Repository navigation
Closed
Labels
Description
It looks like build issues related to numpy for python < 3.5.
attn @JanSchulz
Looking at the history it does not look like there was a merge to master between it working and not working.
https://ci.appveyor.com/project/mdboom/matplotlib/history
worked: https://ci.appveyor.com/project/mdboom/matplotlib/build/1.0.963
failed: https://ci.appveyor.com/project/mdboom/matplotlib/build/1.0.964
Looking at diffs between the logs it looks like there is a change of py 3.4.3 -> 3.4.4 3.4.4 -> 3.4.3
pass env
conda list
# packages in environment at C:\conda\envs\test-environment:
#
backports-abc 0.4 <pip>
backports_abc 0.4 py34_0 defaults
cycler 0.10.0 py34_0 conda-forge
freetype 2.5.5 vc10_0 [vc10] defaults
funcsigs 0.4 py34_0 defaults
jpeg 8d vc10_0 [vc10] defaults
libpng 1.6.17 vc10_1 [vc10] defaults
libtiff 4.0.6 vc10_1 [vc10] defaults
mkl 11.3.1 0 defaults
mock 1.3.0 py34_0 defaults
msinttypes r26 0 conda-forge
msvc_runtime 1.0.1 vc10_0 [vc10] defaults
nose 1.3.7 py34_0 defaults
numpy 1.10.4 py34_0 defaults
pbr
1.3.0 py34_0 defaults
pip 8.0.3 py34_0 defaults
pyparsing 2.0.3 py34_0 defaults
pyqt 4.11.4 py34_5 defaults
python 3.4.4 2 defaults
python-dateutil 2.4.2 py34_0 defaults
pytz 2015.7 py34_0 defaults
qt 4.8.7 vc10_6 [vc10] defaults
setuptools 20.1.1 py34_0 defaults
sip 4.16.9 py34_2 defaults
six 1.10.0 py34_0 defaults
tk 8.5.18 vc10_0 [vc10] defaults
tornado 4.3 py34_0 defaults
wheel 0.29.0 py34_0 defaults
zlib
1.2.8 vc10_2 [vc10] defaults
%CMD_IN_ENV% python setup.py develop
fail env
conda list
# packages in environment at C:\conda\envs\test-environment:
#
backports-abc 0.4 <pip>
backports_abc 0.4 py34_0 defaults
cycler 0.10.0 py34_0 conda-forge
freetype 2.5.5 vc10_0 [vc10] defaults
funcsigs 0.4 py34_0 defaults
libpng 1.6.17 vc10_1 [vc10] defaults
mkl 11.3.1 0 defaults
mock 1.3.0 py34_0 defaults
msinttypes r26 0 conda-forge
nose 1.3.7 py34_0 defaults
numpy 1.10.4 py34_0 defaults
pbr 1.3.0 py34_0 defaults
pip 8.0.3 py34_0 defaults
pyparsing 2.0.3 py34_0 defaults
pyqt
4.10.4 py34_1 defaults
python 3.4.3 0 defaults
python-dateutil 2.4.2 py34_0 defaults
pytz 2015.7 py34_0 defaults
setuptools 20.1.1 py34_0 defaults
six 1.10.0 py34_0 defaults
tk 8.5.18 vc10_0 [vc10] defaults
tornado 4.3 py34_0 defaults
wheel 0.29.0 py34_0 defaults
zlib 1.2.8 vc10_2 [vc10] defaults
%CMD_IN_ENV% python setup.py develop
Configuring Windows SDK "v7.1" for Python 3 on a 64 bit architecture
[snip]
running build_ext
building 'matplotlib.ft2font' extension
Traceback (most recent call last):
File "setup.py", line 308, in <module>
**extra_args
File "C:\conda\envs\_build\lib\distutils\core.py", line 148, in setup
dist.run_commands()
File "C:\conda\envs\_build\lib\distutils\dist.py", line 955, in run_commands
self.run_command(cmd)
File "C:\conda\envs\_build\lib\distutils\dist.py", line 974, in run_command
cmd_obj.run()
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\install.py", line 67, in run
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\install.py", line 109, in do_egg_install
File "C:\conda\envs\_build\lib\distutils\cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "C:\conda\envs\_build\lib\distutils\dist.py", line 974, in run_command
cmd_obj.run()
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\bdist_egg.py", line 161, in run
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\bdist_egg.py", line 147, in call_command
File "C:\conda\envs\_build\lib\distutils\cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "C:\conda\envs\_build\lib\distutils\dist.py", line 974, in run_command
cmd_obj.run()
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\install_lib.py", line 10, in run
File "C:\conda\envs\_build\lib\distutils\command\install_lib.py", line 107, in build
self.run_command('build_ext')
File "C:\conda\envs\_build\lib\distutils\cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "C:\conda\envs\_build\lib\distutils\dist.py", line 974, in run_command
cmd_obj.run()
