None of my accounts work. I see a message claiming I need to “Sign in to reconnect to this account” but, when I do, it just fails.
It’s failing for all Google accounts and for my IMAP account.
Also, in the journal, I see these kinds of errors:
Nov 29 16:32:07 desktop.casa.g02.org gnome-keyring-daemon[2706]: asked to register item /org/freedesktop/secrets/collection/login/3, but it's already registered
Nov 29 16:32:24 desktop.casa.g02.org gnome-keyring-daemon[2706]: asked to register item /org/freedesktop/secrets/collection/login/3, but it's already registered
Nov 29 16:32:37 desktop.casa.g02.org gnome-keyring-daemon[2706]: asked to register item /org/freedesktop/secrets/collection/login/3, but it's already registered
Note that, in my experience, you will see those messages all the flippin’ time, in the normal operation of gnome-keyring-daemon. They seem to signify nothing other than a conceptual disconnect between what the keyring-daemon thinks is an unrecoverable program error, and what one or more clients considers a harmless redundancy not worth the extra code to prevent.
It’s fairly common for API clients to determine that a call to first check whether an item exists in a database, before then executing the operation to insert it when it doesn’t, would take roughly the same amount of time as the ‘insert’ call itself. And to then conclude that they can save time and avoid making two calls instead of one, by just always doing the ‘insert’ call unconditionally, and letting that call fail harmlessly when the item already exists.
Which is great and all, unless the API provider feels the need to log a critical (!!) message every time a harmless redundant insert call is made, even though harmless redundant insert call is both redundant, harmless.
Unrelated See Also
The message that clutter logs every time a non-parentless child is added to an actor:
Another “invalid” call that is made constantly by applications, and probably shouldn’t be protected with a noisy critical-message assertion. At least, not without first checking whether child->priv->parent == selfalready. (IOW, the child is being redundantly added to its existing parent.) A condition which I suspect accounts for the majority of those critical messages complaining about adding already-parented children to actors, and which should (again) be harmless in practice, and therefore safe to silently ignore.
Because it sure would be swell if they were silently ignored!