1 |
Hi all, |
2 |
|
3 |
I want to raise yet again the issue of multiple qt/kde. |
4 |
|
5 |
We have solved the issue of having multiple QT libs installed. However, this |
6 |
is not much help without having multiple kdelibs installed as well, since |
7 |
nearly all qt apps are kde apps as well. |
8 |
|
9 |
The problem here is that while a user typically uses only one kde, he usually |
10 |
has a large amount of supplementary kde-based apps (everything outside |
11 |
kde-base, which insludes koffice, kdevelop, and the other 1000+ kde apps). |
12 |
|
13 |
Another thing to remember is the fact that back in the kde 1->2 move, about |
14 |
1/3 of all kde apps (my estimation) were never ported to kde2, and were more |
15 |
or less abandoned. It was a very long time before the others were ported and |
16 |
achieved the same level of stability and features. So people who use kde as |
17 |
their main desktop may well prefer to stay with kde2.x and the apps they use |
18 |
than move to kde3.x right away. |
19 |
|
20 |
There are several options imo (in no particular order): |
21 |
1. Develop a scheme for kdelibs similar to that used with qt. Place different |
22 |
kdelibs in /usr/lib/kdelibs-$V and juggle paths around. Use eclasses to |
23 |
ensure correct kdelibs is used. |
24 |
Pro: easy to set up, unless unexpected problem emerges. |
25 |
Con: there's not much sense in having mutliple kdelibs like this unless we go |
26 |
for option 2. |
27 |
This is FHS0compatible, but probably falls under some "in the last extreme" |
28 |
clause. |
29 |
|
30 |
2. Revert to putting whole kde's in /opt/kde$V. |
31 |
Pro: easy to handle once working. Best for users, who can mix kde versions |
32 |
and apps at will. |
33 |
Con: goes against the FHS, which started this whole mess in the first place. |
34 |
|
35 |
3. Enforce single kdelibs. This also means enforcing a single qt. |
36 |
Pro: easiest to implement and manage. |
37 |
Con: will greatly inconvinience users. |
38 |
|
39 |
Right now it looks as if we're heading towards (3). I think there's no point |
40 |
in choosing (2) over (1). |
41 |
|
42 |
What do you think? |
43 |
|
44 |
|
45 |
-- |
46 |
|
47 |
Dan Armak |
48 |
Gentoo Linux Developer, Desktop Team |
49 |
Matan, Israel |