- updated to new fedora-usermgmt macros
- cleanups: removed unneeded curlies, use plain command instead of %%__XXX
macro, whitespace cleanup, removed unneeded versioned dependencies
- added a 'Requires(post): group(clamav)' dependencies for -update and
added the corresponding Provides: to -data
- removed the %%_without_milter conditional; you won't gain anything when
milter would be disabled at buildtime
using of implicitly declared functions was reported to cause segfaults
on AMD64 (brought to my attention by Marc Perkel)
- added patch which fixes wrong usage of strncpy(3) in unrarlib.c
- added pseudo-versions for the 'init(...)' provides as a first step for
the support of alternative initmethods
- moved 'freshclam.conf.5' man page into the -update subpackage (#173221)
- ship 'clamd.conf.5' man page in the -server subpackage *too*. The same
file is contained in multiple packages now, but this man-page can not
be removed from the base package because it also applies to 'clamdscan'
there (#173221).
- removed -timeout patch; it is solved upstream
- reverted the -exim changes; they add yet more complexity, their
functionality can go into an own package and they contained flaws
- [milter] reverted the change of the default child_timeout value; it was
set to 5 minutes in 0.86.2 which conflicts with the internal mode where
a timeout must not be set. So, the clamav-milter would not run with the
default configuration
set to 5 minutes in 0.86.2 which conflicts with the internal mode where
a timeout must not be set. So, the clamav-milter would not run with the
default configuration
results in a negative number which is retained by the bash
modulo-operation. So the sleep may get a negative number of seconds
being interpreted as an option. This version makes sure that the
module-operations returns a non-negative value. [BZ #164494, James
Wilkinson]
- added support for a /usr/sbin/clamav-notify-servers.local hook; this file
will be executed (source'd) before all other actions and can abort the
entire processing by invoking 'exit'
in a negative number which is retained by the bash modulo-operation. So
the sleep may get a negative number of seconds being interpreted as an
option. This version makes sure that the module-operations returns a
non-negative value. [BZ #164494, James Wilkinson]
- changed the freshclam updating mechanism (again); now, it consists of a
crontab which does not need to be changed and a helper script
(freshclam-sleep). This helper script is configured by
/etc/sysconfig/freshclam
