Can't install both opencv 3.1.0 and matplotlib (Windows) #27

hadim · 2016-08-30T13:36:26Z

I didn't check this bug on OS X and Linux. Only Windows.

$ conda info
Current conda install:

             platform : win-64
        conda version : 4.1.11
    conda-env version : 2.5.2
  conda-build version : 1.21.14+0.g4dfebe9.dirty
       python version : 3.5.2.final.0
     requests version : 2.11.0
     root environment : XXXX\Miniconda3  (writable)
  default environment : XXXX\Miniconda3
     envs directories : XXXX\Miniconda3\envs
        package cache : XXXX\Miniconda3\pkgs
         channel URLs : https://conda.anaconda.org/conda-forge/win-64/
                        https://conda.anaconda.org/conda-forge/noarch/
                        https://repo.continuum.io/pkgs/free/win-64/
                        https://repo.continuum.io/pkgs/free/noarch/
                        https://repo.continuum.io/pkgs/pro/win-64/
                        https://repo.continuum.io/pkgs/pro/noarch/
                        https://repo.continuum.io/pkgs/msys2/win-64/
                        https://repo.continuum.io/pkgs/msys2/noarch/
          config file : C:\Users\brouhardlab\.condarc
         offline mode : False
    is foreign system : False

$ conda create -n test python --yes
Fetching package metadata ...........
Solving package specifications: ..........

Package plan for installation in environment C:\Users\brouhardlab\Miniconda3\envs\test:

The following NEW packages will be INSTALLED:

    certifi:        2016.8.2-py35_0 conda-forge
    pip:            8.1.2-py35_0    conda-forge
    python:         3.5.2-2         conda-forge
    setuptools:     26.1.0-py35_0   conda-forge
    vs2015_runtime: 14.0.25420-0    conda-forge
    wheel:          0.29.0-py35_0   conda-forge
    wincertstore:   0.2-py35_0      conda-forge

Linking packages ...
[      COMPLETE      ]|##################################################| 100%
#
# To activate this environment, use:
# $ source activate test
#
# To deactivate this environment, use:
# $ source deactivate
#

$ source activate test
$ conda install opencv --yes
Fetching package metadata ...........
Solving package specifications: ..........

Package plan for installation in environment C:\Users\brouhardlab\Miniconda3\envs\test:

The following NEW packages will be INSTALLED:

    jpeg:    9b-vc14_0         conda-forge [vc14]
    libpng:  1.6.24-vc14_0     conda-forge [vc14]
    libtiff: 4.0.6-vc14_6      conda-forge [vc14]
    mkl:     11.3.3-1
    numpy:   1.11.1-py35_1
    opencv:  3.1.0-np111py35_0 conda-forge
    zlib:    1.2.8-vc14_3      conda-forge [vc14]

Linking packages ...
[      COMPLETE      ]|##################################################| 100%
$ conda install matplotlib --yes
Fetching package metadata ...........
Solving package specifications: ....

The following specifications were found to be in conflict:
  - matplotlib
Use "conda info <package>" to see the dependencies for each package.

Now it works if I uninstall opencv and install matplotlib.

$ conda remove opencv --yes
Fetching package metadata ...........
Solving package specifications: ..........

Package plan for package removal in environment XXXXX\Miniconda3\envs\test:

The following packages will be REMOVED:

    opencv: 3.1.0-np111py35_0 conda-forge

Unlinking packages ...
[      COMPLETE      ]|##################################################| 100%
$ conda install matplotlib --yes
Fetching package metadata ...........
Solving package specifications: ..........
        1 file(s) copied.

Package plan for installation in environment XXXXX\Miniconda3\envs\test:

The following NEW packages will be INSTALLED:

    ca-certificates: 2016.8.2-3        conda-forge
    cycler:          0.10.0-py35_0     conda-forge
    freetype:        2.6.3-vc14_1      conda-forge [vc14]
    matplotlib:      1.5.2-np111py35_5 conda-forge
    openssl:         1.0.2h-vc14_2     conda-forge [vc14]
    pyparsing:       2.1.7-py35_0      conda-forge
    pyqt:            4.11.4-py35_7
    python-dateutil: 2.5.3-py35_0      conda-forge
    pytz:            2016.6.1-py35_0   conda-forge
    qt:              4.8.7-vc14_9                  [vc14]
    sip:             4.18-py35_0       conda-forge
    six:             1.10.0-py35_0     conda-forge

The following packages will be DOWNGRADED due to dependency conflicts:

    jpeg:            9b-vc14_0         conda-forge [vc14] --> 8d-vc14_0    [vc14]
    libtiff:         4.0.6-vc14_6      conda-forge [vc14] --> 4.0.6-vc14_2 [vc14]

Unlinking packages ...
[      COMPLETE      ]|##################################################| 100%
Linking packages ...
[      COMPLETE      ]|##################################################| 100%

The text was updated successfully, but these errors were encountered:

hadim · 2016-08-30T13:41:56Z

I found the culprit. qt depends on :

dependencies:
    jpeg 8d
    libpng 1.6*
    libtiff 4.0.*
    openssl 1.0.2*
    zlib 1.2.*

jakirkham · 2016-08-30T14:06:30Z

So the qt package comes from defaults. Hence why it is not listed as coming from conda-forge above. We don't build qt anywhere, but Linux ATM. See issue ( conda-forge/conda-forge.github.io#169 ).

jakirkham · 2016-08-30T14:12:15Z

In any event, this is not an opencv issue. This is an org pinning issue that has been causing pain for some time unfortunately.

