fix: cursor jumps in input two way binding

[fabhari]

Old Fix: Restore input binding selection position (#14649) .
Current Fix: Remove unnecessary cursor manipulation written for (#14649) as the presence of runes no longer requires special handling.

This fix closes #16634

Before submitting the PR, please make sure you do the following

• It's really useful if your PR references an issue where it is discussed ahead of time. In many cases, features are absent for a reason. For large changes, please create an RFC: https://github.com/sveltejs/rfcs
• Prefix your PR title with feat:, fix:, chore:, or docs:.
• This message body should clearly illustrate what problems it solves.
• Ideally, include a test that fails without this PR but passes with it.
• If this PR changes code within packages/svelte/src, add a changeset (npx changeset).

Tests and linting

• Run the tests with pnpm test and lint the project with pnpm lint

[changeset-bot]

🦋 Changeset detected

Latest commit: 0357b5f

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
svelte	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

Playground

pnpm add https://pkg.pr.new/svelte@16649

[7nik]

This breaks a case like this (compare to the latest version)

<input bind:value={() => name, v => name = v.toUpperCase()} />

To elaborate: the input and the variable get desynced.

[fabhari]

Hey @7nik Thanks for highlighting that.
hope have fixed the issue by syncing the set and get callbacks using promises to always execute in an order .had to stop get cb excuting before set cb which led to desync

[svelte-docs-bot]

Preview: https://svelte-dev-git-preview-svelte-16649-svelte.vercel.app/

[7nik]

So, now the input and the variable get desynced because of the "Never rewrite the contents of a focused input" case. But don't we still want to write if it's the latest batch?

[Rich-Harris]

Ah whoops, you're right. So yes, we do need to handle it in the event handler and not in the subsequent effect, and the reason we can't do it immediately is because of a quirk of how each blocks work — the item is a source that gets updated during flush rather than a derived, which means an immediate get may not reflect the current state.

Updated the PR to make that explicit.

Thanks!

[Rich-Harris]

gah the test failure is bizarre. Passes fine when the test runs individually but not when the entire suite runs?

[fabhari]

Hi @Rich-Harris , quick question out of curiosity.

I noticed the latest commit you have changed is the callback method to be async & waiting for requestAnimationFrame().

I am wondering how this approach differs from my simpler solution which waits for SET synchronous callback to end first and the call GET. What potential drawbacks my solution might have. ?

Could you help me understand the benefits and perhaps suggest an even better way to handle this?
Hey @7nik @Rich-Harris can anyone help me understand this difference in handling or negative impact my solution has ?

[7nik]

.then(...) is always executed in a new microtask. But in async Svelte, settling reactivity may take multiple microtasks. This is why tick() should be used.

[fabhari]

great. thanks for the info @7nik