1431137 - Our CredentialRequestOptions doesn't look much like the spec's

Status: RESOLVED INVALID

(Core :: DOM: Web Authentication, enhancement, P2)

Description

[Boris Zbarsky [:bzbarsky]]

We have:

  dictionary CredentialRequestOptions {
    PublicKeyCredentialRequestOptions publicKey;
    AbortSignal signal;
  };

whereas the spec at https://www.w3.org/TR/credential-management-1/ (what our IDL file links to) has:

  dictionary CredentialRequestOptions {
    CredentialMediationRequirement mediation = "optional";
  };
  partial dictionary CredentialRequestOptions {
    boolean password = false;
  };
  partial dictionary CredentialRequestOptions {
    FederatedCredentialRequestOptions federated;
  };

and the editor's draft at https://w3c.github.io/webappsec-credential-management/ has:

  dictionary CredentialRequestOptions {
    CredentialMediationRequirement mediation = "optional";
    AbortSignal signal;
  };
  partial dictionary CredentialRequestOptions {
    boolean password = false;
  };
  partial dictionary CredentialRequestOptions {
    FederatedCredentialRequestOptions federated;
  };

In either case, the spec does not have a PublicKeyCredentialRequestOptions member.  This is a serious compat issue, because {} is a valid value for the spec dictionary, but not for ours.

Comment 1

[J.C. Jones [:jcj] (he/they)]

Thanks, bz; a side-effect of our minimum credman implementation. I'll find the minimum change that fixes this and produce a patch, and see if I can turn on some more of the webIDL web platform tests.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Oh, this is coming from https://w3c.github.io/webauthn/#credentialrequestoptions-extension ?

Seems like this should be fixed in webauthn, then, since it's making {} an invalid value for CredentialRequestOptions, which seems pretty dubious to me.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Actually, it looks like webauthn may be depending on some IDL behavior that https://github.com/heycam/webidl/issues/76 proposes changing, and Gecko implements the post-change behavior.

Comment 4

[Boris Zbarsky [:bzbarsky]]

I filed https://github.com/w3c/webauthn/issues/750

Comment 5

[R. Martinho Fernandes [:rmf]]

Since this actually matches the Webauthn spec as it is today, I'm closing this.