Incompatibility with GnuPG 2 generates bad signatures about gnupg-pkcs11-scd

[mbechler]

And another one in pkdecrypt: Looks like gnupg now expects padding if not indicated otherwise (making decryption fail).

@@ -1283,6 +1285,7 @@
 				&ptext_len
 			)
 		)) != GPG_ERR_NO_ERROR ||
+		(error = assuan_write_status (ctx, "PADDING", "0")) != GPG_ERR_NO_ERROR ||
 		(error = assuan_send_data(ctx, ptext, ptext_len)) != GPG_ERR_NO_ERROR
 	) {
 		goto cleanup;

[alonbl]

Thank you! I hope I can check out this during this weekend.

…

On 30 November 2016 at 17:32, mbechler ***@***.***> wrote: And another one in pkdecrypt: Looks like gnupg now expects padding if not indicated otherwise (making decryption fail). @@ -1283,6 +1285,7 @@ &ptext_len ) )) != GPG_ERR_NO_ERROR || + (error = assuan_write_status (ctx, "PADDING", "0")) != GPG_ERR_NO_ERROR || (error = assuan_send_data(ctx, ptext, ptext_len)) != GPG_ERR_NO_ERROR ) { goto cleanup; — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#3 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABNIrTCWcDijADNo4Hg-W6WS0MWyrwXoks5rDZcVgaJpZM4KxYCB> .

[alonbl]

Hi,
Can you please checkout gnupg-2.1 branch and see if it works for you?
Thanks!

[mbechler]

Thanks for your work. Sorry for the delay.

The corruption while signing is gone but now produces unsupported algorithm as we get into the first branch (hash parameter set) and the hash size won't match as the OID prefix is already included.

Patch adding explicit branches for these cases
gnupg-pkcs11-scd-hashcomp.patch.txt

[alonbl]

Thanks! Can you please send this as pull request?

[alonbl]

Merged, can I release?

[mbechler]

Seems fine to me. I haven't seen any more issues with the patched version. Thanks.