1 |
Mike Gilbert <floppym <at> gentoo.org> writes: |
2 |
|
3 |
|
4 |
> Basically, yes. If [R]DEPEND in /var/db/pkg is different from the |
5 |
> values in the ebuilds in the tree, <at> changed-deps will select it. |
6 |
|
7 |
> > Also, these two similar commands return different results (I have |
8 |
> > bdeps=y in DEFAULT_OPTS btw): |
9 |
> > |
10 |
> > emerge -uND --changed-deps=y world (51 packages) |
11 |
> > emerge <at> changed-deps (11 packages) |
12 |
|
13 |
> > Do you know why those commands give different results? |
14 |
> > The smaller list is a strict subset of the longer one. |
15 |
|
16 |
> That difference is surprising; you would probably need to ask the |
17 |
> portage developers to get a real answer. |
18 |
|
19 |
|
20 |
I do not "own' the code underneath these options, nor would I waste my |
21 |
time on what is undoubtedly broken logic. Portage and associated tools do |
22 |
not have a complete description of the problem with the current VDB. A DAG |
23 |
is the best way to solve the problem of these portage anomalies. |
24 |
|
25 |
Or just rebuild @world and occasionally @system, and then cobble together |
26 |
a script that hopefully finds all the other installed codes and packages.... |
27 |
Still, cruft (some of the old files that should have been removed at some |
28 |
point) remains. |
29 |
|
30 |
I certainly and not saying this is an easy problem to solve; all distros |
31 |
suffer tremendously here. That's why many just 'reinstall' periodically. |
32 |
But without a robust implementation of a DAG or something similar, deep |
33 |
problems will remain unresolved and hopefully, not noticed. Many large data |
34 |
centers just 'reload' entire rows of racked systems too. |
35 |
|
36 |
These are big problems on clusters/clouds. The current approach is to build |
37 |
many 'customized frameworks' for different classes of problems, on top of |
38 |
poorly implemented distros with a collection of packages. That approach |
39 |
works well, if the packages are relegated to one complex language and |
40 |
scripts. Once a collection of codes, a package if you like, requires several |
41 |
complex programming languages to compile and execute, thus crossing |
42 |
framework boundaries, then the problems exponentiate, and it becomes |
43 |
an admin/security mess. I do believe that the current gentoo approach is |
44 |
going to be robustly application as a pristine solution for this gargantuan |
45 |
cluster problem; but a DAG sort of fundamental tooling is going to be |
46 |
minimally sufficient if stability is to be achieved. Clusters are already |
47 |
making their way to gentoo so these problem in only going to grow. |
48 |
|
49 |
|
50 |
Oddly enough, I feel like and uneducated admin in these matters. But, I do |
51 |
now have several major corps that want me to be a temporary consultant |
52 |
on cluster problems. We shall see. We shall see what's what. |
53 |
|
54 |
|
55 |
wwr, |
56 |
James |