continuumio / anaconda-recipes Goto Github PK

View Code? Open in Web Editor NEW

120.0 120.0 69.0 2.49 MB

Continuum managed recipes for free anaconda packages.

License: Other

Batchfile 14.79% Shell 12.41% Python 12.28% C++ 7.53% CMake 0.27% C 52.70% QMake 0.02%

anaconda-recipes's People

Contributors

Stargazers

Watchers

Forkers

gillins msarahan jakirkham pmuller mdboom anukat2015 jhunkeler astaric libardo1 data-exp-lab hrk2109 djhoese rlugojr lexual tomkooij oleksandr-pavlyk parejkoj digideskio astrofrog volumie ales-erjavec sjoerdapp arielsmt rainwoodman chromium-24 183amir tkelman zhufenggnss bsipocz honnibal stackless-dev nehaljwani qwhelan barkinet mbargull mandeep nikicc markotoplak joaopaulopacheco ifarhankhan shizzz477 kzwkt panickervinod ahdezm absarf xmnlab menshikh-iv sandychapman jjhelmus www3838438 enterstudio mfrigillana professorzhang global-localhost global19 global19-atlassian-net ruchirasachinthana python-repository-hub woodpile37 specialized806 tpsaint

anaconda-recipes's Issues

pytables missing package on win32 for py35 np19

Would be very glad if you could provide a build, thank you!

Add missing components to `ipython-notebook` 4.x package

The following were components in the original iPython 3.x along with their post split counterparts.¹

Checked component appear in the ipython-notebook 4.x package. Unchecked ones are missing when installing ipython-notebook. It would be nice if the missing counterparts were included in the ipython-notebook package to make this transition a bit more seamless when making the transition.

[1] - http://blog.jupyter.org/2015/04/15/the-big-split/

- Occurred after the blog post was published. Also, will be split further.

What is replace.exe?

A few recipes make use of Scripts\replace.exe for windows builds, which doesn't seem to be from anaconda. Is there a package to install that provides this?

Update ipyparallel to 4.1.0

libevent missing Python dependency

Appears the libevent recipe is not open sourced AFAICT or at least it is not in this repo. In any event, I'm going to put this issue here as I can't think of a more appropriate place to put it ATM. Should add libevent does appear in the anaconda channel. So this seems reasonable.

After recently working on a libevent recipe in PR ( conda-forge/staged-recipes#1491 ), it has become clear to me that libevent depends on Python. This is because it installs a script to the bin directory called event_rpcgen.py. Not only does it depend on Python, but that script appears to hard-code Python 2. Thus libevent should depend on Python 2.x for run and build.

Qt 5.6.0 broken because of permission errors in qt.conf (Fixed)

Hi,
After updating an existing Anaconda distribution that was installed your with shiny 400M installer, Qt was bumped to 5.6.0:

$ ./conda list qt
# packages in environment at /opt/anaconda-testing:
#
pyqt                      5.6.0                    py35_0  
qt                        5.6.0                         0

But there is a problem: when this environment is installed in /opt, some permissions are just brainless. Indeed, after this update, anaconda3/bin/qt.conf (it is just a directory example) is set to 750 permissions:

$ stat qt.conf
  File: 'qt.conf'
  Size: 157             Blocks: 8          IO Block: 4096   regular file
Device: 801h/2049d      Inode: 2816378     Links: 1
Access: (0750/-rwxr-x---)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2016-10-06 15:09:38.961687000 +0200
Modify: 2016-10-06 15:10:22.752769805 +0200
Change: 2016-10-06 15:10:22.752769805 +0200
 Birth: -

With Qt 4.8.7, it was 664 which is much smarter.

Consequences:
You just broke 50 machines in my infrastructure and my scientists are now unable to use other Python tools (like pyzo). In fact, the Python shell fails to load modules (xcb for example) because the paths are not defined. Yes, regular users cannot read that bloody file because they are in the other group! It took 1,5 day for me to debug it, thanks to strace(1).

Fixes
Maybe you should learn UNIX permissions / rights. This is extremely unprofessional for a company! This is why I will not recommend your twisted snake env. Another new reason why I hate this kind of "packager manager", where the maintainers are unable to correctly maintain packages. Or perhaps, I should not update packages to make sure everything works…

Oh BTW, rollbacking doesn't work because a symlink already exists. I had to remove it by hand. I will not open another issue because I'm pretty sure we will find another bunch of flaws.

Recipe license

Could the license or terms of use of these recipes be specified? Are in the public domain similar to the recipes in https://github.com/conda/conda-recipes or does another license apply?

mkl-devel recipe?

Is there any chance you can share the recipe you use for mkl-devel? I have other packages that depend on mkl that I want to build and I'd like to make sure that they play along nicely with the Anaconda mkl runtime package.

I am guessing that there are licensing issues with distributing mkl-devel because gaining access to the source code and linking against it requires acceptance of the Intel license. However, having knowledge of how Continuum builds their mkl shouldn't violate that if I go get the source from Intel myself... maybe?

Matplotlib Python 3.6 builds missing

The Matplotlib builds for Python 3.6 appear to be missing from the defaults channel. Are there currently issues building this?

Shapely depends on numpy ABI

Shapely uses Cython to speed up some operations, and therefore depends on the numpy ABI. Please see https://github.com/conda-forge/shapely-feedstock for a (longstanding) recipe that correctly builds Shapely.

conda-build-all 1.0.2

Noticed that conda-build-all 1.0.1 was added to defaults and wanted to suggest that 1.0.2 be added as well. It fixed a bug with conda-build 2.x compatibility and some issues with R support. Would be nice if it were available given conda-build-all is already in defaults.

scipy test fails when mkl-rt is installed

If I build a conda environment with scipy and nose, and then do

python -c "import scipy; scipy.test('fast')"

it works. But if I add mkl-rt to the environment, then I get a failure, see below.

At some point I was running into isses with the new mkl numpy/scipy, and googling around led me to think that installing mkl-rt was a good idea, but I did get messages that implied it had some of the same files as the mkl package, which made me nervous. What is the difference? Why would someone want mkl-rt? It is hard to remove that package also, if you remove mkl-rt, then I think you remove some files that are also in mkl, as certain parts of scipy won't import anymore. I don't see a good way to 're-install' mkl to get it back.

Anyways, maybe nobody is using mkl-rt, but thought I would share: here is the failure

======================================================================
FAIL: test_iterative.test_convergence(<function bicgstab at 0x7f0b8409a500>, <nonsymposdef>)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/reg/neh/home/davidsch/.conda/envs/scipy/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/reg/neh/home/davidsch/.conda/envs/scipy/lib/python2.7/site-packages/scipy/sparse/linalg/isolve/tests/test_iterative.py", line 196, in check_convergence
    assert_equal(info,0)
  File "/reg/neh/home/davidsch/.conda/envs/scipy/lib/python2.7/site-packages/numpy/testing/utils.py", line 379, in assert_equal
    raise AssertionError(msg)
AssertionError: 
Items are not equal:
 ACTUAL: -10
 DESIRED: 0

My conda environment is

(scipy) psanaphi103: ~ $ conda list
# packages in environment at /reg/neh/home/davidsch/.conda/envs/scipy:
#
libgfortran               3.0.0                         1    defaults
mkl                       11.3.3                        0    defaults
mkl-rt                    11.1                         p0    defaults
nose                      1.3.7                    py27_1    defaults
numpy                     1.11.1                   py27_0    defaults
openssl                   1.0.2h                        1    defaults
pip                       8.1.2                    py27_0    defaults
python                    2.7.12                        1    defaults
readline                  6.2                           2    defaults
scipy                     0.18.0              np111py27_0    defaults
setuptools                27.2.0                   py27_0    defaults
sqlite                    3.13.0                        0    defaults
tk                        8.5.18                        0    defaults
wheel                     0.29.0                   py27_0    defaults
zlib                      1.2.8                         3    defaults

below is output from conda info.

