[Mimedefang] MIMEDefang 2.73-BETA-2 is available

Matt Garretson mattg at assembly.state.ny.us
Thu Jan 19 17:52:07 EST 2012


On 1/19/2012 2:21 PM, David F. Skoll wrote:
> I might just document in the man page that you should
> chmod /var/spool/MIMEDefang if necessary.

That'd be fine by me.


However I've just noticed a problem I'm having with the beta 
that I'd missed the first time 'round.

It looks like, even with the new -G option, the umask of all 
the mimedefang processes is 077 when the "tmp" and "Work" 
dirs are created. This results in spool permissions like this:

# cd /var/spool/MIMEDefang
# ll -d .
drwxr-x---. 10 defang defang 280 2012-01-19 17:39 ./

# ll -d mdefang-q0JMXxUM009518
drwxr-x---. 4 defang defang 160 2012-01-19 17:34 mdefang-q0JMXxUM009518/

# ll mdefang-q0JMXxUM009518
total 60
-rw-r-----. 1 defang defang   759 2012-01-19 17:34 COMMANDS
-rw-r-----. 1 defang defang  1972 2012-01-19 17:34 HEADERS
-rw-r-----. 1 defang defang 45241 2012-01-19 17:34 INPUTMSG
-rw-------. 1 defang defang    66 2012-01-19 17:34 RESULTS
drwx------. 2 defang defang    40 2012-01-19 17:34 tmp/
drwx------. 2 defang defang    80 2012-01-19 17:34 Work/
 
The "Work" subdir is not group readable, so clamd can't scan it.

Might this be because of the increased "paranoia" in mimedefang-multiplexor.c ?
It seems to use the more lax umask just for creating control files, but 
then reverts to 077 just before looping.  AFAICT this is different than 
the 2.72 behavior.




More information about the MIMEDefang mailing list