I wrote:
>> My first guess is that openssl does not work correctly when
>> the length is not divisible by eight. Does this certificate actually
>> *work* ?
Ben Laurie:
>It shouldn't be a problem to have odd-sized moduli.
Yup, actually the private key for the public key certified by the
Mysterious Polish Certificate is (computed using my factorization from
yesterday):
d = 30583757702412054338248862564530603930167546267161
64632993976708185854045053662176785522483239260125
48772660617022493381389757894937929607030767904489
473
Usually RSA implementations can only handle a modulus that
is a product of two primes (this one has seven). No problems
with the public key ops, but the computation of Phi(n) is a bit more
complicated and the usual CRT private key trick won't work.
Further observations:
o The public key of the recipient of this certificate actually has a
_negative_ modulus n, which was converted to -n by OpenSSL !
o The issuer "oi-wbd" is apparently Osrodek Informatyki - Wojew�dzki Bank
Danych .. what ever that is.
My second guess that someone simply messed with this cert with a hex
editor. No bugs in OpenSSL implied.
Cheers,
- mj
Markku-Juhani O. Saarinen <[EMAIL PROTECTED]> University of Jyv�skyl�, Finland