(scipy) psanaphi103: ~ $ conda info
Current conda install:

               platform : linux-64
          conda version : 4.2.7
       conda is private : False
      conda-env version : 4.2.7
    conda-build version : 2.0.1
         python version : 2.7.12.final.0
       requests version : 2.10.0
       root environment : /reg/g/psdm/sw/conda/inst/miniconda2-dev-rhel7  (read only)
    default environment : /reg/neh/home/davidsch/.conda/envs/scipy
       envs directories : /reg/neh/home/davidsch/.conda/envs
                          /reg/g/psdm/sw/conda/inst/miniconda2-dev-rhel7/envs
          package cache : /reg/neh/home/davidsch/.conda/envs/.pkgs
                          /reg/g/psdm/sw/conda/inst/miniconda2-dev-rhel7/pkgs
           channel URLs : file:///reg/g/psdm/sw/conda/channels/system-rhel7/linux-64/
                          file:///reg/g/psdm/sw/conda/channels/system-rhel7/noarch/
                          file:///reg/g/psdm/sw/conda/channels/psana-rhel7/linux-64/
                          file:///reg/g/psdm/sw/conda/channels/psana-rhel7/noarch/
                          file:///reg/g/psdm/sw/conda/channels/external-rhel7/linux-64/
                          file:///reg/g/psdm/sw/conda/channels/external-rhel7/noarch/
                          https://repo.continuum.io/pkgs/free/linux-64/
                          https://repo.continuum.io/pkgs/free/noarch/
                          https://repo.continuum.io/pkgs/pro/linux-64/
                          https://repo.continuum.io/pkgs/pro/noarch/
                          https://conda.anaconda.org/conda-forge/linux-64/
                          https://conda.anaconda.org/conda-forge/noarch/
                          https://conda.anaconda.org/scikit-beam/linux-64/
                          https://conda.anaconda.org/scikit-beam/noarch/
                          file:///reg/g/psdm/sw/conda/channels/testing-rhel7/linux-64/
                          file:///reg/g/psdm/sw/conda/channels/testing-rhel7/noarch/
            config file : /reg/g/psdm/sw/conda/inst/miniconda2-dev-rhel7/.condarc
           offline mode : False

if you google that error, you'll see references to issues with using optimization flags when building scipy: https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/505993

qt 4.8.7 build 3 renamed moc to moc-qt4, breaks software

I just built an environment that depends on qt4. It picked up qt 4.8.7_3 from defaults. This build seems to rename the bin/moc file to bin/moc-qt4. This brakes software that depends on the qt package that used the moc program. My previous environment used qt 4.8.7_1, build 1, where the executable was just called moc - this seems like a big change for a build number. You can see this change from the file list in conda-meta, and trying to run moc in the respective environments.

We are excited about conda, but I run into problems because conda changes from version to version, and a package in defaults changes. I need to isolate our production installation from these changes until I can test them more - so this is a simple question, how can I programmatically query the package list in defaults, and download what I want - including versions of conda, conda-build, conda-env. Specifically what is the URL I would use, I'm thinking I will set up my own file based channel of the packages from defaults that I want to use in production - I could just download the packages from defaults in a development environment, and move them, but in this scenario, I would look for the downloaded packages in the environment pkgs area, and trust that they are the same thing I got from defaults - which they won't be if I download the same package name from a different channel before I copy

Can you please add the recipes you have for numpy and scipy?

I didn't know this repo existed. Thanks for creating it!

Recipe for more modern GCC

I would like to create a conda Bioconda package that relies on a more modern version of GCC.

The software I want to package is Flexbar which relies on the SeqAn library >=2.2.0 which in turn relies on GCC >=4.9.2 for C++14 features. This problem also exists for other SeqAn apps such as Yara and RazerS 3 which are currently packaged by downloading pre-built binaries by the SeqAn authors.

Without an updated GCC (or alternatively clang++), this software cannot be compiled. This will become more and more of a problem as more software relies on modern C++ features.

I asked in the Bioconda issue tracker (bioconda/bioconda-recipes#3224) and also @johanneskoester directly and the recommendation was to ask here for input on this (and also to ping @msarahan).

Regarding on my pull request conda-forge/staged-recipes#2089 in conda-forge, Johannes told me that versions should not be part of the package name and suggested this problem should be discussed here.

Please advise.

Related seqan/seqan issue: seqan/seqan#1547

pillow 4.0.0 required by anaconda 4.3.0 misses a dependence to `olefile`

Pillow 4.0.0 now requires a dependence on a Python package called olefile (just added to this recipe container). If you pip install something on a conda environment with anaconda 4.3.0 pre-installed that requires pillow, then it will try to download olefile from PyPI. Ideally, you should just depend on it.

ping @183amir

Update-request: line-profiler

I hope it's okay to simply ask here: Is it possible to update the line_profiler package?

There has been a 2.0 release which fixed some ipython 5.x compatibility issues.

Clarify how this repository stands against conda/conda-recipes

Hi, just noticed this repo. It would be nice if there was some clarification in the README.md that explains:

Are these recipes also the recipes maintained internally by ContinuumIO?
The relationship between this repo, conda/conda-recipes and conda-forge? (e.g. why a third one?)
Is one of the aforementioned repos considered "leading"?

libxslt 1.1.29 introduces system dependencies which should be packaged (Linux)

libxslt was updated to 1.1.29 and it is now linking with additional system libraries which are not widely available across distributions:

$ conda create -n test libxslt
$ source activate test
(test) $ ldd $CONDA_PREFIX/lib/libexslt.so
	linux-vdso.so.1 =>  (0x00007ffde9957000)
	libxslt.so.1 => /home/guilherme/Work/miniconda/envs/test/lib/./libxslt.so.1 (0x00007f2ee49b8000)
	libxml2.so.2 => /home/guilherme/Work/miniconda/envs/test/lib/./libxml2.so.2 (0x00007f2ee465d000)
	libz.so.1 => /home/guilherme/Work/miniconda/envs/test/lib/./libz.so.1 (0x00007f2ee4447000)
	libiconv.so.2 => /home/guilherme/Work/miniconda/envs/test/lib/./libiconv.so.2 (0x00007f2ee4158000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f2ee3e1f000)
	librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f2ee3c15000)
	libgcrypt.so.11 => not found
	libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f2ee3a11000)
	libgpg-error.so.0 => /lib/x86_64-linux-gnu/libgpg-error.so.0 (0x00007f2ee37fd000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f2ee3436000)
	/lib64/ld-linux-x86-64.so.2 (0x000055ab7009d000)
	libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f2ee3218000)

I don't know about libgpg-error.so.0, but libgcrypt.so.11 should definitely not come from system as its API/ABI seems to constantly change (currently Ubuntu 16.10 ships libgcrypt.so.20).