File "setup.py", line 149, in run
return BuildExtCommand.run(self)
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\build_ext.py", line 49, in run
File "C:\conda\envs\_build\lib\distutils\command\build_ext.py", line 339, in run
self.build_extensions()
File "C:\conda\envs\_build\lib\distutils\command\build_ext.py", line 448, in build_extensions
self.build_extension(ext)
File "c:\conda\envs\_build\lib\site-packages\setuptools-20.1.1-py3.4.egg\setuptools\command\build_ext.py", line 174, in build_extension
File "C:\conda\envs\_build\lib\distutils\command\build_ext.py", line 500, in build_extension
include_dirs=ext.include_dirs,
File "C:\conda\conda-bld\work\setupext.py", line 850, in __get__
result = obj._hooks[self._name]() + result
File "C:\conda\conda-bld\work\setupext.py", line 870, in include_dirs_hook
import numpy
File "C:\conda\envs\_build\lib\site-packages\numpy\__init__.py", line 180, in <module>
from . import add_newdocs
File "C:\conda\envs\_build\lib\site-packages\numpy\add_newdocs.py", line 13, in <module>
from numpy.lib import add_newdoc
File "C:\conda\envs\_build\lib\site-packages\numpy\lib\__init__.py", line 8, in <module>
from .type_check import *
File "C:\conda\envs\_build\lib\site-packages\numpy\lib\type_check.py", line 11, in <module>
import numpy.core.numeric as _nx
File "C:\conda\envs\_build\lib\site-packages\numpy\core\__init__.py", line 14, in <module>
from . import multiarray
ImportError: DLL load failed: The specified module could not be found.
C:\conda\conda-bld\work>if errorlevel 1 exit 1
Command failed: C:\windows\system32\cmd.exe /c call bld.bat
Command exited with code 1
Activity
tacaswell commentedon Mar 6, 2016
Sorry, I flipped the fail/pass env in my initial post. If you are reading this in GH I have fixed it, if you are reading the emails they are flipped.
tacaswell commentedon Mar 6, 2016
I have a suspicion that this might a bug due to the new conda dependency solver as the command
results in different environments
CI: try explicitly including msvc_runtime
jankatins commentedon Mar 6, 2016
I've no clue why this error happen but you guess your suspicion might be correct. It looks like #6116 fix it, thanks!
jankatins commentedon Mar 6, 2016
It seems one of the differences between the linked logs is conda 3.19.3-py27_0 (works) vs 4.0.0-py27_0 . What I don't understand is why this worked before?
There seems to be two different errors: on py2.7 it fails with a missing libpng and on 3.4 with some problems on numpy imports.
CC: @mcg1969 for some advise?
mcg1969 commentedon Mar 6, 2016
I can take a look later today, but we have identified and fixed one issue in 4.0.0 that will be pushed out ASAP.
mcg1969 commentedon Mar 6, 2016
Can you give me some specifics here on what you expect
condato do here? Please keep in mind: the previous conda solver was broken in numerous ways; it's not sufficient to simply say "the behavior has changed", because that was the goal :-) But if you think that conda is not producing the proper package solution here, let's figure that out. It may indeed be a new or remaining issue in the solver, and it might also be that it's now interpreting the metadata properly, and that metadata needs adjusting. (We've already seen that in one instance.)jankatins commentedon Mar 6, 2016
@mcg1969 It's more of a "if this is now ok, we are ok" thing :-) IMO there are two things:
msvc_runtime-> PR in #6116
mcg1969 commentedon Mar 7, 2016
This seems to be related to the
msvc_runtimeissue: ContinuumIO/anaconda-issues#443 (comment)I think that's something that's going to be fixed very soon, possibly with a metadata-only fix.
mcg1969 commentedon Mar 9, 2016
We did find a couple of issues related to features (which
msvc_runtimecontrols) that might have affected this issue. It might be worth revisiting this now that conda 4.0.2 is out.Fix multiarray import error on appveyor
Fix multiarray import error on appveyor
anntzer commentedon Mar 19, 2018
Clearly seems resolved now.