>>> If it was C, I would be hacking the code and adding debugging to see >>> where the connection is terminating ... etc. >> >> I'd be (secretly, of course) delighted if this were a reason to move >> from a competitor's product, but that's not the case here. Even if >> this server were written in C, with nice debugging support, and you >> had the source, this wouldn't help you. M-Link's dropping the >> connection on its side because it thinks you're running a protocol >> it doesn't recognise, so you'd need to debug that. >> >> Luckily, M-Link *is* written in C, and I do have the source, as well >> as a debug build that's duplicating the problem, so I'll take a >> look. :-) >> >> > Okay, source not needed - I ran up Wireshark, and looked at what's > going on. > > What happens is: > > I say "<starttls/>". > > You say "<proceed/>". This is good. > > I say "TLS ClientHello". > > You say - outside of TLS - "<failure/>". This, as they say, is not > good - we're certainly not expecting to see that there. > > Can you try to send something to my personal IM account? > [email protected] > > I've got wireshark running, so should be able to see what happens in > the opposite direction.
How odd ... test message sent. -N
signature.asc
Description: OpenPGP digital signature
