fix: default head requests for endpoints when no explicit head method…

[VitaliyR]

… implemented

Changes

I've noticed that /_image endpoint doesn't respond nicely (non 404) to HEAD requests. After investigation, I've found that none of server endpoints respond to the HEAD - they respond as 404.

User must explicitly implement export const HEAD to add support for this endpoint or provide export const ALL.

But typically users won't be doing this, since they typically might want just to add some GET handler and thats it. Since HEAD endpoint is useful for various things, for instance, checking if there is a route there but not caring about it's response - this PR adds default HEAD implementation for such cases.

• In case user would implement it explicitly, their implementation works export const HEAD
• If there are no export const GET handler - HEAD won't be auto-provided as well
• If user implemented ALL -> HEAD won't be auto-provided

Testing

1. TDD: Added test cases, tested before & after
2. Run examples/ssr/ via npm run dev:
   2.1. fetch('http://localhost:4321/api/products') -> 200 (API endpoint)
   2.2. fetch('http://localhost:4321/api/products', { method: 'HEAD' }) -> before 404, after 200
   2.3. fetch('http://localhost:4321/products/2', { method: 'HEAD' }) -> before 200, after 200 (Page endpoint, not affected)

Docs

Not sure about if docs should be updated.
Personally I'd expect this to work outside the box.

[changeset-bot]

🦋 Changeset detected

Latest commit: 3709bf9

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

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

[codspeed-hq]

CodSpeed Performance Report

Merging #13210 will not alter performance

_{Comparing VitaliyR:fix-endpoint-head-requests (3709bf9) with main (3b66955)}

Summary

✅ 6 untouched benchmarks

[ascorbic]

I think the idea of automatically enabling HEAD for GET handlers is good, but I think this needs a different implementation. The spec says that HEAD should return the same headers as the equivalent GET request, but this PR has it returning nothing. I think the right approach would be to call the GET handler and then return a new response with the returned status and headers.

[VitaliyR]

@ascorbic originally I've thought about saving time on the executing the GET handler, but, perhaps, you are right - according to spec we should aim to return the same status && headers, and we can do that only by executing the GET itself.

I've updated the PR, wdyt?

[florian-lefebvre]

Looks good to me!

[github-actions]

This PR is blocked because it contains a minor changeset. A reviewer will merge this at the next release if approved.

[ascorbic]

Updated the changeset and added a line to the docs: withastro/docs#10947

Let's go