Gentoo Archives: gentoo-science

From: fbissey@××××××××××××.nz
To: gentoo-science@l.g.o
Subject: Re: [gentoo-science] sci-mathematics/sage-4.7.1 and numpy version
Date: Wed, 24 Aug 2011 22:38:54
Message-Id: 20110825103356.jwg0ggo0gk4woo44@webmail.slingshot.co.nz
In Reply to: Re: [gentoo-science] sci-mathematics/sage-4.7.1 and numpy version by rose@rz.uni-potsdam.de
1 Quoting rose@××××××××××××××.de:
2
3 > Quoting Steven Trogdon <strogdon@×××××.edu>:
4 >
5 >> On Wed, 24 Aug 2011 12:03:19 +0200
6 >> rose@××××××××××××××.de wrote:
7 >>
8 >>> Hi,
9 >>>
10 >>> I just installed more or less succesful sage. During the installation
11 >>> numpy was downgraded from numpy-1.6.1 to numpy-1.5.1.
12 >>> Now every 'emerge -uvDN world' complaines about;
13 >>>
14 >>> WARNING: One or more updates have been skipped due to a dependency
15 >>> conflict:
16 >>>
17 >>> dev-python/numpy:0
18 >>>
19 >>> (dev-python/numpy-1.6.1::gentoo, ebuild scheduled for merge)
20 >>> conflicts with
21 >>> ~dev-python/numpy-1.5.1 required by
22 >>> (sci-mathematics/sage-4.7.1::sage-on-gentoo, installed)
23 >>>
24 >>> dev-python/pexpect:0
25 >>>
26 >>> (dev-python/pexpect-2.4-r1::sage-on-gentoo, ebuild scheduled for
27 >>> merge) conflicts with
28 >>> ~dev-python/pexpect-2.0 required by
29 >>> (sci-mathematics/sage-4.7.1::sage-on-gentoo, installed)
30 >>>
31 >>> If I try to install couchdb, numpy-1.6.1 will be installed again:
32 >>>
33 >>> root@caiman:/root(4)# emerge -pvuDN couchdb
34 >>>
35 >>> These are the packages that would be merged, in order:
36 >>>
37 >>> Calculating dependencies... done!
38 >>> [ebuild N ] dev-lang/erlang-14.2.3 USE="doc emacs java odbc smp
39 >>> ssl tk wxwidgets -halfword -hipe -kpoll -sctp" 98,075 kB
40 >>> [ebuild N ] dev-db/couchdb-1.1.0 USE="-test" 1,063 kB
41 >>> [ebuild U ] dev-python/numpy-1.6.1 [1.5.1] USE="doc lapack -test" 0 kB
42 >>>
43 >>> Should I mask numpy-1.6.1 or how are the chances that sage will work
44 >>> with numpy-1.6.1?
45 >>>
46 >> At least until this Bug
47 >>
48 >> https://bugs.gentoo.org/show_bug.cgi?id=371099
49 >>
50 >> is resolved (sog) sage requires ~numpy-1.5.1
51 >>
52 >> I'm not sure exactly the impact on your system of masking
53 >> >=numpy-1.6.0 but you
54 >> could try that and see what emerge -puDNv world gives. I don't see which
55 >> component, if any, of couchdb is pulling in numpy or where else numpy is
56 >> needed on your system. Oh, and you will also have to mask >=pexpect-2.4
57 >>
58 >> Steve
59 >>
60 >
61 > Thanks Steve,
62 >
63 > OK, I masked:
64 >
65 >> =dev-python/numpy-1.6.0
66 >> =dev-python/pexpect-2.4
67 >
68 > The following installation of couchdb was no problem, also 'emerge
69 > -puvDN world' did not complain about any conflicts. But I made the
70 > mistake to delete scipy. Now 'emerge scipy' fails with:
71 > ...
72 > RefactoringTool:
73 > /var/tmp/portage/sci-libs/scipy-0.9.0-r1/work/scipy-0.9.0/build/py3k/scipy/sparse/linalg/isolve/setup.py
74 > RefactoringTool:
75 > /var/tmp/portage/sci-libs/scipy-0.9.0-r1/work/scipy-0.9.0/build/py3k/scipy/integrate/setup.py
76 > /var/tmp/portage/sci-libs/scipy-0.9.0-r1/work/scipy-0.9.0/build/py3k/scipy/io/__init__.py
77 > /var/tmp/portage/sci-libs/scipy-0.9.0-r1/work/scipy-0.9.0/build/py3k/scipy/io/matlab/mio4.py
78 > ...
79 > /var/tmp/portage/sci-libs/scipy-0.9.0-r1/work/scipy-0.9.0/build/py3k/scipy/integrate/ode.py
80 > /var/tmp/portage/sci-libs/scipy-0.9.0-r1/work/scipy-0.9.0/build/py3k/scipy/integrate/quadpack.py
81 > blas_opt_info:
82 > blas_mkl_info:
83 > libraries mkl,vml,guide not found in /usr/local/lib64
84 > libraries mkl,vml,guide not found in /usr/local/lib
85 > libraries mkl,vml,guide not found in /usr/lib64
86 > libraries mkl,vml,guide not found in /usr/lib
87 > NOT AVAILABLE
88 >
89 > atlas_blas_threads_info:
90 > Setting PTATLAS=ATLAS
91 > Warning: No configuration returned, assuming
92 > unavailable./usr/lib64/python3.2/site-packages/numpy/distutils/system_info.py:527: UserWarning: Specified path
93 > is
94 > invalid.
95 > warnings.warn('Specified path %s is invalid.' % d)
96 > libraries atlcblas,f77blas,m,atlas not found in /usr/lib64
97 > NOT AVAILABLE
98 > ...
99 > blas_src_info:
100 > NOT AVAILABLE
101 >
102 > /usr/lib64/python3.2/site-packages/numpy/distutils/system_info.py:1411:
103 > UserWarning:
104 > Blas (http://www.netlib.org/blas/) sources not found.
105 > Directories to search for the sources can be specified in the
106 > numpy/distutils/site.cfg file (section [blas_src]) or by setting
107 > the BLAS_SRC environment variable.
108 > warnings.warn(BlasSrcNotFoundError.__doc__)
109 > Traceback (most recent call last):
110 > File "setup.py", line 181, in <module>
111 > ...
112 > File
113 > "/usr/lib64/python3.2/site-packages/numpy/distutils/system_info.py",
114 > line 461, in get_info
115 > raise self.notfounderror(self.notfounderror.__doc__)
116 > numpy.distutils.system_info.BlasNotFoundError:
117 > Blas (http://www.netlib.org/blas/) libraries not found.
118 > Directories to search for the libraries can be specified in the
119 > numpy/distutils/site.cfg file (section [blas]) or by setting
120 > the BLAS environment variable.
121 > * ERROR: sci-libs/scipy-0.9.0-r1 failed (compile phase):
122 > * Building failed with CPython 3.2 in distutils_building() function
123 >
124 >
125 > USE_PYTHON="2.7" 'emerge -v1 scipy'
126 You replied while I was composing another answer :)
127 It looks like you are suffering from the bug mentioned by Steve. It may be
128 that using python-3.2 is a further element trigering it. This is interesting
129 info in and of itself that will need to be added to the bug.
130
131 The goal of Steve's suggestion was to avoid building scipy for python-3.2. It
132 obviously didn't work. So here are my instructions:
133 1) make sure numpy-1.5.1 is installed
134 2) if it is, get the scipy ebuild from
135 https://bugs.gentoo.org/show_bug.cgi?id=371099 and try it.
136
137 Francois