[fix] reading from same response body twice during prerender (#3473)

[Theo-Steiner]

Please don't delete this checklist! 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
• This message body should clearly illustrate what problems it solves.
• Ideally, include a test that fails without this PR but passes with it.

This PR fixes the behavior described in #3473, where building the default template with adapter static fails as during the pre-rendering the same response body was read twice. As @bluwy pointed out in the original issue, this is related to fetch (and accordingly node-fetch) only allowing a single read of a response body and can be circumvented by using the clone() API that is provided for this purpose.
I also wrote a test that fails without this PR, and I think that would be generally useful to have.

Tests

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

test/test.js:953:2 › Load › handles large responses fails due to problems discussed in #591

Changesets

• If your PR makes a change that should be noted in one or more packages' changelogs, generate a changeset by running pnpx changeset and following the prompts. All changesets should be patch until SvelteKit 1.0

[netlify]

✔️ Deploy Preview for kit-demo canceled.

🔨 Explore the source changes: 1cb1794

🔍 Inspect the deploy log: https://app.netlify.com/sites/kit-demo/deploys/61f19cb6c3d6000007dbf0f3

[changeset-bot]

🦋 Changeset detected

Latest commit: 1cb1794

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

This PR includes changesets to release 1 package

Name	Type
@sveltejs/kit	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

[bluwy]

Looks like cloning it has an impact on large responses (probably expected) and caused the tests to fail 🤔

[Theo-Steiner]

Looks like cloning it has an impact on large responses (probably expected) and caused the tests to fail 🤔

it looks like we bumped into #591 here...

[Rich-Harris]

Thank you.

I pushed a commit that changes the behaviour so that the prerenderer gets access to the same buffered body as the code that inlines responses into the HTML (i.e. without cloning the response), though as it was on its way to GitHub I realised that this would fail in the case where the data wasn't buffered during load.

I guess we'll need a hybrid approach in which we use the buffered body if it's available, otherwise buffer it at prerender time. Working on it.