PyQt 5 crashes on Linux

When trying to use PyQt 5 on Linux in a stock miniconda installation it crashed with a seg fault. Isolating the problem, it happens when executing QGuiApplication::font method on Qt code.

Steps to reproduce

~/miniconda2/bin/conda create -n test_qt python=3.5 qt=5 pyqt
source ~/miniconda2/bin/activate test_qt
python -c "from PyQt5.QtWidgets import QApplication; print(QApplication([])); from PyQt5.QtGui import QFont; print(QFont())"

This outputs:

<PyQt5.QtWidgets.QApplication object at 0x7f188c691808>
Segmentation fault (core dumped)

gdb backtrace

#0  0x00007ffff4e0b5b9 in ?? () from /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/PyQt5/../../../libQt5Gui.so.5
#1  0x00007ffff4e0c56d in QGuiApplication::font() () from /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/PyQt5/../../../libQt5Gui.so.5
#2  0x00007ffff4eac5de in QFont::QFont() () from /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/PyQt5/../../../libQt5Gui.so.5
#3  0x00007fffedf7617d in init_type_QFont () from /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/PyQt5/QtGui.so
#4  0x00007fffee954e98 in ?? () from /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/sip.so
#5  0x00007ffff79b6daf in type_call (type=<optimized out>, args=0x7ffff7f81048, kwds=0x0) at Objects/typeobject.c:905
#6  0x00007ffff794f236 in PyObject_Call (func=0x7bddc8, arg=<optimized out>, kw=<optimized out>) at Objects/abstract.c:2165
#7  0x00007ffff7a29314 in do_call (nk=<optimized out>, na=0, pp_stack=0x7fffffffd448, func=0x7bddc8) at Python/ceval.c:4936
#8  call_function (oparg=<optimized out>, pp_stack=0x7fffffffd448) at Python/ceval.c:4732
#9  PyEval_EvalFrameEx (f=<optimized out>, throwflag=<optimized out>) at Python/ceval.c:3236
#10 0x00007ffff7a2cb49 in _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=<optimized out>, args=<optimized out>, argcount=0, kws=0x0, kwcount=0, defs=0x0, defcount=0, kwdefs=0x0, closure=0x0, name=0x0, qualname=0x0)
    at Python/ceval.c:4018
#11 0x00007ffff7a2ccd8 in PyEval_EvalCodeEx (_co=<optimized out>, globals=<optimized out>, locals=<optimized out>, args=<optimized out>, argcount=<optimized out>, kws=<optimized out>, kwcount=0, defs=0x0, defcount=0, kwdefs=0x0, closure=0x0) at Python/ceval.c:4039
#12 0x00007ffff7a2cd1b in PyEval_EvalCode (co=<optimized out>, globals=<optimized out>, locals=<optimized out>) at Python/ceval.c:777
#13 0x00007ffff7a5487a in run_mod (arena=0x69bcc0, flags=0x7fffffffd710, locals=0x7ffff7f2f408, globals=0x7ffff7f2f408, filename=<optimized out>, mod=<optimized out>) at Python/pythonrun.c:976
#14 PyRun_StringFlags (str=0x7ffff7e62050 "from PyQt5.QtWidgets import QApplication; print(QApplication([])); from PyQt5.QtGui import QFont; print(QFont())\n", start=257, globals=0x7ffff7f2f408, locals=0x7ffff7f2f408, flags=0x7fffffffd710) at Python/pythonrun.c:900
#15 0x00007ffff7a54910 in PyRun_SimpleStringFlags (command=0x7ffff7e62050 "from PyQt5.QtWidgets import QApplication; print(QApplication([])); from PyQt5.QtGui import QFont; print(QFont())\n", flags=0x7fffffffd710) at Python/pythonrun.c:421
#16 0x00007ffff7a6e16e in run_command (cf=0x7fffffffd710, command=0x604460 L"from PyQt5.QtWidgets import QApplication; print(QApplication([])); from PyQt5.QtGui import QFont; print(QFont())\n") at Modules/main.c:279
#17 Py_Main (argc=<optimized out>, argv=<optimized out>) at Modules/main.c:707
#18 0x0000000000400add in main (argc=3, argv=0x7fffffffd888) at ./Programs/python.c:65

valgrind output

It shows a lot of things. This is what seems relevant:

==24070== Invalid read of size 8
==24070==    at 0x7BAF5B9: ??? (in /home/guilherme/miniconda2/envs/test_qt/lib/libQt5Gui.so.5.6.2)
==24070==    by 0x7BB056C: QGuiApplication::font() (in /home/guilherme/miniconda2/envs/test_qt/lib/libQt5Gui.so.5.6.2)
==24070==    by 0x7C505DD: QFont::QFont() (in /home/guilherme/miniconda2/envs/test_qt/lib/libQt5Gui.so.5.6.2)
==24070==    by 0xF09D17C: init_type_QFont (in /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/PyQt5/QtGui.so)
==24070==    by 0xE47DE97: ??? (in /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/sip.so)
==24070==    by 0x4F07DAE: type_call (typeobject.c:905)
==24070==    by 0x4EA0235: PyObject_Call (abstract.c:2165)
==24070==    by 0x4F7A313: do_call (ceval.c:4936)
==24070==    by 0x4F7A313: call_function (ceval.c:4732)
==24070==    by 0x4F7A313: PyEval_EvalFrameEx (ceval.c:3236)
==24070==    by 0x4F7DB48: _PyEval_EvalCodeWithName (ceval.c:4018)
==24070==    by 0x4F7DCD7: PyEval_EvalCodeEx (ceval.c:4039)
==24070==    by 0x4F7DD1A: PyEval_EvalCode (ceval.c:777)
==24070==    by 0x4FA5879: run_mod (pythonrun.c:976)
==24070==    by 0x4FA5879: PyRun_StringFlags (pythonrun.c:900)
==24070==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
==24070== 
==24070== 
==24070== Process terminating with default action of signal 11 (SIGSEGV)
==24070==  Access not within mapped region at address 0x0
==24070==    at 0x7BAF5B9: ??? (in /home/guilherme/miniconda2/envs/test_qt/lib/libQt5Gui.so.5.6.2)
==24070==    by 0x7BB056C: QGuiApplication::font() (in /home/guilherme/miniconda2/envs/test_qt/lib/libQt5Gui.so.5.6.2)
==24070==    by 0x7C505DD: QFont::QFont() (in /home/guilherme/miniconda2/envs/test_qt/lib/libQt5Gui.so.5.6.2)
==24070==    by 0xF09D17C: init_type_QFont (in /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/PyQt5/QtGui.so)
==24070==    by 0xE47DE97: ??? (in /home/guilherme/miniconda2/envs/test_qt/lib/python3.5/site-packages/sip.so)
==24070==    by 0x4F07DAE: type_call (typeobject.c:905)
==24070==    by 0x4EA0235: PyObject_Call (abstract.c:2165)
==24070==    by 0x4F7A313: do_call (ceval.c:4936)
==24070==    by 0x4F7A313: call_function (ceval.c:4732)
==24070==    by 0x4F7A313: PyEval_EvalFrameEx (ceval.c:3236)
==24070==    by 0x4F7DB48: _PyEval_EvalCodeWithName (ceval.c:4018)
==24070==    by 0x4F7DCD7: PyEval_EvalCodeEx (ceval.c:4039)
==24070==    by 0x4F7DD1A: PyEval_EvalCode (ceval.c:777)
==24070==    by 0x4FA5879: run_mod (pythonrun.c:976)
==24070==    by 0x4FA5879: PyRun_StringFlags (pythonrun.c:900)

