beetroots</li>
<li>Potatoes</li>
<li>Tomatoes</li>
<li>Strawberries</li>
<li>Almonds</li>
<li>Blackberries</li>
<li>Lemons</li>
<li>Watermelons</li>
<li>Zucchinis</li>
<li>Pomelos</li>
</ul>
<script>
const ul = document.querySelector('ul#groceries');
const lis = ul.querySelectorAll('li');
let anchorIndex = null; // The second bit of state carried around
let selectedIndices = [];
ul.addEventListener('mousedown', (event) => {
const clickedListItem = event.target.closest('li');
if (!clickedListItem) return;
const allSiblings = Array.from(clickedListItem.parentNode.querySelectorAll('li'));
const targetIndex = allSiblings.indexOf(clickedListItem);
let selectedIndices = allSiblings.map((el, i) => {
if (el.classList.contains("selected")) {
return i;
} else {
return null;
}
}).filter((idxOrNull) => idxOrNull !== null);
// Update selection state
const mode = listReselectModeFromEvent(event);
[anchorIndex, selectedIndices] = listReselect({
mode,
selectedIndices,
anchorIndex,
targetIndex
});
// Remove "selected" class on all indices
allSiblings.map((el) => el.classList.remove('selected'));
selectedIndices.forEach((selectedIndex) => allSiblings[selectedIndex].classList.add('selected'));
});
</script>
<h2>HTML multiselects</h2>
<p>Multiselects work similarly. There is one important difference though -
in Finder dragging selected items triggers drag and drop behavior,
but multiselects support drag-select.</p>
<select multiple style='height: 5em'>
<option>Apples</option>
<option>Oranges</option>
<option>Cucumbers</option>
<option>Kiwis</option>
<option>Peaches</option>
<option>Plums</option>
<option>Grapes</option>
<option>Lettuce</option>
<option>Cabbages</option>
<option>Aubergines</option>
<option>Beetroots</option>
<option>Potatoes</option>
<option>Tomatoes</option>
<option>Strawberries</option>
<option>Almonds</option>
<option>Blackberries</option>
<option>Lemons</option>
<option>Watermelons</option>
<option>Zucchinis</option>
<option>Pomelos</option>
</select>
<h2>Observed behavior</h2>
<ol>
<li>Click select unselects all items and makes the <tt>target</tt> selected, and sets it as the <tt>anchor</tt> for the next operation</li>
<li>⌘+ click flips the selected state of <tt>target</tt> and sets it as <tt>anchor</tt> for the next operation</li>
<li>⇧+ click with nothing selected selects from the first item to <tt>target</tt>, always, and sets the <tt>target</tt> to be the anchor for the next operation.</li>
<li>⇧+ click from an existing <tt>anchor</tt> to the <tt>target</tt> makes all items between <tt>anchor</tt> and <tt>target</tt> selected, except for the <tt>anchor</tt> - it keeps the selection state.</li>
<li>⇧+ click to an item further down the list from an unselected anchor will cause all items between anchor and target, including both, to be selected</li>
<li>⇧+ click to an item higher up the list from an <b>selected</b> <tt>anchor</tt> will cause all items between the target and the anchor, including both, to be selected, <b>as well as an item beyound the anchor.</b>
<li>⇧+ click to an item higher up the list from an <b>unselected</b> anchor will cause all items between the target and the anchor, including both, to be selected, <b>as well as one item further down from the anchor. 🤷♂️</b>
</ol>
<p>If this seems complex or intricate: it is. UI is hard.</p>
<h2>Implementation</h2>
<p>It is possible to implement this in a functional way, as a transformation function of the following shape:</p>
<code><pre>
