OpenMacNews wrote:
> but, iiuc, smtps/port 465/tls-on-connect is a somewhat different
> negotiation than 'true' STARTTLS initiation ...
sure, but what happens after STARTTLS is the same, AFAIK.
> e.g., in TBird v1.5, TLS is *specifically* called out, allowing STARTTLS
> on port 143 as is 'should be'.
143 is imap. If you ment smtp (25): This was already possible with TB
1.0x. I used smtps to save a little time on high-latency connections.
> i can only tell you that i can now reproduce the issues on 3 diff Tbird
> boxes/instances ...
> admittedly with just ONE exim/cyrus instance ...
The authenticator is called long before the DATA phase, so this cannot be
the problem. Same applies to TLS.
I cannot think of single way where these things will change what happens
in an ACL, as long as you don't check for them or there's a bug in Exim.
You can try this yourself: Disable the DATA and MIME ACLs and check again.
>> If you want the real reason, run "exim -d -bd 2>&1 | less", send a
>> message with TB and search for "550".
> i *have* been running with:
> ../exim -bd -q15m -d
I didn't see any debug output.
> > Depending on the problem, you
> > might have to modify the debug optinos, e.g. "-d+expand".
> hmmm ... i'd enabled ALL logselector options ... not the same?
Even better, but where's the output? It produces several pages.
