1 |
On Fri, 2018-06-15 at 20:58 +-0700, grozin+AEA-gentoo.org wrote: |
2 |
+AD4 Hello +ACo, |
3 |
+AD4 |
4 |
+AD4 For a long time I was scared to switch to the 17.0 profile - I was |
5 |
+AD4 sure |
6 |
+AD4 that recompiling the world will be a difficult and time-consuming |
7 |
+AD4 affair |
8 |
+AD4 with lots of errors. Finally, I decided to switch my office +AH4-x86 |
9 |
+AD4 desktop |
10 |
+AD4 (a rather old and weak desktop with a single-core pentium E5300). I |
11 |
+AD4 was |
12 |
+AD4 right - now most of useful stuff does not work. I am even more afraid |
13 |
+AD4 to |
14 |
+AD4 do it with my main computer - a +AH4-amd64 laptop with a 4-core intel i7. |
15 |
+AD4 I |
16 |
+AD4 really need it to work normally. |
17 |
+AD4 |
18 |
+AD4 OK, now a question. I'm trying to restore normal working of cblas, |
19 |
+AD4 blas, |
20 |
+AD4 lapack and all programs based on them. I use the science overlay. |
21 |
+AD4 I've |
22 |
+AD4 removed (by force, with -C) cblas-reference, blas-reference, |
23 |
+AD4 lapack-reference and the corresponding virtuals. I try to follow |
24 |
+AD4 |
25 |
+AD4 https://wiki.gentoo.org/wiki/User+AF8-talk:Houseofsuns |
26 |
+AD4 |
27 |
+AD4 I've installed app-admin/eselect::science. Emerging |
28 |
+AD4 sci-libs/blas-reference::science fails with sandbox violation on |
29 |
+AD4 trying to |
30 |
+AD4 mkdir /etc/env.d/alternatives/blas. If I create it by hand and try |
31 |
+AD4 again, |
32 |
+AD4 then sandbox violation on mkdir |
33 |
+AD4 /etc/env.d/alternatives/blas/reference. If |
34 |
+AD4 I remove sandbox from FEATURES in my /etc/portage/make.conf, nothing |
35 |
+AD4 changes - still sandbox violations. |
36 |
+AD4 |
37 |
+AD4 How can I emerge blas-reference::science? |
38 |
+AD4 |
39 |
+AD4 Andrey |
40 |
|
41 |
|
42 |
I have similar errors, see e.g.: https://bugs.gentoo.org/658170, |
43 |
https://bugs.gentoo.org/657984, https://bugs.gentoo.org/658034 or https |
44 |
://bugs.gentoo.org/658026. |
45 |
I have the feeling that theese problems with blas, lapack, atlas etc. |
46 |
are not directly related to the 17.0 profile. I switched to the 17.0 |
47 |
profile in December and begin of January and I was still able to emerge |
48 |
atlas Jan 3rd: |
49 |
|
50 |
rose+AEA-lynxold:/home/rose/Txt/personal(132)+ACQ genlop -t atlas |
51 |
+ACo sci-libs/atlas |
52 |
|
53 |
Sat Jul 8 19:51:01 2017 +AD4APgA+ sci-libs/atlas-3.11.39 |
54 |
merge time: 2 hours, 23 minutes and 17 seconds. |
55 |
|
56 |
Mon Jan 1 15:26:46 2018 +AD4APgA+ sci-libs/atlas-3.11.39 |
57 |
merge time: 2 hours, 10 minutes and 44 seconds. |
58 |
|
59 |
Wed Jan 3 06:00:13 2018 +AD4APgA+ sci-libs/atlas-3.11.39 |
60 |
merge time: 2 hours, 39 minutes and 32 seconds. |
61 |
|
62 |
|
63 |
So maybe the new eselect is responsible for the issues. |
64 |
|
65 |
Juergen |