selectionMode = Single | Continuous | Discontinuous
anchorIndex = ItemIndex | nil
selectedIndices = [ItemIndex] # A list or Set of indices of already selected items
update(selectedIndices, anchorIndex, targetIndex, mode) => [newAnchorIndex, newSelectedIndices]
</pre></code>
<p>On subsequent calls, the <tt>anchorIndex</tt> needs to be passed along since it is one
of the two components of state.</p>
<p>There is some HTML manipulation involved in the demo but the function is not specific to React, Stimulus
or even something vaguely ML-like - it is fairly generic. Should make a decent Redux reducer too.</p>
</body>
<h2>Combination with disabled checkboxes</h2>
<p>If you want to have a combination with checkboxes and checkboxes can be disabled, your best bet would be
to treat <tt>disabled</tt> checkboxes as indices that can be used for multiselection, but to not
highlight them and to filter them out from the result before applying it to the data model. Specifically:
</p>
<code><pre>
const mode = listReselectModeFromEvent(evt);
[anchorIndex, selectedIndices] = listReselect({mode, anchorIndex, selectedIndices, targetIndex})
allSiblings.map((container, i) => {
const checkbox = container.querySelector('input[type="checkbox"]');
if (!checkbox.disabled) {
checkbox.checked = selectedIndices.has(i);
}
});
</pre></code>
Like here:
<ul id='checkboxes'>
<li><input type="checkbox"> Apples</li>
<li><input type="checkbox"> Oranges</li>
<li><input type="checkbox" disabled> Cabbages (sold out)</li>
<li><input type="checkbox"> Almonds</li>
<li><input type="checkbox"> Blackberries</li>
<li><input type="checkbox"> Lemons</li>
<li><input type="checkbox"> Watermelons</li>
</ul>
<script>
(function n() {
const ul = document.querySelector('ul#checkboxes');
const lis = ul.querySelectorAll('li');
let anchorIndex = null; // The second bit of state carried around
let selectedIndices = [];
ul.addEventListener('mousedown', (event) => {
const clickedListItem = event.target.closest('li');
if (!clickedListItem) return;
const allSiblings = Array.from(clickedListItem.parentNode.querySelectorAll('li'));
const targetIndex = allSiblings.indexOf(clickedListItem);
let selectedIndices = allSiblings.map((el, i) => {
if (el.classList.contains("selected")) {
return i;
} else {
return null;
}
}).filter((idxOrNull) => idxOrNull !== null);
// Update selection state
const mode = listReselectModeFromEvent(event);
[anchorIndex, selectedIndices] = listReselect({
mode,
selectedIndices,
anchorIndex,
targetIndex
});
// Remove "selected" class on all indices
allSiblings.map((el) => el.classList.remove('selected'));
// ...set "selected" class on selected items. Also on the one with
// the checkbox since we use the "selected" class as a state marker -
// but that is not a requirement.
selectedIndices.forEach((selectedIndex) => allSiblings[selectedIndex].classList.add('selected'));
// Update checkbox state
allSiblings.map((container, i) => {
const checkbox = container.querySelector('input[type="checkbox"]');
if (!checkbox.disabled) {
checkbox.checked = selectedIndices.has(i);
}
});
event.preventDefault();
});
})();
</script>
<h2>Tests</h2>
<p>These only test the selection state change function, not the DOM manipulation.</p>
<script>
function numericSort(a, b) {
return a - b;
}
// https://stackoverflow.com/a/46491780
function toJsonWithSetSupport(key, value) {
if (typeof value === 'object' && value instanceof Set) {
return [...value].sort(numericSort);
}
