On Thu, Sep 12, 2002 at 09:03:17AM -0600, Craig Kaes wrote:
> Yes, I looked at rewriting bss_file.c and came to the same conclusion
> that you did -- a good last-chance sort of solution (though bss_socket.c
> seems to have done much of this work???). In the end, I went ahead and
> preloaded th
Yes, I looked at rewriting bss_file.c and came to the same conclusion
that you did -- a good last-chance sort of solution (though bss_socket.c
seems to have done much of this work???). In the end, I went ahead and
preloaded the certificate and key. This key/cert combo may be used by
thousand
On Wed, Sep 11, 2002 at 12:41:18PM -0600, Craig Kaes wrote:
> I appreciate that I am now on my own. FWIW, though, your statement
> below is untrue, at least on Solaris 8. Open is not limited in the same
> way that fopen is w/ regard to # of fds. Also, fwiw, nether is socket(2).
To be fair, I
I appreciate that I am now on my own. FWIW, though, your statement
below is untrue, at least on Solaris 8. Open is not limited in the same
way that fopen is w/ regard to # of fds. Also, fwiw, nether is socket(2).
OPEN:
[craigk:/tmp]$ cat /home/craigk/tmp/fopen/foo2.c
#include
#include
#in
-
From: "Craig Kaes via RT" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Wednesday, September 11, 2002 12:51
PM
Subject: [openssl.org #279] Solaris 2.8 issue w/
fopen in openssl-0.9.6g/crypto/bio/bss_file.c
> > I'm running a (jabber) socket multiplexor that
On Wed, Sep 11, 2002 at 09:21:09AM +0200, Craig Kaes via RT wrote:
> OPEN_MAX, the max # of fds allowable to me is honored by fopen and on
> BSD and Gnu this value tracks ulimit values. On Solaris, tho, it
> appears hard coded. To wit:
>
> [craigk:~/tmp/fopen]$ cat foo.c
> #include
> #inclu
I'm running a (jabber) socket multiplexor that provides both SSL and
non-SSL interfaces. The short story is that it manages multiple
long-lived TCP connections. I have no problems connecting via SSL until
the total number of connections exceeds about 250. After this point,
non-ssl connecti
