1 |
On 09/07/2014 00:33, gottlieb@×××.edu wrote: |
2 |
> On Tue, Jul 08 2014, Neil Bothwick wrote: |
3 |
> |
4 |
>> On Tue, 08 Jul 2014 18:02:56 -0400, gottlieb@×××.edu wrote: |
5 |
>> |
6 |
>>> For some reason my mostly-stable-slightly-testing system |
7 |
>>> is trying to merge *UNstable* dev-lang/perl. |
8 |
>> |
9 |
>>> Specifically it wants me to unmask |
10 |
>>> =dev-lang/perl-5.20.0-r1 |
11 |
>> |
12 |
>> That's not testing, it's masked. |
13 |
> |
14 |
> Right; that is what I thought I said. Oh now I see I wrote *UNstable*. |
15 |
> I thought that was a synonym for masked. Will purge that term from my |
16 |
> gentoo vocabulary. |
17 |
> |
18 |
>> % eix -e perl |
19 |
>> [I] dev-lang/perl |
20 |
>> Available versions: 5.16.3(0/5.16) (~)5.18.2-r1(0/5.18){tbz2} |
21 |
>> [M](~)5.20.0-r1(0/5.20) {berkdb debug doc gdbm ithreads} |
22 |
>> |
23 |
>>> [nomerge ] perl-core/CPAN-Meta-2.132.510 |
24 |
>>> [ebuild U ~] virtual/perl-File-Temp-0.230.400-r1 [0.230.400] 0 kB |
25 |
>>> [ebuild U ] virtual/perl-File-Spec-3.400.0-r1 [3.400.0] 0 kB |
26 |
>>> [ebuild r U #] dev-lang/perl-5.20.0-r1:0/5.20 [5.16.3:0/5.16] |
27 |
>> |
28 |
>> It would appear that CPAN-Meta is the culprit, see what is pulling |
29 |
>> that in. |
30 |
> |
31 |
> That was listed in the voluminous emerge output. Here is the relevant |
32 |
> portion. It starts with xdg-utils (required by lots) and then a |
33 |
> bunch of perl. |
34 |
> |
35 |
> The following mask changes are necessary to proceed: |
36 |
> (see "package.unmask" in the portage(5) man page for more details) |
37 |
> # required by virtual/perl-ExtUtils-Command-1.180.0-r1 |
38 |
> # required by perl-core/ExtUtils-MakeMaker-6.820.0 |
39 |
> # required by virtual/perl-ExtUtils-MakeMaker-6.820.0 |
40 |
> # required by perl-core/CPAN-Meta-2.132.510 |
41 |
> # required by virtual/perl-CPAN-Meta-2.132.510 |
42 |
> # required by perl-core/Module-Build-0.400.700 |
43 |
> # required by virtual/perl-Module-Build-0.400.700 |
44 |
> # required by dev-perl/File-MimeInfo-0.170.0 |
45 |
> # required by x11-misc/xdg-utils-1.1.0_rc2[perl] |
46 |
> # required by www-client/chromium-35.0.1916.153 |
47 |
> # required by @selected |
48 |
> # required by @world (argument) |
49 |
> |
50 |
> |
51 |
>>> # Masked for testing, will defenitely cause major breakage |
52 |
>>> # Single modules with 5.20 versions added here too -dilfridge |
53 |
>>> =dev-lang/perl-5.20.0-r1 |
54 |
>> |
55 |
>> See, masked, not keyworded, and apparently with good reason. |
56 |
> |
57 |
> Right. Sorry for the masked / *UNstable* confusion. |
58 |
> I was surprised eix doesn't even show this version, but that is minor. |
59 |
|
60 |
|
61 |
I rather suspect you are falling foul of portage's recent weird and |
62 |
wonderful behaviour trying to satisfy deps[1]. Faling that, you might |
63 |
just have gotten a bad sync while someone was adding perl-5.20 to the tree |
64 |
|
65 |
Plus your eix does not match emerge, meaning you have recently sync'ed |
66 |
and not run eix-update after. |
67 |
|
68 |
Sync again and re-run emerge -avuND world, see if the latest tree has |
69 |
your issue fixed. |
70 |
|
71 |
If not, run it with --backtrack=0 and again with --backtrack=100, let's |
72 |
see if helping portage out makes a difference. |
73 |
|
74 |
Finally, my favourite trick to find settings I forgot about: |
75 |
|
76 |
egrep -ir "perl|cpan" /etc/portage |
77 |
|
78 |
|
79 |
|
80 |
[1] yes, it is getting better as we move forward. But the console output |
81 |
is still odd at times, it's still slow as molasses, and it really looks |
82 |
like sometimes it's way to smart for it's own good. |
83 |
|
84 |
|
85 |
|
86 |
-- |
87 |
Alan McKinnon |
88 |
alan.mckinnon@×××××.com |