On 2006-07-14, at 0:53, Mark Crispin wrote:
One other possibility that came to mind is that a common third-
party patch makes the c-client library be a shared library, instead
of a static part of the program binary. This really doesn't save
all that much on modern systems (even with static linking, multiple
instances of the same program share their code pages). More to the
point, I've seen cases where the shared library of c-client was for
some other version than what the application was built to use, and
the resulting version skew resulted in crashes.
So, if c-client is a shared library, try it with a static link. If
the problem vanishes, so much the better...
Interesting. As you might remember, Mark, I'm also a FreeBSD user,
and can tell you that the cclient port is implemented as a shared
library, as it's a separate port. So differing versions could be an
issue perhaps. I didn't realize it was intended to be a static library.
So, Jaime, you definitely want to make sure they're synchronized:
FreeBSD> pkg_info | egrep '(cclient|imap-uw)'
cclient-2004g,1 Mark Crispin's C-client mail access routines
imap-uw-2004g_1,1 University of Washington IMAP4rev1/POP2/POP3 mail
servers
I never see signal 11 on any of the imap-uw systems I manage. But do
see signal 6, SIGABRT I think, though pretty rarely, ie: "kernel: pid
43264 (imapd), uid 1021: exited on signal 6"
-rory
_______________________________________________
Imap-uw mailing list
[email protected]
https://mailman1.u.washington.edu/mailman/listinfo/imap-uw
