Forcing a variant to a variant

[Scott Haneda]

On Jun 4, 2009, at 5:47 PM, Jeremy Lavergne wrote:

>> Duplication will have to happen on name, desc, long desc,  I  
>> assume?  Version probably does not matter in any strict sense, as  
>> it is not really downloading anything?
>
> Right
>
>> So with +server and +clamav_milter what do you do?  I am now seeing  
>> the need for these ports:
>>
>> clamav - base port, gets you cpam
>> clamav-server - depends on clamav but installs a startup item
>> clamav-clamav_milter - depends on clamav, but installs clamav_milter
>> clamav-server-milter - depends on clamav, gets you milter and the  
>> startup item
>>
>> I need some guidance here, I am seeing this very quickly getting  
>> out of hand where there are ports with more than one variant.
>
> The variants used descend through the dependency tree:  If you  
> install +milter on clamav-server, +milter will get used on clamav.   
> At least, this is what I believe:  Try it and find out.
>
>> If I need assp to depend on clam, and clam to have +server and  
>> +clam_av_milter, what do you suggest?  * I do not need milter at  
>> this time, just wondering.
>>
>> I am leaning on just not including the clamav as a variant for  
>> assp, but it is so logical to have it at the same time.
>
> This should all follow from using the variants in the server port.


I am not sure i understand this:
port install assp
Within that port, assp will depend on clamav, but since I can not  
define a variant for clamav +server, I have to make a clamav-server  
port.  Fine up to that point.

Now, what is I want clamav +server +clamav_milter available to the  
assp port?
My port install command would be port install assp still, there would  
still be a call in there to clamav-server, where does the milter go in?

Thanks
-- 
Scott * If you contact me off list replace talklists@ with scott@ *