Installed packages

$ conda list --show-channel-urls
# packages in environment at /home/guilherme/miniconda2/envs/test_qt:
#
dbus                      1.10.10                       0    defaults
expat                     2.1.0                         0    defaults
fontconfig                2.12.1                        2    defaults
freetype                  2.5.5                         2    defaults
glib                      2.50.2                        1    defaults
gst-plugins-base          1.8.0                         0    defaults
gstreamer                 1.8.0                         0    defaults
icu                       54.1                          0    defaults
jpeg                      9b                            0    defaults
libffi                    3.2.1                         1    defaults
libgcc                    5.2.0                         0    defaults
libiconv                  1.14                          0    defaults
libpng                    1.6.27                        0    defaults
libxcb                    1.12                          1    defaults
libxml2                   2.9.4                         0    defaults
openssl                   1.0.2k                        0    defaults
pcre                      8.39                          1    defaults
pip                       9.0.1                    py35_1    defaults
pyqt                      5.6.0                    py35_2    defaults
python                    3.5.2                         0    defaults
qt                        5.6.2                         3    defaults
readline                  6.2                           2    defaults
setuptools                27.2.0                   py35_0    defaults
sip                       4.18                     py35_0    defaults
sqlite                    3.13.0                        0    defaults
tk                        8.5.18                        0    defaults
wheel                     0.29.0                   py35_0    defaults
xz                        5.2.2                         1    defaults
zlib                      1.2.8                         3    defaults

conda info

Current conda install:

               platform : linux-64
          conda version : 4.2.13
       conda is private : False
      conda-env version : 4.2.13
    conda-build version : 2.0.12
         python version : 2.7.12.final.0
       requests version : 2.12.1
       root environment : /home/guilherme/miniconda2  (writable)
    default environment : /home/guilherme/miniconda2/envs/test_qt
       envs directories : /home/guilherme/miniconda2/envs
          package cache : /home/guilherme/miniconda2/pkgs
           channel URLs : https://repo.continuum.io/pkgs/free/linux-64
                          https://repo.continuum.io/pkgs/free/noarch
                          https://repo.continuum.io/pkgs/pro/linux-64
                          https://repo.continuum.io/pkgs/pro/noarch
                          https://conda.anaconda.org/conda-forge/linux-64
                          https://conda.anaconda.org/conda-forge/noarch
            config file : /home/guilherme/miniconda2/.condarc
           offline mode : False

I tried to reproduce in C++:

#include <QtGui/QFont>
#include <QtWidgets/QApplication>

int main(int argc, char *argv[]) {
    QGuiApplication qapp(argc, argv);
    qapp.font();
    QFont font;
    printf("Successfully created\n");
}

and compile and run with:

g++ -o qt_example qt_example.cpp -I$CONDA_PREFIX/include -I$CONDA_PREFIX/include/qt -L $CONDA_PREFIX/lib -lQt5Widgets -lQt5Gui -lQt5Core -licui18n -licudata -licuuc
LD_LIBRARY_PATH=$CONDA_PREFIX/lib ./qt_example

but it works just fine.

All info above was obtained on an Ubuntu 16.10 and conda packages Python 3.5, Qt 5.6.2, PyQt 5.6.0 but I've also tried some other configuration:

python 2.7, 3.5 and 3.6
qt 5.6.0 and 5.6.2 (all builds)
pyqt 5.6.0 (all builds)
Ubuntu 16.04 VM, CentOS 6 chroot

Anaconda builds python-2.7.11-5 and later drop bin/python2.7 as not linked, resulting in incorrect sys.prefix

The current python recipe specifies a

build:
nolink:
bin/python

which should probably refer to bin/python2.7 instead.
bin/python is itself a symbolic link to python2.7.

Because python2.7 is linked to the package directory on install

python -> python2.7
python2 -> python2.7
python2.7 -> /home/username/.conda/envs/.pkgs/python-2.7.11-5/bin/python2.7

python itself thinks it is running out of the package directory
and sets sys.prefix accordingly.

If python2.7 is installed as a copy,
then python sets sys.prefix to the conda environment path.

It should look like this:

python -> python2.7
python2 -> python2.7
python2.7 (a copy)

Versions 2.7.11 build 0 and before contain bin/python2.7 in the file that excludes links.

python-2.7.11-0/info$ cat no_softlink
bin/python
bin/python-config
bin/python2
bin/python2.7

-tim

NumPy windows build is broken

The numpy.distutils.system_info.get_info function is broken on the windows build of NumPy. I need this function to get the location of BLAS and LAPACK installations. But when I call get_info with the Anaconda NumPy build I just get an error:

>>> from numpy.distutils.system_info import get_info
>>> get_info('blas_opt')
C:\Users\23190281\Miniconda2\lib\site-packages\numpy\distutils\system_info.py:63
9: UserWarning: Specified path C:\Minonda\envs\_build\Library\lib is invalid.
  warnings.warn('Specified path %s is invalid.' % d)
C:\Users\23190281\Miniconda2\lib\site-packages\numpy\distutils\system_info.py:63
9: UserWarning: Specified path C:\Minonda\envs\_build\Library\include is invalid
.
  warnings.warn('Specified path %s is invalid.' % d)
C:\Users\23190281\Miniconda2\lib\site-packages\numpy\distutils\system_info.py:16
46: UserWarning:
    Atlas (http://math-atlas.sourceforge.net/) libraries not found.
    Directories to search for the libraries can be specified in the
    numpy/distutils/site.cfg file (section [atlas]) or by setting
    the ATLAS environment variable.
  warnings.warn(AtlasNotFoundError.__doc__)
C:\Users\23190281\Miniconda2\lib\site-packages\numpy\distutils\system_info.py:16
55: UserWarning:
    Blas (http://www.netlib.org/blas/) libraries not found.
    Directories to search for the libraries can be specified in the
    numpy/distutils/site.cfg file (section [blas]) or by setting
    the BLAS environment variable.
  warnings.warn(BlasNotFoundError.__doc__)
C:\Users\23190281\Miniconda2\lib\site-packages\numpy\distutils\system_info.py:16
58: UserWarning:
    Blas (http://www.netlib.org/blas/) sources not found.
    Directories to search for the sources can be specified in the
    numpy/distutils/site.cfg file (section [blas_src]) or by setting
    the BLAS_SRC environment variable.
  warnings.warn(BlasSrcNotFoundError.__doc__)

For the full details, see this issue: cvxgrp/scs#72

h5py windows requirement for hdf5 (1.8.15.1)

