[guest - Sat May 14 03:32:51 2005]:
Something hit the enter key for me before I finished typing
I meant to say that the assertion will go off any time the high order
part of the number is zero (amoung other things). I sort of suspect
that was not the intent and that it was meant to check
The assert statement in bn_div_words (bn_asm.c) is kind of silly. The
bug report that came to me is:
When built for debugging, the command "ssh-keygen -t rsa1 -N '' -q -f
PATH" fails with the message "Assertion failed: ((i == BN_BITS2) || (h >
(BN_ULONG)1< BN_div_word ==> bn_div_words.
BN_ULO
Roger,
What I have found is that OpenSsl allocates s3->rbuf and s3->wbuf as soon as the SSL session is created, however uses these buffers as scratch buffer while doing handshake, or doing any other SSL opertion. Once the operation is done, those buffers could be safely freed since then do not ho
openssl cannot sign large files (including detached signature, at least
in DER and PEM formats). Attached patch (against 20050511 snapshot)
fixes this problem.
--
Artem Chuprina
RFC2822: Jabber: [EMAIL PROTECTED]
__
OpenSSL P
[EMAIL PROTECTED] - Sun Mar 6 19:39:25 2005]:
> Hi,
>
> Somewhen between 0.9.7 and 0.9.8 (I first noticed it around 6 months
> ago),
> PKCS12 creation was refactored, and the new code left out the pkcs12
> -CSP
> implementation (the option is still accepted, but does nothing
> useful). I'm
> n
Hello!
I use openssl-0.9.8 dev. There is a memory leak on verification of
signed messages. This leak exist in snapshots 20050404 and 20050511.
I've tested it using Debian Linux.
I sign a file in default format with
export OPENSSL_DEBUG_MEMORY=on
LD_LIBRARY_PATH=. apps/openssl smime -sign -inkey
