1 |
Hi, |
2 |
|
3 |
Duncan wrote: |
4 |
> I'm repeating what Simon said, but sometimes it helps to get it said a |
5 |
[snip] |
6 |
> The immediate user-level fix, as Simon says, is kill the |
7 |
[snip] |
8 |
> Here's the background. As Simon explained, in ordered to be able to have |
9 |
|
10 |
Just a personal note on this: What's the point in repeating everything? |
11 |
You only bore people with it. I think we can assume that the original |
12 |
poster either gets it the first time, or he'll surely ask again and give |
13 |
details about what he didn't understand. |
14 |
|
15 |
> First, there's the binary compatibility issue, both with legacy apps and |
16 |
> moving forward. It's simply harder to get binaries from other sources to |
17 |
> work, when they are designed with one set of assumptions about locations, |
18 |
> but installed on a system with a rather different set of assumptions. As |
19 |
> I said, there's the 32-bit legacy thing, but consider continuing effect |
20 |
> going forward, as well. If we stayed the way we are, every binary-only |
21 |
> installation a Gentoo-amd64 user ever made, including the new amd64 |
22 |
> binaries, when they become common, would be a fight. Despite the temporary |
23 |
|
24 |
Why? When the binary gets linked, ldd just looks into various places and |
25 |
decides which library to take depending on the bitness. There's no |
26 |
fight, there's /etc/ld.so.conf. Sorry, but that paragraph is fubar. |
27 |
|
28 |
> mostly duplicated common code, it's ALWAYS more efficient to keep in line |
29 |
> with the rest of the community, so the maintenance and further development |
30 |
|
31 |
I disagree: There are bad standards, which should be replaced by better |
32 |
ones. You can't do that by keeping in line with the rest. |
33 |
|
34 |
> Thus, the upstream compatibility issue resolves to this: the closer we |
35 |
> are to accepted LSB/FHS standard, the less work we will have to do making |
36 |
> changes so that open source packages which generally target the loose |
37 |
|
38 |
Wrong. Correct would be: the closer we are to UPSTREAM, the less work we |
39 |
will have to do. A nice example: QT |
40 |
|
41 |
> Together, these two reasons means the only /possible/ sensible alternative |
42 |
> is to bite that bullet, and exchange some temporary misery now, while the |
43 |
> conversion is in progress, for a /far/ more standards compliant multilib |
44 |
> setup, going forward. |
45 |
|
46 |
What misery? |
47 |
|
48 |
> Currently, Gentoo-amd64 has just about finished moving 64-bit libs out of |
49 |
> lib. We have both a lib32, and a lib64, with lib being a symlink to |
50 |
> lib64, for those remaining packages (such as skencil, apparently) that |
51 |
> have so far fallen thru the cracks, and continue to make the early |
52 |
> Gentoo-amd64 assumption that lib is the 64-bit shared object location. |
53 |
> With 2005.1, I'm expecting FEATURES=multilib-strict will be the default, |
54 |
> to finally weed out as many of the few remaining back packages as |
55 |
> possible. With 2006.0, it's possible that lib -> lib64 symlink can |
56 |
> finally be broken, and any remaining packages then WILL get bugs filed, |
57 |
> when someone tries to use them and has issues. With luck, by 2006.1, it |
58 |
> should be safe to start doing basically the same thing with the lib32 -> |
59 |
> lib move, as we will have just got thru doing with the lib -> lib64 move. |
60 |
> First, there will be a symlink between the two, in another release or two, |
61 |
> the main packages will be changed and it'll be time to activate the |
62 |
> multilib-strict test for anything still ending up in lib32 instead of lib. |
63 |
> By 2007.0, therefore, if luck holds, that multilib-strict test can be the |
64 |
> default, to catch all the stragglers possible, and 2007.1 should then |
65 |
> hopefully be able to remove lib32, relegating it to the annuls of |
66 |
> Gentoo-amd64 history. (Those .1 mentions assume that Gentoo continues |
67 |
> with the twice-yearly releases, thus, they mean second-half.) |
68 |
|
69 |
Where do you get this information from? We (read: we, the Gentoo/AMD64 |
70 |
developer team) never published such a roadmap, nor do we have one |
71 |
ourselves. Please, don't cook up a story just to enlarge your |
72 |
pen^W^W^Wmails. |
73 |
|
74 |
> However, that's really only half the issue. The other half is portage, |
75 |
> which currently can't track 32-bit dependencies separate from its 64-bit |
76 |
> dependencies. We had hoped that portage would have dual-bitness support |
77 |
> by 2005.1, but that's now looking impossible, since the new portage |
78 |
> remains only in CVS, there hasn't yet been even the usual hard masked for |
79 |
> testing alpha snapshot releases, let alone the -preX versions, which would |
80 |
> be the first ones that could even theoretically leave hard mask testing, |
81 |
> for ~arch testing. Thus, it'll be 2006.0 at the earliest, before portage |
82 |
> will be able to handle dual-bitness tracking, keeping the 32-bit |
83 |
> dependencies separate from the 64-bit dependencies. Until then, |
84 |
> Gentoo-amd64 32-bit support will remain spotty at best, requiring users |
85 |
> either install from tarball, or use a 32-bit chroot with its own portage |
86 |
> tracking 32-bit dependencies, for pretty much anything beyond the core |
87 |
> system libraries (sandbox, gcc, glibc, now from-source supported, other |
88 |
> 32-bit core libraries managed as binary-only packages or from the chroot). |
89 |
|
90 |
You can take comfort with the fact that even if portage already had this |
91 |
functionality, there wouldn't be real multilib now. |
92 |
|
93 |
Regards, |
94 |
|
95 |
-- |
96 |
Simon Stelling |
97 |
Gentoo/AMD64 Operational Co-Lead |
98 |
blubb@g.o |
99 |
-- |
100 |
gentoo-amd64@g.o mailing list |