The "relativly" recent change in the h5py recipe (restricted to hdf5 1.8.15.1 on windows) leads to alternating updates and downgrades between h5py 2.6 and hdf5 1.8.16.0 and h5py 2.7 and hdf 1.8.15.1 on my computer:

The following packages will be UPDATED:

    bottleneck:   1.2.0-np112py35_0  --> 1.2.0-np111py35_0
    hdf5:         1.8.15.1-vc14_4    [vc14] --> 1.8.16-vc14_0      [vc14]
    matplotlib:   2.0.0-np112py35_0  --> 2.0.0-np111py35_0
    numba:        0.31.0-np112py35_0 --> 0.31.0-np111py35_0
    numexpr:      2.6.2-np112py35_0  --> 2.6.2-np111py35_0
    pandas:       0.19.2-np112py35_1 --> 0.19.2-np111py35_1
    pywavelets:   0.5.2-np112py35_0  --> 0.5.2-np111py35_0
    scikit-image: 0.13.0-np112py35_0 --> 0.13.0-np111py35_0
    scipy:        0.19.0-np112py35_0 --> 0.19.0-np111py35_0
    statsmodels:  0.8.0-np112py35_0  --> 0.8.0-np111py35_0

The following packages will be DOWNGRADED due to dependency conflicts:

    h5py:         2.7.0-np112py35_0  --> 2.6.0-np111py35_1
    numpy:        1.12.1-py35_0      --> 1.11.3-py35_0

and next time:

The following packages will be UPDATED:

    bottleneck:   1.2.0-np111py35_0  --> 1.2.0-np112py35_0
    h5py:         2.6.0-np111py35_1  --> 2.7.0-np112py35_0
    matplotlib:   2.0.0-np111py35_0  --> 2.0.0-np112py35_0
    numba:        0.31.0-np111py35_0 --> 0.31.0-np112py35_0
    numexpr:      2.6.2-np111py35_0  --> 2.6.2-np112py35_0
    numpy:        1.11.3-py35_0      --> 1.12.1-py35_0
    pandas:       0.19.2-np111py35_1 --> 0.19.2-np112py35_1
    pywavelets:   0.5.2-np111py35_0  --> 0.5.2-np112py35_0
    scikit-image: 0.13.0-np111py35_0 --> 0.13.0-np112py35_0
    scipy:        0.19.0-np111py35_0 --> 0.19.0-np112py35_0
    statsmodels:  0.8.0-np111py35_0  --> 0.8.0-np112py35_0

The following packages will be DOWNGRADED due to dependency conflicts:

    hdf5:         1.8.16-vc14_0      [vc14] --> 1.8.15.1-vc14_4    [vc14]

An old version of `libtiff` needs a hotfix

I am creating an environment like this:

conda create -n test1 bob.io.image=2.0.6=py34_jpeg9_libpng1.6.21_5
Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata .........
Solving package specifications: ..........

Package plan for installation in environment /home/amir/miniconda/envs/test1:

The following NEW packages will be INSTALLED:

    bob.blitz:     2.0.8-np111py34_7               conda-forge
    bob.core:      2.1.2-py34_5                    conda-forge
    bob.extension: 2.2.0-py34_1                    conda-forge
    bob.io.base:   2.0.8-py34_3                    conda-forge
    bob.io.image:  2.0.6-py34_jpeg9_libpng1.6.21_5 conda-forge
    boost:         1.60.0-py34_0                   defaults   
    giflib:        5.1.2-1                         conda-forge
    hdf5:          1.8.15.1-2                      defaults   
    icu:           54.1-0                          defaults   
    jpeg:          9b-0                            conda-forge
    libblitz:      0.10-0                          conda-forge
    libpng:        1.6.22-0                        defaults   
    libtiff:       4.0.2-0                         defaults   
    mkl:           11.3.3-0                        defaults   
    numpy:         1.11.1-py34_0                   defaults   
    openssl:       1.0.2h-1                        defaults   
    pip:           8.1.2-py34_0                    defaults   
    python:        3.4.5-0                         defaults   
    readline:      6.2-2                           defaults   
    setuptools:    23.0.0-py34_0                   defaults   
    sqlite:        3.13.0-0                        defaults   
    tk:            8.5.18-0                        defaults   
    wheel:         0.29.0-py34_0                   defaults   
    xz:            5.2.2-0                         defaults   
    zlib:          1.2.8-3                         defaults

but the libtiff package that is downloaded is not compatible with jpeg 9b:

ldd -d ~/miniconda/envs/test1/lib/libtiff.so
    linux-vdso.so.1 (0x00007ffebbdbe000)
    libjpeg.so.62 => not found
    libz.so.1 => /home/amir/miniconda/envs/test1/lib/./libz.so.1 (0x00007fc6789b9000)
    libm.so.6 => /usr/lib/libm.so.6 (0x00007fc6786b5000)
    libc.so.6 => /usr/lib/libc.so.6 (0x00007fc678313000)
    /usr/lib64/ld-linux-x86-64.so.2 (0x000055fc4fd42000)
undefined symbol: jpeg_resync_to_restart    (/home/amir/miniconda/envs/test1/lib/libtiff.so)

This is causing my build to fail here: conda-forge/staged-recipes#751
As you can see libtiff from the defaults channel is not properly pinned on a jpeg version.

conda info libtiff
libtiff 4.0.2 0
---------------
file name   : libtiff-4.0.2-0.tar.bz2
name        : libtiff
version     : 4.0.2
build number: 0
build string: 0
channel     : defaults
size        : 1001 KB
date        : 2013-02-09
fn          : libtiff-4.0.2-0.tar.bz2
license     : BSD-like
license_family: BSD
md5         : 6e61cca80f0ccb1ca487e617de42679a
priority    : 1
schannel    : defaults
url         : https://repo.continuum.io/pkgs/free/linux-64/libtiff-4.0.2-0.tar.bz2
dependencies:

qt 4.8.7_4 breaks pyqt 4.11.4 py34_4

I have a package that depends on pyqt. After installing it in a fresh python 3.4 environment on my mac (and some coworkers) I get the following error when trying to run my application:

Traceback (most recent call last):
  File "/Users/davidh/anaconda/envs/sift_production/lib/python3.4/site-packages/cspov/__main__.py", line 26, in <module>
    app_object = app.use_app('pyqt4')
  File "/Users/davidh/anaconda/envs/sift_production/lib/python3.4/site-packages/vispy/app/_default_app.py", line 47, in use_app
    default_app = Application(backend_name)
  File "/Users/davidh/anaconda/envs/sift_production/lib/python3.4/site-packages/vispy/app/application.py", line 49, in __init__
    self._use(backend_name)
  File "/Users/davidh/anaconda/envs/sift_production/lib/python3.4/site-packages/vispy/app/application.py", line 218, in _use
    raise RuntimeError(msg)
RuntimeError: Could not import backend "PyQt4":
dlopen(/Users/davidh/anaconda/envs/sift_production/lib/python3.4/site-packages/PyQt4/QtOpenGL.so, 2): Library not loaded: @rpath/libpng16.16.dylib
  Referenced from: /Users/davidh/anaconda/envs/sift_production/lib/libQtGui.4.8.7.dylib
  Reason: Incompatible library version: libQtGui.4.dylib requires version 39.0.0 or later, but libpng16.16.dylib provides version 34.0.0

After some googling I found that someone had success with downgrading qt to a earlier build, qt=4.8.7=1. After doing this everything runs fine.

