| 1 |
On Wed, 2004-10-27 at 22:58 +0200, Pieter Van den Abeele wrote: |
| 2 |
> On 27 Oct 2004, at 22:14, Chris Gianelloni wrote: |
| 3 |
> |
| 4 |
> > Also, why is it this is the first time I have *ever* heard about this |
| 5 |
> > when there have been "anti-GRP" discussions for as long as GRP has been |
| 6 |
> > around and not once have I heard mention of any alternative, either |
| 7 |
> > in-the-works, or even just planned? |
| 8 |
> |
| 9 |
> Releng didn't exist yet when I wrote the proposal. I wrote a bunch of |
| 10 |
> proposals back then, one of them was dropping stage2 and stage3 for |
| 11 |
> release and replacing them with extra grp. I also suggested using |
| 12 |
> quickpkg for recreating binaries from one "live-grp" cd. I also |
| 13 |
> remember klieber and drobbins having endless arguments at the time |
| 14 |
> about whether GRP was evil or good (and gentoo being in an impasse |
| 15 |
> because drobbins got his root access removed). |
| 16 |
|
| 17 |
That explains quite a bit. My idea is also to use quickpkg to get the |
| 18 |
system contents with one slight difference, we would still be providing |
| 19 |
a stage3 tarball. The reason for this is simple, we remove a *ton* of |
| 20 |
"system" packages for the LiveCD that we simply don't need but that |
| 21 |
should be on a running Gentoo system. Now, it might be that when we |
| 22 |
start looking at building a complete environment LiveCD that we need to |
| 23 |
retain these packages. In that case, we would be able to build the |
| 24 |
stages using quickpkg, also. |
| 25 |
|
| 26 |
As I have said before, I don't really like GRP much myself, but I like |
| 27 |
what it provides for our users. If we can provide nearly the same |
| 28 |
functionality (or better) then I would definitely jump at the |
| 29 |
opportunity. |
| 30 |
|
| 31 |
> Back then each arch (only x86, ppc and sparc I think) had its own |
| 32 |
> livecd and stages script. On ppc, we never built GRP as a separate |
| 33 |
> target, we just built X-enabled livecds and GRP resulted from that |
| 34 |
> process. (live-grp being almost a logical consequence). I'm quite sure |
| 35 |
> I referred to the live-grp idea in the emails to releng explaining my |
| 36 |
> catalyst enhancements (one of them being that GRP should be used to |
| 37 |
> build livecd-stage1, because on ppc we never build grp as a separate |
| 38 |
> target...). |
| 39 |
|
| 40 |
Well, I am thinking along the same lines, except that livecd-stage1 |
| 41 |
would be the place for all the packages to be added and there would be |
| 42 |
no GRP build, at all. Essentially it is the same thing as doing it the |
| 43 |
other way around. |
| 44 |
|
| 45 |
I am CC'ing -releng on this as I think we've diverged enough from -core |
| 46 |
discussions and we should continue this there. I welcome anyone |
| 47 |
interested in this and other release discussions to join the |
| 48 |
gentoo-releng mailing list. |
| 49 |
|
| 50 |
-- |
| 51 |
Chris Gianelloni |
| 52 |
Release Engineering - Operational/QA Manager |
| 53 |
Games - Developer |
| 54 |
Gentoo Linux |
| 55 |
|
| 56 |
-- |
| 57 |
gentoo-releng@g.o mailing list |
Archives