HTTP status 422 should (IMHO) be "Unprocessable Content" not "Unprocessable Entity"

[hacklschorsch]

I am trying to write a spec-conforming microservice. The spec requires that I return HTTP 422 (Unprocessable Content) in a specific situation:

https://www.rfc-editor.org/rfc/rfc9578.txt

If any of these conditions are not met, the Issuer MUST return an HTTP 422 (Unprocessable Content) error to the Client.

But when I return a 422, node.js attaches Unprocessable Entity instead.

node/lib/_http_server.js

Line 158 in 57bc177

422: 'Unprocessable Entity', // RFC 4918 11.2

This issue for the (unrelated) Rust hyperium HTTP library explains it better than I can: hyperium/http#664

Both of the following RFCs define 422 but with a one-word difference:

• RFC 4918, "Proposed Standard", June 2007, "Unprocessable Entity"
• RFC 9110, "Internet Standard", June 2022, "Unprocessable Content"

Arguments in Favor of RFC 9110

• RFC 9110 is titled "HTTP Semantics" which is a direct fit for this library, whereas RFC 4918 is scoped to WebDAV.
• RFC 9110 is an Internet Standard while RFC 4918 is only a Proposed Standard.
• MDN uses RFC 9110 for the 422 status code
• RFC 9110 is newer (2022 vs. 2007)

[pimterry]

This seems convincing to me, but it'd be good to know what the rest of @nodejs/http think, and even though it's just a small difference it will clearly be a breaking change that might need careful handling.

I think the practical change for end users would just be the default status message for 422 responses, and the value of the exposed http.STATUS_CODES constant (both defined here).

There's also various related constants here and within llhttp and Undici though that we should also update (probably preserving the others as aliases for backward compat, if we can), although I don't think those are directly exposed publicly AFAICT.

Would you be open to creating a PR if everybody's on board?

[ljharb]

It's been "unprocessable entity" in practice for well over a decade; i'm not sure i see the value of this breaking change.

[marco-ippolito]

This would be a semver major breaking change, and I'm afraid it's gonna break a lot of apps

[ShogunPanda]

I also agree, it's semver-major.
FWIW we don't currently support any of the 911x RFCs.

[hacklschorsch]

Would you be open to creating a PR if everybody's on board?

Sure thing!

But it seems that won't be required :)

Thank you all for your consideration!