1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA1 |
3 |
|
4 |
On Saturday 26 June 2004 17:06, Robin H. Johnson wrote: |
5 |
> On Sat, Jun 26, 2004 at 03:06:52PM +0900, Jason Stubbs wrote: |
6 |
> > Issue #1: Ignoring dependencies |
7 |
> > The "cleaning" of the vardb to trick portage is IMO a bad thing. There is |
8 |
> > obviously enough in a stage 1 to be able to build up all of system, but |
9 |
> > according to the data portage it is impossible. This then requires a hack |
10 |
> > as indicated above to attempt the merge anyway. However, this hack |
11 |
> > affects users of installed systems as well as portage will go ahead and |
12 |
> > attempt a compilation that is guaranteed to fail. |
13 |
> |
14 |
> What's your solution for this? |
15 |
|
16 |
Well, I don't actually understand why the var db needs to be cleaned. I can't |
17 |
see why a combination of --nodeps and --onlydeps on package x and y can't |
18 |
solve whatever the problem is, though... |
19 |
|
20 |
> > Issue #2: Lack of dependency information |
21 |
> > Looking at the above, linux-headers doesn't need bzip2 to unpack, |
22 |
> > ncompress and db don't need glibc to build and almost nothing needs gcc |
23 |
> > to compile. It gets much worse when doing emerge -ep world. If it's too |
24 |
> > much of a PITA to fix this for all packages, portage could ensure that |
25 |
> > all of system is installed before anything else, but this needs to be |
26 |
> > 100% explicit for the base system. I don't believe it's as difficult as |
27 |
> > it sounds - a few eclasses such as bz2files, csource, cppsource, etc |
28 |
> > should be sufficient. However, without accurate information, parallel |
29 |
> > emerges are just a daydream. |
30 |
> |
31 |
> I've got no objections to putting the correct information into the tree, |
32 |
> but portage needs to be able to handle circular dependancies much better |
33 |
> first. One other glaring problem in your listing is sys-devel/make. One |
34 |
> other thing that will bite at some points is the things that some |
35 |
> upstream developers put into their configure scripts, that are decidedly |
36 |
> not always present on a system (perl and rpm to name the worst offenders |
37 |
> I've seen). If you can tell me I won't run into any problems by adding |
38 |
> in the deps for virtual/libc virtual/cc sys-devel/make etc into my |
39 |
> packages, I'll go around and correct them now. |
40 |
|
41 |
I don't think that the change would bring about any more bad dependency |
42 |
ordering than already exists, but I can't be sure. Perhaps, the solution is |
43 |
to fix the dependency resolver then fix the packages and then enable parallel |
44 |
emerges. That sound acceptable? |
45 |
|
46 |
> virtual/cc is something I think should exist, as some people want to use |
47 |
> dev-lang/icc or dev-lang/ccc instead of sys-devel/gcc. |
48 |
> |
49 |
> Also, where does the adding of dependancies stop? Eg I have something |
50 |
> that uses kernel-headers and glibc, so do I just specify glibc and |
51 |
> ignore kernel-headers as that's needed by glibc? |
52 |
|
53 |
That's a design question that's open to debate. Portage can do it either way, |
54 |
but I would suggest that the package depend on both kernel-headers and glibc |
55 |
for robustness in the tree. Unlikely in this case, but it ensures that a |
56 |
change to glibc's dependencies don't break the packages that depend on it. |
57 |
|
58 |
Regards, |
59 |
Jason Stubbs |
60 |
-----BEGIN PGP SIGNATURE----- |
61 |
Version: GnuPG v1.2.4 (GNU/Linux) |
62 |
|
63 |
iQCVAwUBQN1Bh1oikN4/5jfsAQIn1AP+JpYCphzGsFLJlm6hnntLfGaJ+KWjjb61 |
64 |
UNR4sw7q/qRq4GfNliBfNetb16K6jjBRseZUrt9p5ofdjgu6nCxEyfl5F+0QyPSH |
65 |
sAZq0eqFcobARU+TUPGGBeJG82ve46BJDCTPF4ieK3dFm4/0sfRYHz+DAYQnNFG6 |
66 |
eq507/u1bRM= |
67 |
=3rkH |
68 |
-----END PGP SIGNATURE----- |
69 |
|
70 |
-- |
71 |
gentoo-dev@g.o mailing list |