feat: single-pass hydration

[Rich-Harris]

I'm a firm believer that hydration is a better way to make a server-rendered app interactive than the various contortions people perform in order to avoid it, as long as you can do it fast.

This PR reduces DOM size and significantly improves hydration speed (as in, like, 40% faster).

The key insight is that node.nextSibling is (surprisingly!) the most expensive part of hydration. Right now, we're doing that twice: once to build up the hydrate_nodes array, and once as we create effects and add event listeners. In this PR, we skip the first part — we keep a single hydrate_node updated as we hydrate the application.

One neat side-effect of this change is that each blocks no longer need hydration comments for each item, just for the each block itself.

There's a fair bit of tidying up that needs to happen, and possibly some room for further optimisations.

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.

Tests and linting

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

[trueadm]

There's a lot going on here, but it looks good to me.

[Rich-Harris]

Just realised hydration boundaries are unbalanced in some cases which could lead to incorrect behaviour in the case of mismatches — need to add a couple more tests

[vipero07]

#12422 I think is related to this change.