Per the WebAuthn spec (https://w3c.github.io/webauthn/#sctn-op-make-cred , step 3), errors returned from `navigator.credentials.get()` correspond to the following - `InvalidStateError`: the user consented to create a new credential, but there are no authenticators that match the `authenticatorSelection` but aren't in `excludeCredentials`. - `NotAllowedError`: the user did not consent to create a new credential The most obvious case for "did not consent to create a new credential" is if the user pressed "cancel". However, it seems Safari also does the same if trying to register a platform authenticator (Touch ID or Face ID) that is already registered. While other browsers allow the user to interact with the prompt, e.g. Safari on macOS 11.0 Big Sur flashes `Do you want to allow "example.com"' to use Touch ID` (with the options "Don't allow" and "OK") but immediately replaces it with `You have already set up Touch ID for this website.` (with only the option "OK"). This makes it impossible to tell if the user intended to create a credential, and for the RP (website) to offer a more useful option based on `InvalidStateError`. I don't know if this is a bug, but I wanted to make sure it was not an accident. The user experience (with the flashing prompt) is not great, so I think there is room for some improvement, either in the browser, or in the website's reaction to the error. It's unclear to me if any of the bugs here cover this issue: https://bugs.webkit.org/buglist.cgi?bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&field0-0-0=product&field0-0-1=component&field0-0-2=alias&field0-0-3=short_desc&field0-0-4=status_whiteboard&field0-0-5=content&list_id=6674120&order=changeddate%20DESC%2Cbug_status%2Cpriority%2Cassigned_to%2Cbug_id&query_format=advanced&type0-0-0=substring&type0-0-1=substring&type0-0-2=substring&type0-0-3=substring&type0-0-4=substring&type0-0-5=matches&value0-0-0=webauthn&value0-0-1=webauthn&value0-0-2=webauthn&value0-0-3=webauthn&value0-0-4=webauthn&value0-0-5=%22webauthn%22
<rdar://problem/72484635>
Created attachment 441440 [details] Patch
Comment on attachment 441440 [details] Patch r=me
Created attachment 441621 [details] Patch
Committed r284413 (243182@main): <https://commits.webkit.org/243182@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 441621 [details].
Re-opened since this is blocked by bug 231936
Reverted in https://trac.webkit.org/changeset/284429/webkit
Created attachment 441781 [details] Patch
Created attachment 441790 [details] Patch
Comment on attachment 441790 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=441790&action=review > Source/WebKit/ChangeLog:11 > + A fixme to add this was inadvertently removed in https://bugs.webkit.org/attachment.cgi?id=393180&action=prettypatch nit: extra space before URL > Source/WebKit/ChangeLog:13 > + Reviewed by NOBODY (OOPS!). This is supposed to come *before* the description, not after.
Created attachment 441798 [details] Patch
Comment on attachment 441798 [details] Patch r=me (and Chris)
Waiting for iOS-wk2 to finish before adding cq+
Created attachment 441826 [details] Patch
Committed r284532 (243274@main): <https://commits.webkit.org/243274@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 441826 [details].
Comment on attachment 441826 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=441826&action=review > Source/WebKit/ChangeLog:2 > +2021-10-19 John Pascoe <j_pascoe@apple.com> > + [WebAuthn] Obtain consent to create new credential when platform authenticator in excludedCredentials Trivial formatting mistake: there should be a blank line here.