Edit: Should issues like this go on conda/conda or here?

ipywidgets is several versions out of date

Looks like the last build in defaults is 4.1.1, latest is 5.2.2.

cmake and gfortran troubles

I can without a problem build my project outside of conda build but inside it crashes (macOS Sierra).

build.sh

#!/bin/bash
env
cd build
cmake ..

CMakeError.txt
CMakeOutput.txt
meta.txt

Packages require rebuilding with conda-build > 2.0

From @lparsons on February 23, 2017 22:51

I've run into an issues with a couple of packages (from the r channel) which give me the error:

PaddingError: Placeholder of length '80' too short in package [channel]::[package]
The package must be rebuilt with conda-build > 2.0.

While these are coming from the r channel, I believe this is the new home for these packages? Or at least could be?

The packages are:

r::r-base-3.2.1-0
r::r-base-3.2.2-0

Copied from original issue: conda-forge/r-base-feedstock#12

Adding obvious-ci to defaults

Not sure if this is the right place to ask this or not. Please feel free to redirect me to a more appropriate forum if not.

Given that conda-build-all is in defaults, was curious to see if obvious-ci could also be added to defaults as well. As the package does seem to be used by a larger community than just conda-forge, it seems reasonable for it to be in the defaults collection. Also nothing has changed substantially since the last release, which was ~1yr ago, as it is pretty stable. So the maintenance burden should be pretty light.

libgdal 2.1.0 build fails with an undefined reference to kealib error

Due to this issue, I tried to build libgdal recipe with conda-build. I got this error:

/home/.../opt/anaconda3/conda-bld/work/gdal-2.1.0/libgdal.so: undefined reference to `kealib::KEAImageIO::setImageMetaData(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >)'

my config:

             platform : linux-32
        conda version : 4.1.11
    conda-env version : 2.5.2
  conda-build version : 1.21.9
       python version : 3.5.2.final.0
     requests version : 2.10.0
     root environment : /home/.../opt/anaconda3  (writable)
  default environment : /home/.../opt/anaconda3
     envs directories : /home/.../opt/anaconda3/envs
        package cache : /home/.../opt/anaconda3/pkgs
         channel URLs : https://repo.continuum.io/pkgs/free/linux-32/
                        https://repo.continuum.io/pkgs/free/noarch/
                        https://repo.continuum.io/pkgs/pro/linux-32/
                        https://repo.continuum.io/pkgs/pro/noarch/
          config file : /home/.../.condarc
         offline mode : False
    is foreign system : False

Requests breaks ssl if crypto is not up-to-date

The latest requests can completely break conda!

The latest requests requires cryptography>1.3.4: https://github.com/kennethreitz/requests/blob/master/setup.py#L97

'security': ['pyOpenSSL>=0.14', 'cryptography>=1.3.4', 'idna>=2.0.0'],

Issue upstream: https://github.com/kennethreitz/requests/issues/3710

I had this version of cryptography installed:
cryptography 0.9.1

Stack trace:

 conda install conda
Fetching package metadata ...An unexpected error has occurred.
Please consider posting the following information to the
conda GitHub issue tracker at:

    https://github.com/conda/conda/issues



Current conda install:

               platform : osx-64
          conda version : 4.2.13
       conda is private : False
      conda-env version : 4.2.13
    conda-build version : 2.0.12
         python version : 2.7.13.final.0
       requests version : 2.12.4
       root environment : /opt/anaconda  (writable)
    default environment : /opt/anaconda
       envs directories : /opt/anaconda/envs
          package cache : /opt/anaconda/pkgs
           channel URLs : https://repo.continuum.io/pkgs/pro/osx-64
                          https://repo.continuum.io/pkgs/pro/noarch
                          https://repo.continuum.io/pkgs/free/osx-64
                          https://repo.continuum.io/pkgs/free/noarch
            config file : /opt/anaconda/.condarc
           offline mode : False



`$ /opt/anaconda/bin/conda install conda`




    Traceback (most recent call last):
      File "/opt/anaconda/lib/python2.7/site-packages/conda/exceptions.py", line 479, in conda_exception_handler
        return_value = func(*args, **kwargs)
      File "/opt/anaconda/lib/python2.7/site-packages/conda/cli/main.py", line 145, in _main
        exit_code = args.func(args, p)
      File "/opt/anaconda/lib/python2.7/site-packages/conda/cli/main_install.py", line 80, in execute
        install(args, parser, 'install')
      File "/opt/anaconda/lib/python2.7/site-packages/conda/cli/install.py", line 238, in install
        prefix=prefix)
      File "/opt/anaconda/lib/python2.7/site-packages/conda/api.py", line 24, in get_index
        index = fetch_index(channel_urls, use_cache=use_cache, unknown=unknown)
      File "/opt/anaconda/lib/python2.7/site-packages/conda/fetch.py", line 300, in fetch_index
        repodatas = [(u, f.result()) for u, f in zip(urls, futures)]
      File "/opt/anaconda/lib/python2.7/site-packages/concurrent/futures/_base.py", line 403, in result
        return self.__get_result()
      File "/opt/anaconda/lib/python2.7/site-packages/concurrent/futures/thread.py", line 55, in run
        result = self.fn(*self.args, **self.kwargs)
      File "/opt/anaconda/lib/python2.7/site-packages/conda/fetch.py", line 75, in func
        res = f(*args, **kwargs)
      File "/opt/anaconda/lib/python2.7/site-packages/conda/fetch.py", line 117, in fetch_repodata
        timeout=(6.1, 60))
      File "/opt/anaconda/lib/python2.7/site-packages/requests/sessions.py", line 501, in get
        return self.request('GET', url, **kwargs)
      File "/opt/anaconda/lib/python2.7/site-packages/requests/sessions.py", line 488, in request
        resp = self.send(prep, **send_kwargs)
      File "/opt/anaconda/lib/python2.7/site-packages/requests/sessions.py", line 609, in send
        r = adapter.send(request, **kwargs)
      File "/opt/anaconda/lib/python2.7/site-packages/requests/adapters.py", line 423, in send
        timeout=timeout
      File "/opt/anaconda/lib/python2.7/site-packages/requests/packages/urllib3/connectionpool.py", line 594, in urlopen
        chunked=chunked)
      File "/opt/anaconda/lib/python2.7/site-packages/requests/packages/urllib3/connectionpool.py", line 350, in _make_request
        self._validate_conn(conn)
      File "/opt/anaconda/lib/python2.7/site-packages/requests/packages/urllib3/connectionpool.py", line 835, in _validate_conn
        conn.connect()
      File "/opt/anaconda/lib/python2.7/site-packages/requests/packages/urllib3/connection.py", line 330, in connect
        cert = self.sock.getpeercert()
      File "/opt/anaconda/lib/python2.7/site-packages/requests/packages/urllib3/contrib/pyopenssl.py", line 324, in getpeercert
        'subjectAltName': get_subj_alt_name(x509)
      File "/opt/anaconda/lib/python2.7/site-packages/requests/packages/urllib3/contrib/pyopenssl.py", line 171, in get_subj_alt_name
        ext = cert.extensions.get_extension_for_class(
    AttributeError: 'Extensions' object has no attribute 'get_extension_for_class'

My fix: conda install -f /opt/anaconda/pkgs/requests-2.10.0-py27_0.tar.bz2

tensorflow mac builds

Hi is there any reason that opencv=3 and tensorflow do not have mac builds?