183amir · 2016-08-30T21:04:23Z

Opencv can have another branch that builds with jpeg 8 and would not break anybody's workflow.
I suggested this before but it was disregarded.

shelper · 2016-10-11T15:53:26Z

then can anyone add opencv with jpeg8 to conda-forge?

patricksnape · 2016-10-11T18:15:54Z

You are welcome to add a PR, unfortunately I don't have time to debug this.

183amir · 2016-10-11T21:16:09Z

then can anyone add opencv with jpeg8 to conda-forge?

This is possible but discouraged. The solution so far is to bring what you have compiled under jpeg8 here to conda-forge and compile it with jpeg9.
More similar discussions here: conda-forge/boost-feedstock#15 and conda-forge/jasper-feedstock#12

shelper · 2016-10-12T17:42:59Z

conda-forge has jpeg9 for 32bit, but not 64bit,
at least on my windows 7 64bit, when i tried conda install -c conda-forge jepg it only installs version 8d
see below:

jpeg 9b vc10_0

file name : conda-forge::jpeg-9b-vc10_0.tar.bz2
name : jpeg
version : 9b
build number: 0
build string: vc10_0
channel : conda-forge/win-64
size : 297 KB
arch : x86_64
binstar : {'package_id': '57113611752ed06038dfe854', 'owner_id': '5528f42ce1dad12
974506e8d', 'channel': 'main'}
features : vc10
fn : jpeg-9b-vc10_0.tar.bz2
license : Custom free software license
machine : x86_64
md5 : 16d13c0ff3f70aeb3985c41e90a9889b
operatingsystem: win32
platform : win
priority : 2
schannel : conda-forge
subdir : win-64
target-triplet: x86_64-any-win32
url : https://conda.anaconda.org/conda-forge/win-64/jpeg-9b-vc10_0.tar.bz2
dependencies:

jpeg 9b vc14_0

file name : conda-forge::jpeg-9b-vc14_0.tar.bz2
name : jpeg
version : 9b
build number: 0
build string: vc14_0
channel : conda-forge/win-64
size : 301 KB
arch : x86_64
binstar : {'package_id': '57113611752ed06038dfe854', 'owner_id': '5528f42ce1dad12
974506e8d', 'channel': 'main'}
features : vc14
fn : jpeg-9b-vc14_0.tar.bz2
license : Custom free software license
machine : x86_64
md5 : 84096b1f3fddf28e165117590fcad7b4
operatingsystem: win32
platform : win
priority : 2
schannel : conda-forge
subdir : win-64
target-triplet: x86_64-any-win32
url : https://conda.anaconda.org/conda-forge/win-64/jpeg-9b-vc14_0.tar.bz2
dependencies:

jpeg 9b vc9_0

file name : conda-forge::jpeg-9b-vc9_0.tar.bz2
name : jpeg
version : 9b
build number: 0
build string: vc9_0
channel : conda-forge/win-64
size : 294 KB
arch : x86_64
binstar : {'package_id': '57113611752ed06038dfe854', 'owner_id': '5528f42ce1dad12
974506e8d', 'channel': 'main'}
features : vc9
fn : jpeg-9b-vc9_0.tar.bz2
license : Custom free software license
machine : x86_64
md5 : b73350bd1d7e5381760ab5e9bf4edb58
operatingsystem: win32
platform : win
priority : 2
schannel : conda-forge
subdir : win-64
target-triplet: x86_64-any-win32
url : https://conda.anaconda.org/conda-forge/win-64/jpeg-9b-vc9_0.tar.bz2
dependencies:

jakirkham · 2016-10-12T17:52:38Z

If there are issues with the jpeg packages, please raise them at the jpeg issue tracker. Thanks.

hadim mentioned this issue Aug 30, 2016

Issue with jpeg dependency on Windows conda-forge/qt-feedstock#2

Closed

jakirkham closed this as completed Aug 30, 2016

nzjrs mentioned this issue Apr 12, 2018

imshow segfaults loopbio/opencv-feedstock#13

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Can't install both opencv 3.1.0 and matplotlib (Windows) #27

Can't install both opencv 3.1.0 and matplotlib (Windows) #27

hadim commented Aug 30, 2016

hadim commented Aug 30, 2016

jakirkham commented Aug 30, 2016

jakirkham commented Aug 30, 2016

183amir commented Aug 30, 2016

shelper commented Oct 11, 2016

patricksnape commented Oct 11, 2016 •

edited

Loading

183amir commented Oct 11, 2016

shelper commented Oct 12, 2016 •

edited

Loading

jpeg 9b vc10_0

jpeg 9b vc14_0

jpeg 9b vc9_0

jakirkham commented Oct 12, 2016

Can't install both opencv 3.1.0 and matplotlib (Windows) #27

Can't install both opencv 3.1.0 and matplotlib (Windows) #27

Comments

hadim commented Aug 30, 2016

hadim commented Aug 30, 2016

jakirkham commented Aug 30, 2016

jakirkham commented Aug 30, 2016

183amir commented Aug 30, 2016

shelper commented Oct 11, 2016

patricksnape commented Oct 11, 2016 • edited Loading

183amir commented Oct 11, 2016

shelper commented Oct 12, 2016 • edited Loading

jpeg 9b vc10_0

jpeg 9b vc14_0

jpeg 9b vc9_0

jakirkham commented Oct 12, 2016

patricksnape commented Oct 11, 2016 •

edited

Loading

shelper commented Oct 12, 2016 •

edited

Loading