On 2007.04.26., at 1:33, Sam Varshavchik wrote:
>> The attached dirty patch makes starttls work on Digital UNIX 4.0
>> (Tru64 4.0) amongst more others.
>
> … and stops it from working everywhere else, since you return from
> main() if the rand generator is initialized.
Told you it's dirty.
Gábor Bérczi (Gabucino) writes:
Confirmed, and fixed. starttls.c doesn't check if randomness is
available, and neither does it seed it when it isn't.
This is a bug in courier-imap !
The attached dirty patch makes starttls work on Digital UNIX 4.0
(Tru64 4.0) amongst more others.
… and s
Confirmed, and fixed. starttls.c doesn't check if randomness is
available, and neither does it seed it when it isn't.
This is a bug in courier-imap !
The attached dirty patch makes starttls work on Digital UNIX 4.0
(Tru64 4.0) amongst more others.
starttls.diff
Description: Binary data
Ok I didn't write "set -a" into my initscript.
Next:
Apr 24 19:16:41 gabucino imapd: couriertls: connect: error:
140B544E:SSL routines:SSL_GET_NEW_SESSION:ssl session id callback failed
I think couriertcpd doesn't seed the random generator on systems
which don't have urandom. :(
On 2007.04.2
Yo.
I compiled courier-authlib with DEC C compiler, works fine. I
compiled courier-imap with gcc, which also works fine. Had to link a
custom snprintf.o for each of the libs and binaries, because OSF1
libc lacks snprintf+vsnprintf.
courier works fine on port 143, but SSL does not. I used th