Sveltekit throws developer warning when using correct fetch from load-function

[MathiasWP]

Describe the bug

In #4958 a helper warning was introduced if the window.fetch method is used instead of the fetch method provided by the load function. However, this warning is logged even though the correct fetch method is used.

Reproduction

https://github.com/MathiasWP/sveltekit-window-fetch-warning-bug

Logs

No response

System Info

System:
    OS: macOS 11.4
    CPU: (8) x64 Intel(R) Core(TM) i5-8259U CPU @ 2.30GHz
    Memory: 21.08 MB / 8.00 GB
    Shell: 5.8 - /bin/zsh
  Binaries:
    Node: 16.13.0 - /usr/local/bin/node
    Yarn: 1.22.10 - /usr/local/bin/yarn
    npm: 8.1.0 - /usr/local/bin/npm
  Browsers:
    Brave Browser: 101.1.38.111
    Chrome: 101.0.4951.64
    Edge: 101.0.1210.53
    Firefox: 100.0
    Safari: 14.1.1
  npmPackages:
    @sveltejs/adapter-auto: next => 1.0.0-next.44 
    @sveltejs/kit: next => 1.0.0-next.338 
    svelte: ^3.44.0 => 3.48.0

Severity

annoyance

Additional Information

No response

[elliott-with-the-longest-name-on-github]

Not sure why this is happening, but I can confirm that commit d7cb8b8 (warn if load uses window.fetch) is where this was introduced, i.e. this is not a regression caused by a later commit.

[mrkishi]

Note: this is about initial_fetch using the patched window.fetch instead of the original.

[elliott-with-the-longest-name-on-github]

@mrkishi

You're totally correct -- I just nailed it down. The easy fix is this: https://github.com/tcc-sejohnson/kit/tree/sejohnson-fix-fetch-warning-on-initial-fetch

The slightly-harder fix is to pass fetch to initial_fetch (adjust the initial_fetch signature to accept fetch).

Given that this is a dev-mode thing for ergonomics, I don't see much of an issue just checking the stack trace for initial_fetch as seen in the linked branch. If we want to do that, I'll open a PR. If not, happy to make the adjustments necessary to not use the patched window.fetch.

I'm not entirely sure how to test the changes, though -- this would only happen before the app is started...

[Rich-Harris]

I have a fix in a branch, one sec...