| 1 |
Hi all, |
| 2 |
|
| 3 |
I want to raise the issue of supporting old versions of KDE. |
| 4 |
|
| 5 |
I am about to add kde-2.2.2 ebuilds to cvs (2.2.2 is to be released later |
| 6 |
today) and new revisions of the kde-2.2.1 ebuilds. Both sets will support |
| 7 |
multi-kdelibs, which allows various combinations of kdelibs packages to be |
| 8 |
used together. Unfortunately, this support isn't backward-compatible. |
| 9 |
|
| 10 |
The question is, should I continue to support older KDEs? How old, and how |
| 11 |
much support? |
| 12 |
|
| 13 |
Support for the new multi-kdelibs and multi-qt schemes, and for the upcoming |
| 14 |
new dependency specification format (portage 2?), can only be really added |
| 15 |
one way - by making inheriting (eclassing) ebuilds for all versions of kde to |
| 16 |
be supported. Right now only 2.2.1 (and the upcoming 2.2.2) have inheriting |
| 17 |
ebuilds. |
| 18 |
|
| 19 |
I will probably support only the latest versions (e.g. 2.2.1, 2.2.2). 2.2.0 |
| 20 |
can be added for completeness, since it's fairly recent and was the latest |
| 21 |
KDE in the early days of rc6. Older, already installed versions will still |
| 22 |
work - insofar they are currently working - but any small update, possibly |
| 23 |
even the emerging of some kde apps, will demand a complete upgrade to the |
| 24 |
latest qt/kde. |
| 25 |
|
| 26 |
I could also add support for all kde versions (incl. 2.1.x). This will |
| 27 |
require a lot more work and so I will only do it if enough people respond |
| 28 |
here saying they have these older versions. In which case, they'll also have |
| 29 |
to give their reasons for not upgrading to kde-2.2.x. |
| 30 |
|
| 31 |
Doing this would be tedious and would require a lot more testing. I'm |
| 32 |
mentioning it in the interests of fairness to the many people who (may) still |
| 33 |
use older KDEs. So, speak up now or be silent forever. |
| 34 |
|
| 35 |
If I did add such support, a user would still have to remerge all his current |
| 36 |
kde. So as far as I can see there are no reasons not to upgrade. |
| 37 |
|
| 38 |
After I test & commit the multi-kdelibs ebuilds (including some eclass |
| 39 |
updates and other misc stuff), I'm considering deleting older KDE ebuilds |
| 40 |
from cvs. The 2.2_beta ebuilds obviously have to go, and I'm also considering |
| 41 |
deleting all but the latest 2.1.x and 2.2.x ebuilds - that is, all but |
| 42 |
2.1.1/2 and 2.2.2. Again, tell me if you know of any reasons not to do so. |
| 43 |
|
| 44 |
Once everything is settled, tested and committed, I'll post an appropriate |
| 45 |
message on gentoo-announce. |
| 46 |
|
| 47 |
-- |
| 48 |
|
| 49 |
Dan Armak |
| 50 |
Gentoo Linux Developer, Desktop Team |
| 51 |
Matan, Israel |
Archives