$ conda search --platform osx-64 opencv
Fetching package metadata .........
opencv                       2.4.8                np17py27_2  defaults
$ conda search --platform osx-64 tensorflow           
Fetching package metadata .........

NSIS error: could not find ${NSISDIR}\Contrib\Modern UI 2\MUI.nsh

Using conda constructor with the latest Miniconda install results in an error while invoking makensis

...
!include: could not find: "C:\Users\USERNAME\Miniconda3\NSIS\Contrib\Modern UI 2\MUI2.nsh"
!include: error in script: "C:\Users\USERNAME\Miniconda3\NSIS\Include\MUI2.nsh" on line 1
Error in script "C:\Users\USERNAME\AppData\Local\Temp\tmp5p19o9oq\main.nsi" on line 12 -- aborting creation process
Traceback (most recent call last):
  File "C:\Users\USERNAME\Miniconda3\Scripts\constructor-script.py", line 5, in <module>
    sys.exit(constructor.main.main())
  File "C:\Users\USERNAME\Miniconda3\lib\site-packages\constructor\main.py", line 177, in main
    verbose=opts.verbose, cache_dir=opts.cache_dir)
  File "C:\Users\USERNAME\Miniconda3\lib\site-packages\constructor\main.py", line 90, in main_build
    create(info)
  File "C:\Users\USERNAME\Miniconda3\lib\site-packages\constructor\winexe.py", line 187, in create
    check_call(args)
  File "C:\Users\USERNAME\Miniconda3\lib\subprocess.py", line 291, in check_call
    raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['C:\\Users\\USERNAME\\Miniconda3\\NSIS\\makensis.exe', '/V2', 'C:\\Users\\USERNAME\\AppData\\Local\\Temp\\tmp5p19o9oq\\main.nsi']' returned non-zero exit status 1.

Update request: gensim 0.13.4.1

May I ask to update the gensim in Anaconda to the next release?
The users are asking in Gensim mailing list.

Add pyside recipe

pyside is available through the default channel, but it's recipe is not in this repo. It would be great to add this at some point.

OpenCV and caffe builds compatible with anaconda 4.3.0

I would like to create an environment with anaconda 4.3.0 and some other packages (caffe and opencv) which we would use to develop bob at idiap (https://www.idiap.ch/software/bob/) but as the new version of anaconda has updated its dependencies, the rest of the stack here has not been updated. To be precise, anaconda 4.3.0 depends on jpeg 9 while opencv still uses jpeg 8 and anaconda 4.3.0 depends on mkl 2017.0.1 while caffe depends on mkl 11.3.3.

It would be great if Continuum could release a new version of these packages too which would be compatible with the newest version of anaconda.

I have also opened a pull request here: #80 which would update opencv to use jpeg 9. If needed, I can open another one to update caffe's dependencies.

ping @anjos

Recipes for official ContinuumIO builds not in Anaconda

I was curious to see the rpy2 recipe that is available via the official R channel, and was wondering if there were any plans to make packages that aren't in Anaconda or the main default channel publicly available?

For rpy2 in particular, this was one of the classically difficult packages to build on Windows. Up until recently there were only conda packages for OSX and Linux, but recently a Windows build was posted to the R channel. It however only supports py34 and py35 so I was looking for a template to build a py27 version from (or you can also consider this an official request to build a python 2.7 package for Windows 😃 ).

Readline package appears broken (Linux)

It appears the copy of readline that now fills the channels, possibly the same one provided here, is broken. To see what I mean, try downloading any of the docker images and install ipython with conda. The colors are missing, tab completion fails, using up/down arrows to traverse history fails. This is normally due to a broken readline or a copy of python that failed to link against readline. This really hurts the usability of things like the terminal based iPython.

spyder meta.yaml dependencies

I have had some trouble installing spyder into existing conda environments. I seem to be able to have worked around this by updating the packages that spyder depends on, I think all that really mattered was qt, namely going from 4.8.5 to 4.8.7. It suggests to me that the meta.yaml for spyder is not resctrictive enough on the dependencies.

For instance, the py35 error was

line 260, in set_editor
    from spyderlib.qt.QtWebKit import QWebView
  File "...//python3.5/site-packages/spyderlib/qt/QtWebKit.py", line 10, in <module>
    from PyQt4.QtWebKit import *  # analysis:ignore
ImportError: ...//python3.5/site-packages/PyQt4/QtNetwork.so: undefined symbol: _ZTI10QSslSocket

Python 2.6 recipe

Would be great if the Python 2.6 recipe was open sourced (assuming there is a conda-build recipe). Am curious to see what sort of patches were applied and so forth.

xz recipe does not copy all relevant license files

The xz package's license is spread across multiple files:
COPYING
COPYING.GPLv2
COPYING.GPLv3
COPYING.LGPLv2.1

However, the conda recipe meta.yaml file refers to only one of the license files: COPYING

As a result, the conda package does not include a copy of the license, which is required by the GPL license(s).

proj4 on Windows 7

I get this error when I try to install proj4 on Windows 7 in a Python 3.5 environment.

conda install proj4

The following specifications were found to be in conflict:
  - proj4 -> @vc10
  - proj4 -> @vc9

The conda package from SciTools works.

conda install -c scitools proj4

(linux-64) readline 6.2 has undefined symbols

I ran into a problem when I attempted to link against readline. The package coming from the defaults channel is not linked against ncurses and tinfo, so you end up with this:

$ ldd -r $CONDA_PREFIX/../../lib/libreadline.so.6.2
	linux-vdso.so.1 =>  (0x00007ffd2b35f000)
	libc.so.6 => /lib64/libc.so.6 (0x00007fb61ab23000)
	/lib64/ld-linux-x86-64.so.2 (0x0000003aa5400000)
undefined symbol: PC	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: UP	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: BC	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: tgetflag	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: tgetent	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: tputs	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: tgoto	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: tgetnum	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)
undefined symbol: tgetstr	(/internal/1/jhunk/miniconda3/lib/libreadline.so.6.2)

As a test I installed readline from conda-forge and linked against it without any problems. It doesn't seem to be linked against Continuum (or conda-forge's) ncurses, but maybe that's by design.

From "conda-forge" readline-6.3 package:

$ ldd -r $CONDA_PREFIX/lib/libreadline.so.6.3
	linux-vdso.so.1 =>  (0x00007fff831a9000)
	libncurses.so.5 => /lib64/libncurses.so.5 (0x00007fc1a1a80000)
	libtinfo.so.5 => /lib64/libtinfo.so.5 (0x00007fc1a185e000)
	libc.so.6 => /lib64/libc.so.6 (0x00007fc1a14ca000)
	libdl.so.2 => /lib64/libdl.so.2 (0x00007fc1a12c6000)
	/lib64/ld-linux-x86-64.so.2 (0x0000003aa5400000)

I would prefer to avoid providing conda-forge as a solution to my coworkers, because it introduces a lot more than just "readline". It could cause conflicts/bugs with the software developed to work specifically with the defaults channel.

