[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: Questions about splitting kdebase into a plethora of packages
From:       Lokheed <lokheed () gmail ! com>
Date:       2005-02-11 1:59:48
Message-ID: 420C1194.6070400 () gmail ! com
[Download RAW message or body]

How do so many people fail to see the problem with this. First they are 
called meta ebuilds. Second the problem lies when you have 300+ packages 
and want say 127 out of them. If you cant use a meta ebuild, then you 
HAVE to get 300+. If you dont, then you are stuck putting in EACH 
individual ebuild one by one if they do not have any dependencies set.

Do you think it would be fun to put in 127 ebuilds by hand? Then it 
comes to the problem of what if I dont want everything from kdebase, 
again I will break the meta ebuild and have to put in everything by 
hand, one by one. How does no one see the problem with this?

Then another problem lies with what kind of meta ebuild will you be 
doing? Are there going to be crippled releases such as the atrocious 
gnome-light? Where will the line be?

Gentoo's power is USE flags. It would have been simple to release a meta 
ebuild for kdebase, kdemultimedia, kdeutils, kdegraphics, and so one 
with USE flags that govern what gets installed.

Then you will only need ONE ebuild for kdebase and so on. This way, a 
user can set the use flags via /etc/portage/package.use and you ONLY get 
what you want and you wont have to support 300+ ebuilds. Does that not 
make sense? It is completely within the means of the Gentoo devs to 
handle it this way and then you wont have breakage since you can set all 
USE flags by default and have the user remove them if they choose too. 
This is how it SHOULD have been done.

Granted you have ebuilds that will hold an extreme amount of USE flags 
but it certainly beats what they hav done now. How is this NOT a good idea?

My comments seem to have gone above your head as well because they 
already have a meta ebuild for all of kde but it contains more than just 
kdebase, which means its useless if you dont want ALL of kde. How many 
times do I have to say this??? I feel like I am the last sane man here. 
Am I saying it incorrectly? Why cant people see the problem with meta 
ebuilds?

Jason Keirstead wrote:

>On February 10, 2005 07:58 pm, Lokheed wrote:
>  
>
>>Yes konqueror without kio-slaves is available in the jumbled mess of
>>ebuilds. They have also split each individual kio-slave to add to
>>further confusion.
>>
>>I understand what you are saying and it has set aside my fears of
>>stability as it seems KDE can handle being split into modules like this
>>with the main problem being usability and missing services/features.
>>    
>>
>
>The solution seems pretty clear to me - do what debian has done for years, and 
>make some ghost ebuilds for the kde packages (kdelibs, kdebase, kdenetwork, 
>etc), that contain no files themselves, but rather, depend on the other 
>ebuilds that make up the sum of that package. Thus, if you 'emerge kdebase', 
>you get everyting. But if all I want is Konqueror, I can just emerge that.
>
>
>  
>
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]