Subject: Re: O-BSD uam_load errors + wrap/encrypt?
From: Jean-Philippe Rey (jprey@ads.ecp.fr)
Date: Thu Mar 22 2001 - 12:18:08 EST
At 10:33 -0600 22/03/01, Karen A Swanberg wrote:
>
>For every variation, this comes up in my logs when the daemons are
>starting:
>
>afpd[] uam_load(uams_guest.so): mod_symbol error
>afpd[] uam_load(uams_clrtxt.so): mod_symbol error
>afpd[] uam_load(uams_dhx.so): mod_symbol error
Now that you mention it, I also encountered this problem. I solved it by adding
#define DLSYM_PREPEND_UNDERSCORE 1
in the config.h file, and then gmake again.
>Jean-Philippe Rey suggested not adding in the --with-tcp-wrappers during
>the compile. Is this because they won't work in this beta, or does the
>compile adds them in manually? I would very much like to use them. The
>whole reason I'm up here on the bleeding beta edge is to use a mostly
>secure version of netatalk, I have a perfectly functional version of
>1.4b2+asun2.1.3, it just doesn't encrypt anything, but it does use
>wrappers. If wrappers don't work now, would people recommend staying with
>a wrapped / non-encrypted netatalk, or go non-wrapped / encrypted
>(assuming I can get this problem ironed out)? I have no solid evidence of
>sniffers on my dept. network, but I'm at a big University, and I'm
>starting the migration to SSH as well, so it seems silly to insist on SSH
>when my netatalk connections are cleartext.
My netatalk server already resides in a protected network, so
wrappers are not so important for me. When I saw that it won't work
out of the box, I didn't try to make it work. But it was before I add
the DLSYM_PREPEND_UNDERSCORE line, that may solve also the wrapper
library problem. Give it a try.
-- Jean-Philippe Rey Délégué informatique - École Centrale Paris jprey@ads.ecp.fr 92295 Châtenay-Malabry Cedex - France
This archive was generated by hypermail 2b28 : Sun Oct 14 2001 - 03:04:35 EDT