Travis CI failing in multiple ways #17
Description
Travis CI is consistently failing on slycot builds. Here is a summary of the Travis CI build failures from PR #13:
- Travis CI succeeds testing
slycotin python 2.7 - Travis CI fails in testing
slycotin python 3.3, due to inability of conda to resolve dependencies
3.4, 3.5 - Travis CI success in testing
python-controlin python 2.7, 3.5, 3.6 (test results)
The dependency issue in python 3.3 and 3.4 gives the following error:
conda build --python "$TRAVIS_PYTHON_VERSION" conda-recipe
Adding in variants from internal_defaults
INFO:conda_build.variants:Adding in variants from internal_defaults
Adding in variants from config.variant
INFO:conda_build.variants:Adding in variants from config.variant
Attempting to finalize metadata for slycot
INFO:conda_build.metadata:Attempting to finalize metadata for slycot
Solving environment: failed
failed to get install actions, retrying. exception was: The following specifications were found to be in conflict:
- numpy[version='>=1.13.3'] -> python[version='>=2.7,<2.8.0a0'] -> readline=7
- numpy[version='>=1.13.3'] -> python[version='>=2.7,<2.8.0a0'] -> tk=8.6
- python=3.3
A different error occurs in python 3.5. The various dependencies seem to work OK there, but there is an error when building the slycot library:
/usr/bin/gfortran -Wall -g -shared <long list of files> -llapack -lblas -lpython3.5m -lgfortran -o build/lib.linux-x86_64-3.5/slycot/_wrapper.cpython-35m-x86_64-linux-gnu.so
/usr/bin/ld: cannot find -llapack
/usr/bin/ld: cannot find -lblas
/usr/bin/ld: cannot find -lpython3.5m
collect2: error: ld returned 1 exit status
This must be some sort of error in the build environment.
The above build issues are not a problem in my local environment (MacOS 10.12.6), so likely something having to do with the Travis CI build environment (which is done under Ubuntu Trusty, I think).
Finally, there is yet a different error in PR #13, where the issue there appears to be in testMinrealBrute, which generates the following error:
FAIL: testMinrealBrute (control.tests.minreal_test.TestMinreal)
Traceback (most recent call last):
File "/Users/murray/Dropbox/macosx/src/python-control/murrayrm/control/tests/minreal_test.py", line 62, in testMinrealBrute
raise e
File "/Users/murray/Dropbox/macosx/src/python-control/murrayrm/control/tests/minreal_test.py", line 57, in testMinrealBrute
ht1.den[0][0], ht2.den[0][0])
File "/Users/murray/Dropbox/macosx/src/python-control/murrayrm/control/tests/minreal_test.py", line 35, in assert_numden_almost_equal
np.testing.assert_array_almost_equal(n1, n2)
File "/Users/murray/anaconda/envs/test_slycot/lib/python3.6/site-packages/numpy/testing/utils.py", line 962, in assert_array_almost_equal
precision=decimal)
File "/Users/murray/anaconda/envs/test_slycot/lib/python3.6/site-packages/numpy/testing/utils.py", line 715, in assert_array_compare
raise AssertionError(msg)
AssertionError:
Arrays are not almost equal to 6 decimals
(shapes (3,), (4,) mismatch)
x: array([ -1.33214 , -10.368378, -13.703545])
y: array([ -1.33214 , -13.297947, -36.505109, -30.136079])
I'm going to try to get a working version of all of this up and going, so that we can get the integration tests working again. Will document progress here.