[Pkg-clamav-devel] LogRotate, create user and the option in the config

[Scott Kitterman]

On May 2, 2014 9:58:47 AM EDT, Andreas Cadhalpun <andreas.cadhalpun at googlemail.com> wrote:
>Hi,
>
>On 28.04.2014 23:42, Andreas Cadhalpun wrote:
>> On 25.04.2014 21:44, Sebastian Andrzej Siewior wrote:
>>> On 2014-04-13 14:28:03 [-0400], Scott Kitterman wrote:
>>>> On April 13, 2014 12:24:55 PM EDT, Andreas Cadhalpun
>>>> <andreas.cadhalpun at googlemail.com> wrote:
>>>>> Currently only clamav-base/LogRotate and clamav-base/Bytecode are
>>>>> shared
>>>>> between the packages. All the other options of clamav-base are
>only
>>>>> used
>>>>> by clamav-daemon.
>>>>> So duplicating the LogRotate and Bytecode options to the packages
>that
>>>>> use these would allow to remove all templates from clamav-base.
>>>>>
>>>>> Then clamav-base would only be used to create the clamav user.
>>>>>
>>>>> What do you think about this?
>>>>
>>>> I don't know what happens on an installed system if existing
>>>> templates are moved between packages. Can someone test this?
>>>
>>> The idea sounds good, we just need to make sure that we don't break
>>> anything right?
>>> In that case, Andreas, could you please prepare this split in a
>topic
>>> branch and
>>> then we can test this and make sure that the options are correctly
>>> migrated
>>> from clamav-base to clamd and other package?
>>
>> This is prepared in the branch aca_templates.
>> I couldn't find any problems on fresh installation or upgrade (in
>> jessie), so it should be fine. But it wouldn't hurt, if you
>review/test
>> this a bit more, e.g. installation/upgrade in wheezy and squeeze.
>
>I've tested this in wheezy and squeeze and found one problem:
>The value of clamav-base/debconf is not saved directly in the
>clamd.conf 
>file and thus was not migrated correctly.
>I implemented a fix by checking the beginning of the clamd.conf file
>for 
>'##', which is only in the clamd.conf.sample.
>
>So if none of you found further problems, I think this can be merge to 
>unstable.
>
>By the way, when testing in squeeze, I noticed that neither the pkill
>-F 
>nor the start-stop-daemon --status options exist in squeeze, so the
>init 
>file update should not go to squeeze.

Squeeze is EOL at the end of this month, so I think that's fine (I'm happy to let someone else care about squeeze-lts  if they want).

Scott K