1 |
Lars Weiler wrote: |
2 |
> Another upgrade for 2006.1 might be Xorg-7.1. ppc does not |
3 |
> rely on any binary-driver which is affected by the |
4 |
> ABI-change. This version of X has been tested by several |
5 |
> developers and found to be stable. But we need to check |
6 |
> Donnie's opinion about that move. |
7 |
> |
8 |
> A pro-argument would be, that users have to do one big |
9 |
> change with 2006.1 (toolchain, X) and can rest for a long |
10 |
> time. A good upgrade-path for Xorg-7.1 is needed for the |
11 |
> users who won't to install their system newly. |
12 |
|
13 |
I'm all for arches that aren't help back by slowness of binary driver |
14 |
updates to stabilize on Xorg 7.1. So far, I've heard interest from PPC |
15 |
and MIPS. |
16 |
|
17 |
It seems that all of ~arch is now ported to modular X, but I'm unsure |
18 |
about stable. My current "unported ebuilds" scripts don't seem to |
19 |
support checking for unported stable ebuilds, just ~arch. They're |
20 |
available [1] as find_all_broken_modular.sh, |
21 |
find_broken_modular_package.sh and find_broken_modular_package.py. |
22 |
|
23 |
I think the Python one needs fixing, but I'm not much for Python. I |
24 |
suspect this line may be the problem, but I don't know: "cpv = |
25 |
portage.portdb.xmatch("bestmatch-visible", cp)". |
26 |
|
27 |
Kevin Quinn's suggested that hardened should mask modular X until we can |
28 |
get the weirdness with the hardened toolchain + X sorted (seems that |
29 |
binutils 2.17 will do that). |
30 |
|
31 |
The upgrade path seems reasonable to me right now. I've written a guide |
32 |
[2], but most of the steps are straightforward. The guide mainly exists |
33 |
to help with troubleshooting issues such as XKB for which configurations |
34 |
will break. |
35 |
|
36 |
Thanks, |
37 |
Donnie |
38 |
|
39 |
1. http://www.gentoo.org/proj/en/desktop/x/x11/modular-x-howto.xml |