1 |
On Wed, 2007-03-28 at 16:29 +0000, Nelson Batalha wrote: |
2 |
> I don't know python, but I'm just curious why doesn't catalyst, in this |
3 |
> stage, just export variables like $target/use to their default locations |
4 |
> like make.conf and use portage as normal? (therefore correcting this issue) |
5 |
> (just asking, not suggesting :P). |
6 |
|
7 |
That is essentially what we're going to do. The problem is that there |
8 |
are certain situations where we don't want the changes made to "stick" |
9 |
in the make.conf and possibly break it or pollute it. We also have to |
10 |
work out quite a few other things that are being done in the |
11 |
environment, rather than in the configurations. As I said, it's an |
12 |
enormous undertaking and not something to be taken lightly. We will |
13 |
likely be completely breaking catalyst when we do this, so we're waiting |
14 |
until we branch to 2.1 before touching it. |
15 |
|
16 |
> What else is it doing at this stage other then exporting variables and call |
17 |
> emerge, that requires this change in portage? Wouldn't catalyst and the |
18 |
> livecd's be easier to maintain like this? |
19 |
|
20 |
Huh? |
21 |
|
22 |
Portage has nothing to do with it. The problem is that environment |
23 |
trumps all, and catalyst uses the environment. It's as simple as that. |
24 |
The fix, however, is not simple and not something I would accept at this |
25 |
point, even if someone else wrote a *perfect* patch for it, since it |
26 |
changes how catalyst behaves in a manner which I would prefer save for |
27 |
the next minor release, rather than a point release. |
28 |
|
29 |
-- |
30 |
Chris Gianelloni |
31 |
Release Engineering Strategic Lead |
32 |
Alpha/AMD64/x86 Architecture Teams |
33 |
Games Developer/Council Member/Foundation Trustee |
34 |
Gentoo Foundation |