1 |
Mike Arthur posted <200604201612.50223.mike@×××××××××××××.uk>, excerpted |
2 |
below, on Thu, 20 Apr 2006 16:12:50 +0100: |
3 |
|
4 |
> I won't be turning off IA32 emulation, still need 32-bit |
5 |
> packages, just rather manage them myself in the chroot. |
6 |
|
7 |
OK. /That's/ what you meant in the part of the OP I didn't quite follow! |
8 |
<g> |
9 |
|
10 |
> Is there any ETA/plans for a multi-arch portage working, i.e. merging |
11 |
> 32-bit and 64-bit non-bin packages with portage. |
12 |
|
13 |
It has been on the list for some time, but no ETA that I'm aware of yet. |
14 |
|
15 |
FWIW, portage-2.1 is on the way (regularly updated pre's in ~arch for some |
16 |
time now, talk of an rc on the way), with the new features that have been |
17 |
getting all the work (better security/signing, a portage API level that an |
18 |
ebuild can set to tell portage that it needs to be at least that level to |
19 |
properly handle the ebuild, better cache and metadata management, cleaner |
20 |
and more modular code, and confcache, to name some of them). |
21 |
|
22 |
In particular the portage API level thing will allow faster portage |
23 |
development without having to wait nearly a year to be sure everyone has |
24 |
upgraded from their old and incompatible versions before the old code can |
25 |
be safely removed from the tree and portage. The goal is to then speed up |
26 |
portage development, with faster releases every few months. One of those |
27 |
releases will probably introduce multi-arch tracking. However, |
28 |
successfully implementing it has some serious complexities (I didn't |
29 |
appreciate how complex proper dependency resolution can be until I joined |
30 |
the portage-dev list and started reading -- I have some /serious/ respect |
31 |
for the code hackers that dare to tread that way, now, and that's just |
32 |
for a /single/ arch/abi!), and it's not going to be an easy job. I expect |
33 |
we're still looking at 8 months to a year out, before it'd be anywhere |
34 |
close to stable. That assumes a portage hacker gets seriously interested |
35 |
in the problem, as well. If not, it could be longer, tho it'll likely |
36 |
happen eventually. |
37 |
|
38 |
OTOH, as time progresses, the need for it will become less urgent, as more |
39 |
and more stuff will be available for the now default 64-bit platform. At |
40 |
this point, that's already the case for most open source stuff, with a few |
41 |
exceptions such as Open Office. With MS now onboard the 64-bit bandwagon, |
42 |
even if they /are/ a year or so behind, in a couple years most of the |
43 |
slaveryware closed source stuff like games and codecs will be available in |
44 |
64-bit as well. Thus, ironically, we'll probably be finally getting |
45 |
multi-arch package manager support, just as it's no longer really needed. |
46 |
|
47 |
-- |
48 |
Duncan - List replies preferred. No HTML msgs. |
49 |
"Every nonfree program has a lord, a master -- |
50 |
and if you use the program, he is your master." Richard Stallman in |
51 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
52 |
|
53 |
|
54 |
-- |
55 |
gentoo-amd64@g.o mailing list |