1 |
On 12/09/17 03:23, John Covici wrote: |
2 |
> On Sat, 09 Dec 2017 03:51:03 -0500, |
3 |
> Alan McKinnon wrote: |
4 |
>> |
5 |
>> On 08/12/2017 21:12, John Covici wrote: |
6 |
>>> On Fri, 08 Dec 2017 11:42:16 -0500, |
7 |
>>> Alan McKinnon wrote: |
8 |
>>>> |
9 |
>>>> On 07/12/2017 17:46, John Covici wrote: |
10 |
>>>>> On Thu, 07 Dec 2017 09:37:56 -0500, |
11 |
>>>>> Alan McKinnon wrote: |
12 |
>>>>>> |
13 |
>>>>>> On 07/12/2017 07:44, John Covici wrote: |
14 |
>>>>>>> Hi. In preparing for the profile switch and the emerge -e world, I |
15 |
>> |
16 |
>> |
17 |
>> [snip] |
18 |
>> |
19 |
>> |
20 |
>>>> No, I don't think you should revert the profile change. I understood |
21 |
>>>> from your mail than you had not done that yet, and typed accordingly. |
22 |
>>>> |
23 |
>>>> I think Michael is on the right track with backtrack - set it to |
24 |
>>>> something very high like 1000, see if that gets to a solution. |
25 |
>>> |
26 |
>>> |
27 |
>>> I did switch back, but the only way I could do a "successful" update |
28 |
>>> was to mask off 5.26 and then it skipped the update and would have |
29 |
>>> been successful. If I switch to the new profile, I can do nothing as |
30 |
>>> far as perl goes. I will show the output of just trying to emerge |
31 |
>>> below, it seems there were many many packages still requiring 5.24. |
32 |
>> |
33 |
>> No, that's not right. The tree is consistent and portage can figure out |
34 |
>> how to get from perl-5.24 to perl-5.26 |
35 |
>> |
36 |
>> You probably have a difference locally, I would search through |
37 |
>> /etc/portage looking for entries that mask some perl modules and peg |
38 |
>> them to 5.24 versions. |
39 |
>> |
40 |
>> Failing that, maybe you have a package installed that depends on a 5.24 |
41 |
>> version of some module and this is the ripple effect |
42 |
>> |
43 |
>> Perhaps run emerge with "--verbose-conflicts" and also "emerge -e world" |
44 |
>> and post the results |
45 |
>> |
46 |
>> |
47 |
>>> This is with the new profile and backtrack set to 500. |
48 |
>>> |
49 |
>>> instances within a single package slot have been pulled |
50 |
>>> !!! into the dependency graph, resulting in a slot conflict: |
51 |
>>> |
52 |
>>> dev-lang/perl:0 |
53 |
>>> |
54 |
>>> (dev-lang/perl-5.26.1-r1:0/5.26::gentoo, ebuild scheduled for merge) |
55 |
>>> pulled in by |
56 |
>>> =dev-lang/perl-5.26* required by |
57 |
>>> (virtual/perl-ExtUtils-Manifest-1.700.0-r4:0/0::gentoo, installed) |
58 |
>>> ^ ^^^^^ |
59 |
>>> dev-lang/perl (Argument) |
60 |
>>> (and 13 more with the same problems) |
61 |
>>> |
62 |
>>> (dev-lang/perl-5.24.3:0/5.24::gentoo, installed) pulled in by |
63 |
>>> =dev-lang/perl-5.24* required by |
64 |
>>> (virtual/perl-Term-ANSIColor-4.40.0-r1:0/0::gentoo, installed) |
65 |
>>> ^ ^^^^^ |
66 |
>>> dev-lang/perl:0/5.24= required by |
67 |
>>> (dev-perl/XML-Twig-3.520.0:0/0::gentoo, installed) |
68 |
>>> ^^^^^^^^ |
69 |
>>> (and 260 more with the same problems) |
70 |
>>> |
71 |
>>> NOTE: Use the '--verbose-conflicts' option to display parents omitted |
72 |
>>> above |
73 |
>>> |
74 |
>>> It may be possible to solve this problem by using package.mask to |
75 |
>>> prevent one of those packages from being selected. However, it is also |
76 |
>>> possible that conflicting dependencies exist such that they are |
77 |
>>> impossible to satisfy simultaneously. If such a conflict exists in |
78 |
>>> the dependencies of two different packages, then those packages can |
79 |
>>> not be installed simultaneously. |
80 |
>>> |
81 |
>>> For more information, see MASKED PACKAGES section in the emerge man |
82 |
>>> page or refer to the Gentoo Handbook. |
83 |
> |
84 |
> hmmm, nothing masked as far as perl modules, I will look at |
85 |
> verbose-conflicts and maybe write down all those modules and start |
86 |
> unmerging and see if eventually portage can figure out something -- I |
87 |
> don't really want to do that, however I will look at the conflicts |
88 |
> and see what I can find. |
89 |
> |
90 |
> |
91 |
|
92 |
I had a lot of problems with the perl updates as well, and could not get |
93 |
it to resolve. I wasted over an hour trying to resolve it (my poor |
94 |
Celeron would take 5-10 minutes trying to calculate dependencies, and I |
95 |
had to do this 6-7 times.) |
96 |
|
97 |
Note, what I did worked for me and may not work for you, so use this |
98 |
advice at your own risk: I emerged the new perl with --nodeps, and |
99 |
invoked `perl-cleaner all` to fix the mess afterwards. It had everything |
100 |
resolved in < 10 minutes. I didn't suffer any system breakage from using |
101 |
the sledgehammer approach, but others may not be so lucky... so, as I |
102 |
said, try it at your own risk. |
103 |
|
104 |
Dan |