Debounce the layout function in the list component

[YonatanKra]

when removing an item.

Closes #1927

See my remarks blow regarding the outcome of this performance improvement, and proposed changes to this PR.

[YonatanKra]

Performance gains

The following measurements were taken for the test (test:debug, http://localhost:9876/debug.html?grep=mwc-list%3A%20mwc-list%20performance%20issue) before and after applying the change.
Before:

After:

Can see from the image that the remove process took a roughly 1/9 of the time after the change.

[YonatanKra]

Caveat: The debounced updates are no longer part of the lists'updateComplete hook.

Can try to add it to the _getUpdateComplete somehow if needs be.

Another solution can be to allow the user to add a flag asyncRemoval or syncRemoval, and commence the layout accordingly.

Let me know if and how you'd like to proceed with this one.

[e111077]

please update CHANGELOG.md

[e111077]

super thanks for the PR!

[e111077]

This seems like a simple enough implementation so let's go with this for now. Part of our code submission process includes importing this code into google's monorepo, and google-wide tests should help us run into migration issues other users may also have; this is where we can determine whether we need a flag. Though hooking into getUpdateComplete might be a requirement

[YonatanKra]

@e111077
Thanks for the comments. I've addressed all of them.
Let me know if something's missing.

[e111077]

final stretch! Great work so far!

[YonatanKra]

Resolved all mentioned issues 😀

[e111077]

I'm also debugging a failure in a YouTube test that uses select. A little afraid it may be caused by the debounce vs a buffer, but it's still too early to tell

[YonatanKra]

I'm also debugging a failure in a YouTube test that uses select. A little afraid it may be caused by the debounce vs a buffer, but it's still too early to tell

It is a breaking change, as it changes the interface (from sync to async). So it might be this test assumes the remove event should have sync effect.

[e111077]

Haven't forgotten about this; still trying to untangle their esoteric code