fix: clone Headers before mutating them during prerendering

[Conduitry]

Fixes #9566.

I'm not sure how best to test this. I didn't see an easy way to make a given Headers object immutable, so I guess I'm going to have to call undici for real. If I try event.fetch somewhere and it refers to a local endpoint, that won't result in a real network request, and we won't run into the issue where undici makes the headers immutable. I don't think there's an HTTP server running during prerendering anyway. I'd prefer not to do something like export const GET = () => fetch("https://www.google.com/"); in my test. Any ideas?

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.

Tests

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

Changesets

• If your PR makes a change that should be noted in one or more packages' changelogs, generate a changeset by running pnpm changeset and following the prompts. Changesets that add features should be minor and those that fix bugs should be patch. Please prefix changeset messages with feat:, fix:, or chore:.

[changeset-bot]

🦋 Changeset detected

Latest commit: a70dc2b

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

[Conduitry]

It looks like I can hack it with Object.defineProperty(response.headers, "set", { value: null });. Now to try to figure out where this test ought to go.

[Conduitry]

Pushed a test which I've confirmed fails on master.

[benmccann]

lgtm. (assuming creating new Response doesn't cause a streaming response body to get buffered)

[Conduitry]

@benmccann I'm not sure whether it does- but I don't think that would really matter, because this is only during prerendering, where we're buffering the whole response and writing it out to a file anyway.

[benmccann]

true. Then yeah, this looks good to me

[benmccann]

@Conduitry it looks like this PR will need a rebase

[Rich-Harris]

assuming creating new Response doesn't cause a streaming response body to get buffered

it won't