return value;
}
function assertTransition(optionsForListReselect, expectedOutcome) {
const result = listReselect(optionsForListReselect);
const actual = JSON.stringify(result, toJsonWithSetSupport);
const expected = JSON.stringify(expectedOutcome, toJsonWithSetSupport);
const input = JSON.stringify(optionsForListReselect, toJsonWithSetSupport);
const isGood = actual === expected;
const cssClass = isGood ? 'test-good' : 'test-bad';
const visualComparison = isGood ? `Was: ${expected}` : `Must: ${expected} was ${actual}`;
document.write(`<div class='${cssClass}'>${input}<br />${visualComparison}</div>`);
return result;
}
</script>
<h3>Single selection</h3>
<script>
assertTransition(
{mode: MODE_SINGLE, selectedIndices: [], anchorIndex: null, targetIndex: 0},
[0, [0]]
);
assertTransition(
{mode: MODE_SINGLE, selectedIndices: [], anchorIndex: null, targetIndex: 5},
[5, [5]]
);
assertTransition(
{mode: MODE_SINGLE, selectedIndices: [5], anchorIndex: 5, targetIndex: 5},
[5, [5]]
);
assertTransition(
{mode: MODE_SINGLE, selectedIndices: [0,1,2,3], anchorIndex: null, targetIndex: 1},
[1, [1]]
);
assertTransition(
{mode: MODE_SINGLE, selectedIndices: [0,1,2,3], anchorIndex: 3, targetIndex: 1},
[1, [1]]
);
assertTransition(
{mode: MODE_SINGLE, selectedIndices: [0,1,2,3], anchorIndex: 7, targetIndex: 1},
[1, [1]]
);
</script>
<h3>Discontinuous selection</h3>
<script>
assertTransition(
{mode: MODE_DISCONTINUOUS, selectedIndices: [], anchorIndex: null, targetIndex: 1},
[1, [1]]
);
assertTransition(
{mode: MODE_DISCONTINUOUS, selectedIndices: [1], anchorIndex: null, targetIndex: 1},
[1, []]
);
assertTransition(
{mode: MODE_DISCONTINUOUS, selectedIndices: [0,1,2,3], anchorIndex: null, targetIndex: 1},
[1, [0,2,3]]
);
assertTransition(
{mode: MODE_DISCONTINUOUS, selectedIndices: [0,1,2,3], anchorIndex: null, targetIndex: 8},
[8, [0,1,2,3,8]]
);
assertTransition(
{mode: MODE_DISCONTINUOUS, selectedIndices: [0,1,2,3], anchorIndex: 0, targetIndex: 0},
[0, [1,2,3]]
);
</script>
<h3>Continuous selection (this is where it gets tricky)</h3>
<script>
let mode = MODE_CONTINUOUS;
assertTransition(
{mode, selectedIndices: [], anchorIndex: null, targetIndex: 3},
[3, [0,1,2,3]]
);
assertTransition(
{mode, selectedIndices: [0,1,], anchorIndex: 0, targetIndex: 3},
[0, [0,1,2,3]]
);
assertTransition(
{mode, selectedIndices: [0,1,2,3], anchorIndex: 0, targetIndex: 1},
[0, [0,1]]
);
assertTransition(
{mode, selectedIndices: [0,2,5,6], anchorIndex: 0, targetIndex: 7},
[0, [0,1,2,3,4,5,6,7]]
);
assertTransition(
{mode, selectedIndices: [0,2], anchorIndex: 2, targetIndex: 4},
[2, [0,2,3,4]]
);
assertTransition(
{mode, selectedIndices: [0,1,2,3], anchorIndex: 0, targetIndex: 1},
[0, [0,1,]]
);
assertTransition(
{mode, selectedIndices: [0,1,2,5], anchorIndex: 0, targetIndex: 3},
[0, [0,1,2,3,5]]
);
assertTransition(
{mode, selectedIndices: [0,2,4], anchorIndex: 4, targetIndex: 2},
[4, [0,2,3,4]]
)
// Weird behavior which happens when you Cmd+deselect an item in-between, then Cmd+reselect it
// and then Shift+select outside that item
assertTransition(
{mode, selectedIndices: [2,3,4,5], anchorIndex: 3, targetIndex: 1},
[3, [1,2,3]]
);
</script>
<h2>Show me the code!</h2>
<p>There is less code than one would think. But it took a long wile to get to it. See here:</p>
<script>
const scriptContents = document.querySelector("script#entire").text;
document.write("<pre><code>" + scriptContents + "</pre></code>");
</script>
</html>
Also see: Tab Triggers