On Fri, Oct 28, 2005 at 04:07:32PM +0300, Claudiu Dragalina-Paraipan wrote:
> I have noticed that BIO_do_connect doesn't start the connection upon
> call, instead the connection is established on the first BIO_puts, in
> my case.
> I am not sure if this is a bug or not, but I have traced it to
> bss_conn.c, function conn_ctrl, line 470:
> 
> --- snip ---
> case BIO_C_DO_STATE_MACHINE:
>  /* use this one to start the connection */
>  if (!(data->state != BIO_CONN_S_OK))
>    ret=(long)conn_state(b,data);
>  else
>    ret=1;
>  break;
> --- snip ---
 
Hello,

what's up with this problem ? I noticed it too when changing from 0.9.7 to
0.9.8.
few people seem to react about it, not even the author of the commit which
triggered the bug (Nils Larsch), to whom I already sent a mail a few weeks ago
(I haven't received an answer yet, unfortunately).

here's the commit in question. it didn't move for the past 6 months...
I'm quite surprised this problem got unnoticed for so long.

http://cvs.openssl.org/filediff?f=openssl/crypto/bio/bss_conn.c&v1=1.28&v2=1.29


the commit log ("added parentheses") and the nature of the bug makes me
worry that no or little testing was done before commiting.

it would be nice to provide some feedback about that since it causes a
serious functional regression when using BIO_s_connect (at least in my
case).

cheers,

-- 
Gabriel Forté <[EMAIL PROTECTED]>

______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [email protected]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to