[Classpathx-crypto] Re: NESSIE RSA-PSS submission

[Jakob Jonsson]

Hi Raif,

I think the inconsistency lies in how you define the leftmost position of an
octet string; this could be either "position 0" (as in C and Java) or
"position 1". We have adopted the second convention; the octet on position n
is the nth leftmost octet rather than the (n+1)st leftmost octet. The length
of the PS string preceding the 01 octet is emLen-hLen-sLen-2, so the 01
octet is on position emLen-hLen-sLen-1.

Yet, I admit that this convention should have been explicitly stated in the
document.

Best regards,
Jakob Jonsson
RSA Laboratories

----- Original Message -----
From: "Raif S. Naffah" <address@hidden>
To: <address@hidden>; <address@hidden>
Cc: "GNU Crypto Developers" <address@hidden>; "Paulo S. L. M.
Barreto" <address@hidden>
Sent: Saturday, January 05, 2002 11:52 PM
Subject: NESSIE RSA-PSS submission


> Gentlemen,
>
> while implementing this scheme in Java, i found an inconsistency in the
> specification of the EMSA-PSS-Decode(M, EM, emBits) primitive (document B
> of your submission, page 11), at step #10.
>
> this should read "...or if the octet at position emLen -hLen -sLen -2 is
> not equal to 01..." and not "...emLen -hLen -sLen -1."
>
>
> cheers;
> rsn