[csw-pkgsubmissions] newpkgs clamav, libclam6, libclam6_devel

[Jake Goerzen]

On 12/24/10 03:50, Peter FELECAN wrote:
> Philip Brown<phil at bolthole.com>  writes:
>
>    
>> On 12/23/10, Peter FELECAN<pfelecan at opencsw.org>  wrote:
>>      
>>> Philip Brown<phil at bolthole.com>  writes:
>>>
>>>        
>>>> seems pretty non-ambiguous to me.
>>>> "run time only". == "only data generated at run time".
>>>>          
>>> This mean that if the data is not generated at run-time it cannot be put
>>> there?
>>>
>>> What about the files distributed by SUNWapch2r? They are in /var/apache2
>>> and there are files and directories not generated at run-time. Isn't it?
>>> Is that in contradiction with your interpretation?
>>>        
>> Let me point out that you are referencing a Sun package, and comparing
>> it to CSW packages.
>> Sun does not always do things sanely.
>> For example, shipping the *config* files for smf, under /var.
>> idiotic.
>>      
> Ah, the eternal ending argument: Sun does "idiotic" things...
>
>    
>> SUNWapch2r seems to be similarly stupid.
>>
>> btw: directories under /var/opt are no problem.
>>      
> If you take 30s you can see that is not only directories but also files.
>
>    
>>> The maintainer explained very clearly why it must be in the place that
>>> the files are but maybe this is a case of "whatever the maintainer feels
>>> like"? Maybe disrupting a clamav installation is of the "best" quality
>>> for the corporate world?
>>>        
>> Its not a matter of "no you cant configure the program to *use* /var/opt".
>> It's merely a matter of not "shipping them *in* /var/opt".
>> ie, "dont put files in /var/opt as part of your 'prototype' file".
>> There isnt a problem with a postinstall script copying files into
>> /var/opt/csw, from elsewhere, as I already explicitly said on this
>> thread.
>>
>> It's fairly trivial for "the maintainer" to toss together a
>> postinstall script to copy them from somewhere else. Should take all
>> of 10 minutes.
>>      
> This is what I dislike the most: "fairly trivial", "should take 10
> minutes", this way you decide what the given maintainer should do with
> his time.
>
>    

I didn't realize this discussion was going on in pkgsubmissions since I 
wasn't subscribed (but I am now).  Here is a idea using the gar recipe 
to work this out:

create a new postinstall script

vi files/CSWclamav.postinstall

#!/bin/sh

MAIN = /opt/csw/share/doc/clamav/main.cvd
DAILY = /opt/csw/share/doc/clamav/daily.cvd

if [ ! -f /var/opt/csw/clamav/db/main.cvd ] ; then
   cp $MAIN /var/opt/csw/clamav/db
fi

if [ ! -f /var/opt/csw/clamav/db/daily.cvd ] ; then
   cp $DAILY /var/opt/csw/clamav/db
fi

Add to gar recipe:

DISTFILES += CSWclamav.postinstall

Add to gar recipe in the "post-install-modulated:"

@mv $(DESTDIR)/var/opt/csw/clamav/db/*.cvd 
$(DESTDIR)/opt/csw/share/doc/clamav/

This way when gmake package is ran the files will be prototype'ed in 
/opt/csw/share/doc/clamav.  Then, during pkgadd if the files don't exist 
the postinstall script will copy them there and clamd can start up.

Jake