1 |
El sáb, 30-06-2012 a las 13:46 -0400, Ian Stakenvicius escribió: |
2 |
> -----BEGIN PGP SIGNED MESSAGE----- |
3 |
> Hash: SHA256 |
4 |
> |
5 |
> On 30/06/12 01:30 PM, Pacho Ramos wrote: |
6 |
> > El sáb, 30-06-2012 a las 13:17 -0400, Ian Stakenvicius escribió: |
7 |
> >> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 |
8 |
> >> |
9 |
> >> On 30/06/12 11:16 AM, Mike Frysinger wrote: |
10 |
> >>> On Saturday 30 June 2012 07:22:39 Zac Medico wrote: |
11 |
> >>>> On 06/30/2012 04:07 AM, Pacho Ramos wrote: |
12 |
> >>>>> I would like to discuss a bit more issues like: |
13 |
> >>>>> https://bugs.gentoo.org/show_bug.cgi?id=423087 |
14 |
> >>>>> |
15 |
> >>>>> Even if there are "a lot" of packages that can cause this |
16 |
> >>>>> breakage when downgraded, I think it should be prevented |
17 |
> >>>>> and package managers shouldn't try to downgrade this kind |
18 |
> >>>>> of packages as they will later cause a total breakage. |
19 |
> >>>>> People is not supposed to know that downgrading some |
20 |
> >>>>> package system will, for example, have an unusable gcc. |
21 |
> >>>> |
22 |
> >>>> It seems like a die in pkg_pretend would serve pretty well. |
23 |
> >>> |
24 |
> >>> doing it on a per-ebuild basis doesn't make much sense. a |
25 |
> >>> simple version compare (like we do in glibc as an exception to |
26 |
> >>> this rule because of its much wider implication) is incorrect: |
27 |
> >>> the new version might not introduce any new symbols compared to |
28 |
> >>> the old one, and even if it has, other packages might not have |
29 |
> >>> been linked against the new symbols. -mike |
30 |
> >> |
31 |
> >> Instead of preventing downgrade wouldn't it make more sense to |
32 |
> >> figure out a way to force a rebuild on @system or @toolchain or |
33 |
> >> whatever bits are broken as soon as the downgrade occurs, rather |
34 |
> >> than just making it a one-way ticket? If we could sort this out |
35 |
> >> (and sub-slots may help with this, but probably we'll need some |
36 |
> >> extra work too) then we could probably support switching from |
37 |
> >> ~arch to arch at a whim.. Not necessarily a bad goal. |
38 |
> >> |
39 |
> > |
40 |
> > The problem is that, in this kind of breakage, gcc breaks as soon |
41 |
> > as zlib is downgraded and, then, user cannot compile anything, |
42 |
> > needing to manually find missing zlib lib from any other |
43 |
> > distributions binaries, put it in the system and re-emerge zlib :| |
44 |
> > |
45 |
> |
46 |
> ..but preserved-libs would keep that from happening wouldn't it? ie, |
47 |
> the lib itself would stick around until gcc isn't using it anymore... |
48 |
> |
49 |
> so it'd just be a matter of an interim issue until preserved-libs is |
50 |
> in stable portage ... and i'm guessing something that would suffice |
51 |
> here is a blockage on downgrades of anything belonging to the contents |
52 |
> of /var/db/edb/{sys-devel/gcc,sys-libs/glibc}-[0-9]*/NEEDED ? |
53 |
> |
54 |
> (apologies for the bad hack:) |
55 |
> |
56 |
> cat /var/db/pkg/{sys-devel/gcc,sys-libs/glibc}-[0-9]*/NEEDED \ |
57 |
> |sed -e 's#^/[^ ]* ##' |sed -e "s/,/\n/g" |sort -u \ |
58 |
> |xargs equery b |awk '{print $1}' |sort -u \ |
59 |
> |sed 's/^/</' |
60 |
> |
61 |
> ^^^ that's your build-preserving package.mask , yes? |
62 |
> -----BEGIN PGP SIGNATURE----- |
63 |
> Version: GnuPG v2.0.19 (GNU/Linux) |
64 |
> |
65 |
> iF4EAREIAAYFAk/vO4gACgkQ2ugaI38ACPBUvwEApQAljVOj492q2/bhttriqWgz |
66 |
> iu8FZdsh1EHMeYaHxi0A/iZNY28W9NT5ynO6B42CAxpYpWym2SIc4JflTu/7IK1h |
67 |
> =3pcd |
68 |
> -----END PGP SIGNATURE----- |
69 |
> |
70 |
> |
71 |
|
72 |
Looks like preserve-libs should be extended to handle this cases: |
73 |
https://bugs.gentoo.org/show_bug.cgi?id=423087#c5 |