Pylint dependencies are insufficient

Consider subsection from output of conda info pylint:

pylint 1.6.4 py27_1
-------------------
file name   : pylint-1.6.4-py27_1.tar.bz2
name        : pylint
version     : 1.6.4
build string: py27_1
build number: 1
channel     : defaults
size        : 316 KB
arch        : x86_64
date        : 2017-01-26
license     : GPL
license_family: GPL2
md5         : 0f1473fd07f18289f4008b8f58167e34
noarch      : None
platform    : darwin
url         : https://repo.continuum.io/pkgs/free/osx-64/pylint-1.6.4-py27_1.tar.bz2
dependencies:
    astroid >=1.4.1,<1.5.0
    isort >=4.2.5
    python 2.7*
    six

Actually, Pylint requires additional backports to work correctly.

extras_require = {}
extras_require[':sys_platform=="win32"'] = ['colorama']
extras_require[':python_version=="2.7"'] = ['configparser', 'backports.functools_lru_cache']

Same issue will arise in next release with singledispatch backport

Current conda install:

               platform : osx-64
          conda version : 4.3.9
       conda is private : False
      conda-env version : 4.3.9
    conda-build version : not installed
         python version : 2.7.13.final.0
       requests version : 2.12.4
       root environment : /Users/rogalski/anaconda  (writable)
    default environment : /Users/rogalski/anaconda
       envs directories : /Users/rogalski/anaconda/envs
          package cache : /Users/rogalski/anaconda/pkgs
           channel URLs : https://repo.continuum.io/pkgs/free/osx-64
                          https://repo.continuum.io/pkgs/free/noarch
                          https://repo.continuum.io/pkgs/r/osx-64
                          https://repo.continuum.io/pkgs/r/noarch
                          https://repo.continuum.io/pkgs/pro/osx-64
                          https://repo.continuum.io/pkgs/pro/noarch
            config file : /Users/rogalski/.condarc
           offline mode : False
             user-agent : conda/4.3.9 requests/2.12.4 CPython/2.7.13 Darwin/16.4.0 OSX/10.12.3
                UID:GID : 501:20

See pylint-dev/pylint#1296 for additional insight.

Question regarding conda-lua-science

Hello, everyone I've been trying to install conda-lua-science packages on a centos 6.6. It installs fine without any problems but when I test it I always get an error. I don't know if this is the right place to ask this question, let me know?

The error that I'm getting is the following one.

/home/user/anaconda2/share/lua/5.1/trepl/init.lua:384: loop or previous error loading module 'torch'
stack traceback:
    [C]: in function 'error'
    /home/user/anaconda2/share/lua/5.1/trepl/init.lua:384: in function 'require'
    [string "_RESULT={require('torch')}"]:1: in main chunk
    [C]: in function 'xpcall'
    /home/user/anaconda2/share/lua/5.1/trepl/init.lua:652: in function 'repl'
    ...e/user/anaconda2/lib/luarocks/rocks/trepl/scm-1/bin/th:199: in main chunk
    [C]: at 0x00405100

Thanks!

Please compile sqlite with readline support

It's so so so so much painful to use sqlite3 without the readline support. Here is an open issue already: ContinuumIO/anaconda-issues#307. People are suffering. Help them. Be hope :(

anaconda-navigator update error=CondaOSError: OS error: win32 hard link failed

The following NEW packages will be INSTALLED:

anaconda-navigator: 1.3.1-py35_0

Proceed ([y]/n)? y

Linking packages ...
CondaOSError: OS error: failed to link (src='C:\Anaconda\pkgs\anaconda-navigator-1.3.1-py35_0\Lib/site-packages/anaconda_navigator/static/images/spinner-green-16x16.gif', dst='C:\Anaconda\Lib/site-packages/anaconda_navigator/static/images/spinner-green-16x16.gif', type=1, error=CondaOSError: OS error: win32 hard link failed
)

Qmake invocations cause crash with XCode >= 8

Projects that use qmake crash when XCode => 8 is installed, issuing the following message:

+ /Users/daniel/anaconda/conda-bld/qscintilla2-qt5_1478527333650/_b_env_placehold_/bin/qmake qscintilla.pro -spec macx-llvm
Project ERROR: Xcode not set up properly. You may need to confirm the license agreement by running /usr/bin/xcodebuild.
Command failed: /bin/bash -x -e /Users/daniel/anaconda/conda-bld/qscintilla2-qt5_1478527333650/work/QScintilla_gpl-2.9.3/conda_build.sh

After Googling this error message, I found that the cause is that the invocation of xcrun changed to xcodebuild, see this StackOverflow question. I tried looking for the mentioned entry-to-change in anaconda/mkspecs/features/mac/default_pre.prf but it appears Anaconda uses a custom default_pre.prf file which does not contain the specified line: isEmpty($$list($$system("/usr/bin/xcrun -find xcrun 2>/dev/null")))

ipywidgets repo name conflict

We recently decided to name the IPython widget repo ipywidgets (see https://github.com/ipython/ipywidgets), and we have the ipywidgets name on pypi (https://pypi.python.org/pypi/ipywidgets/). Unfortunately, this conflicts with this repo's ipywidgets package from Jake Vanderplas.

OS X: gcc -fopenmp requires DLYD_FALLBACK_LIBRARY_PATH to be set

This seems new, and was noticed when some CI builds using miniconda started failing.

System is El Capitan:

Darwin Kevins-MacBook-Pro-2.local 15.6.0 Darwin Kernel Version 15.6.0: Thu Jun 23 18:25:34 PDT 2016; root:xnu-3248.60.10~1/RELEASE_X86_64 x86_64

I created a fresh user account and installed Anaconda/latest/OS X/64 bit/Python 3:

conda --version
conda 4.3.13

I then installed gcc and created the following C file:

#include <stdio.h>

int main(int argc, char **argv)
{
}

Compiling with -fopenmp causes a runtime issue that is fixed by setting DYLD_FALLBACK_LIBRARY_PATH. My understanding is that things are meant to work without declaring these:

Kevins-MacBook-Pro-2:~ condatest$ gcc -o test test.c
Kevins-MacBook-Pro-2:~ condatest$ ./test
Kevins-MacBook-Pro-2:~ condatest$ gcc -fopenmp -o test test.c 
Kevins-MacBook-Pro-2:~ condatest$ ./test
dyld: Library not loaded: @rpath/libgomp.1.dylib
  Referenced from: /Users/condatest/./test
  Reason: image not found
Trace/BPT trap: 5
Kevins-MacBook-Pro-2:~ condatest$ DYLD_FALLBACK_LIBRARY_PATH=$HOME/anaconda3/lib ./test

The issue seems specific to libgomp. Other C libraries work as expected. For example, after "conda install gsl", things work as expected:

Kevins-MacBook-Pro-2:~ condatest$ gcc  -o test test.c -lgsl -lgslcblas
Kevins-MacBook-Pro-2:~ condatest$ ./test