multipart/form-data PATCH request produces http-request snippet with no request body

[beuss]

When documenting a multipart/form-data request using PATCH method, we end up with only headers in http-request snippet.
It looks like switch to multipart is only handled for PUT/POST (org.springframework.restdocs.http.HttpRequestSnippet#getRequestBody). Is there any reason for this? I didn't find any restriction about multipart and PATCH in https://datatracker.ietf.org/doc/html/rfc5789
Anyway, the restriction on PUT/POST looks like a little excessive, if body content is empty but we've parts, it's likely that we want those parts to be included in the generated snippet.

[codespearhead]

Media type multipart/form-data is only defined for the POST method as per RFC 2616:

Note: The "multipart/form-data" type has been specifically defined for carrying form data suitable for processing via the POST request method, as described in RFC 1867 [15]

Which is clarified by errata RFC 5789:

the operation does not modify the resource identified by the Request-URI in a predictable way that's defined by the semantics of the PATCH media type, POST should be considered instead of PATCH or PUT.

That's why MDN web docs says PATCH shouldn't be used for binary data, as stated by this answer:

PATCH is not for uploading binary data (like what you want with the image), see: https://developer.mozilla.org/en-US/docs/web/http/methods/patch (Allowed in HTML Forms: No)

This issue should be closed in my opinion.

[beuss]

1 - Well "specifically designed", doesn't mean "restricted to", I don't see any normative content in the note (no "SHOULD/MUST not be used with other HTTP methods" or so). Moreover HttpRequestSnippet already handles parts for PUT requests

2 - Don't see how multipart/form-data implies that the resource isn't modified in a predictable way

3 - The answer you quote finishes by

Uploading binary data (like images) has to be done with POST always, PUT or PATCH are used to modify the attributes of the ressource.

Don't see why a resource cannot embed binary data. Moreover, MDN isn't a normative document

[wilkinsona]

I agree, @beuss. I see no reason to close this issue.