gpgsm does not support OAEP. Actually it does not make much sense to use this padding scheme at all. It has not advantage over PKCS#1. Thus I change this to a feature request to allow decryption using OAEP

Still does not work on vsd-3.3.7-beta90.9 @ win10. Essentially the same behavior as before:

Do I understand this correctly, that on CLI key generation the public key of the designated revoker should be fetched automatically?

Maybe. EncryptionResult has a list of invalid recipients and I've changed the code to show the Retry dialog only if there's at least one invalid recipient.

Your suggestion sounds ok to me, maybe with a slight change for the message: "Failed to encrypt the notepad because at least on certificate could not be validated."

I tried to add the list of invalid recipients to the message box, but it seems that gpgsm stops the validation of the certificates at the first invalid recipient. I got only the first Bob certificate reported as invalid recipient when I tried to encrypt to both Bob certificates so that it doesn't make sense to list the (incomplete) list of invalid recipients. It also means that Kleopatra cannot update the invalid recipient certificates because it knows only of one invalid certificate.

Ideally the certificate would change, but Kleopatra has no idea that this certificate turned out to be not valid. In fact, Kleopatra doesn't even know that the encryption failed because of some certificate. It could have failed for any other reason (e.g. full disk). Kleopatra only knows that an error occurred and offers to retry with lower security. (I looked at GpgOL and it does the same.)

yes, basically it's what we want.

Current implementation for the case of an S/MIME certificate which turns out to be invalid when it's used for encryption. Is that what we want?

Before making subtickets for each application: I wonder if it is not all Kleopatra anyway? Isn't the security approval dialog basically Kleopatra?

The equivalent for invalid S/MIME certificates are not-certified *PGP certificates.
(Valid/invalid are not ideal as technical terms as they have a broad general meaning, too. I hope my usage here is correct ;-) It is what I gathered from an explanation given by Werner.)