1 |
On Sun, 10 Dec 2017 08:13:09 -0500, |
2 |
Alan McKinnon wrote: |
3 |
> |
4 |
> On 10/12/2017 14:54, John Covici wrote: |
5 |
> > On Sun, 10 Dec 2017 07:36:19 -0500, |
6 |
> > Kent Fredric wrote: |
7 |
> >> |
8 |
> >> [1 <text/plain; US-ASCII (quoted-printable)>] |
9 |
> >> On Sun, 10 Dec 2017 02:17:09 -0500 |
10 |
> >> John Covici <covici@××××××××××.com> wrote: |
11 |
> >> |
12 |
> >>> OK, thanks, I think I will try that. |
13 |
> >> |
14 |
> >> The problem you're facing is that you masked dev-lang/perl, but not any |
15 |
> >> virtual/perl-* or perl-core/-* to compensate. |
16 |
> >> |
17 |
> >> These 3 components work in concert like a single component, as a sort |
18 |
> >> of bodge to compensate for the fact portage has no working "provides" feature, |
19 |
> >> and to compensate for the dependency-system missmatch between how |
20 |
> >> Gentoo works and how CPAN works. |
21 |
> >> |
22 |
> >> Theres' no easy way of fixing this atm, but the short of it is if you're using |
23 |
> >> an ~arch dev-lang/perl, you should be using an ~arch virtual/perl-*, |
24 |
> >> and if you're using an "arch" dev-lang/perl, you should be using only |
25 |
> >> "arch" versions of virtual/perl-* |
26 |
> >> |
27 |
> >> Once you do this, portage may still scream at you, because portage is |
28 |
> >> very much optimised for upgrading, and it tends to think downgrading is |
29 |
> >> an error. |
30 |
> >> |
31 |
> >> So once you get all your masks/keyword changes in place, you should do: |
32 |
> >> |
33 |
> >> emerge -C virtual/perl-* |
34 |
> >> emerge -C perl-core/* |
35 |
> >> |
36 |
> >> (or something to that effect) |
37 |
> >> |
38 |
> >> This looks scary, but generally isn't, because you're not actually removing |
39 |
> >> anything with this, just juggling a few balls and making only older |
40 |
> >> versions of certain things available ( as they're alls shipped in |
41 |
> >> dev-lang/perl ) |
42 |
> >> |
43 |
> >> And then after you do this, portage is more likely to be persuadable |
44 |
> >> into doing the right thing. |
45 |
> >> |
46 |
> >> You can additionally abuse my tool, gentoo-perl-helpers for doing some of this, |
47 |
> >> and some of the steps I've described are automated because they're just |
48 |
> >> that safe and useful. |
49 |
> >> |
50 |
> >> https://wiki.gentoo.org/wiki/Perl#app-admin.2Fgentoo-perl-helpers |
51 |
> >> |
52 |
> >> |
53 |
> >> After putting the right masks in place, do: |
54 |
> >> |
55 |
> >> gentoo-perl gen-upgrade-sets 5.26 5.24 |
56 |
> >> |
57 |
> >> And if you're really lucky, the sets it generates will work the first time :) |
58 |
> >> |
59 |
> >> ( I actually tested this scenario when developing it, but its still an |
60 |
> >> undocumented use on purpose ) |
61 |
> >> |
62 |
> >> GLHF. |
63 |
> > |
64 |
> > I went ahead and did the upgrade which worked, but the emerge from |
65 |
> > perl-cleaner --all did not. I am using ~amd64 and have done so for |
66 |
> > years, so I don't think I need to maks off anything. I seem now to be |
67 |
> > stuck with dev-python/setuptools, so I am now trying to figure out why |
68 |
> > I can't emerge that -- it was triggered by the perl-cleaner --all . |
69 |
> > |
70 |
> |
71 |
> How recent is your tree? |
72 |
> |
73 |
> I had issues with setuptools doing the first run through the 17.0 |
74 |
> upgrade. I never looked into it too closely, I used --keep-going, but |
75 |
> setuptools seemed to think I had a useable python-3.4 |
76 |
> |
77 |
> After the first run through emerge -e world, nuking-python-3.4 and |
78 |
> re-syncing, setuptols worked normally again. |
79 |
> |
80 |
> YMMV of course where you are |
81 |
|
82 |
My tree maybe 30 days old or thereabouts. I could not run the emerge |
83 |
from perl-cleaner because of setup-tools problems. I will see what |
84 |
happens if I run a regular update, but I hate to do that if I am going |
85 |
to do an -e world. |
86 |
|
87 |
-- |
88 |
Your life is like a penny. You're going to lose it. The question is: |
89 |
How do |
90 |
you spend it? |
91 |
|
92 |
John Covici |
93 |
covici@××××××××××.com |