Preact: Element.remove() is unsupported in Internet Explorer

[KrofDrakula]

While testing 8.0-beta in IE11 I've run into a problem in unmountComponent() calling .remove() on base:

https://github.com/developit/preact/blob/8/src/vdom/component.js#L266

Changing that line to:

base.parentNode.removeChild(base);

...fixes the problem. I'm not sure if 8.0 is supposed to support IE, but worth keeping in mind.

[KrofDrakula]

Other instances of this problem:

• https://github.com/developit/preact/blob/8/src/vdom/diff.js#L220
• https://github.com/developit/preact/blob/8/src/vdom/diff.js#L259

[KrofDrakula]

If anyone's trying to work around this problem, you can apply a minimum polyfill for the required methods:

(function() {
  function remove() { this.parentNode && this.parentNode.removeChild(this); }
  if (!Element.prototype.remove) Element.prototype.remove = remove;
  if (Text && !Text.prototype.remove) Text.prototype.remove = remove;
})();

[robertknight]

I'm not sure if 8.0 is supposed to support IE, but worth keeping in mind.

@developit knows the official status but I'd be tempted to suggest that we drop IE 8 support out of the box if it is starting to conflict with the size/simplicity goals of the project.

[KrofDrakula]

This isn't IE8 support, this is IE11 using preact 8.0-beta.

[NekR]

I'm pretty sure IE8 never was supported by Preact. React does support it though.

[developit]

Ah - this was totally a slip of the mind. .remove() is faster than accessing .parentNode.removeChild(), but it looks like we'll have to bite the bullet.

[KrofDrakula]

Hopefully as IE goes the way of the dodo, it will be a thing of the past, as everyone else already supports the living standard:

https://developer.mozilla.org/en-US/docs/Web/API/ChildNode/remove#Browser_compatibility

[developit]

@KrofDrakula Definitely something to keep on the radar. If this was only in IE9 I'd be tempted to let people polyfill, but 11 is just a bit too far inside the support target.

[developit]

Fix released in 8.0.1.

[ascorbic]

This seems to have returned in 10. Should we polyfill now?

[marvinhagemeister]

@ascorbic thanks for the pointer! Made a PR to fix it again 👍