On Thu, 16 Apr 1998, Eric Young wrote:
> > if crashes also for integers larger than 16 bits (but it takes longer to
> > crash)
>
> hmm... so it does, I'll see if I can fix it.
Ok, fixed the bug. Basically I was being a bit loose with my
'fixup BIGNUM->top value' and sometimes it would become negative,
indicating that there was a negative number of words making a bignum.
The main cause in this case is an 'inlined' in
BN_from_montgomery. If you edit line 170, and change
the
#if 0
to
#if 1
to re-enable the BN_rshift() call, I think this will fix the problem.
For my stuff, I've fixed the inline shift.
The fix will be in 0.9.1 (or will that be 0.9.0a?)
thanks for the bug report, I've tightend up quite a few other places
in the bignum library and even put in some compilable assertions :-)
eric
+-------------------------------------------------------------------------+
| Administrative requests should be sent to [EMAIL PROTECTED] |
| List service provided by Open Software Associates, http://www.osa.com/ |
+-------------------------------------------------------------------------+
