Skip to content

Appveyor: use newer conda packages and only run all tests on one platform #6774

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Jul 18, 2016
Merged

Appveyor: use newer conda packages and only run all tests on one platform #6774

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
4104449
Appveyor: use newer conda packages and only run all tests on one plat…
jankatins Jul 17, 2016
4032499
Appveyor: run fast builds first
jankatins Jul 17, 2016
c47edde
Revert "Appveyor: disable building conda packages as they fail the bu…
jankatins Jul 3, 2016
9e0f4af
Appveyor: remove py27 workaround, conda build works
jankatins Jul 7, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
55 changes: 21 additions & 34 deletions appveyor.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,35 +17,37 @@ environment:

matrix:
# for testing purpose: numpy 1.8 on py2.7, for the rest use 1.10/latest
# theoretically the CONDA_INSTALL_LOCN could be only two: one for 32bit, one for 64bit
# but using one for the right python version is hopefully making it fast due to package caching.
- TARGET_ARCH: "x86"
CONDA_PY: "27"
CONDA_NPY: "18"
PYTHON_VERSION: "2.7"
# this variable only influence pdf/svg tests, png tests are run on any platform
# only once, because it basically triples the testing times
# pick the one which seems to make the most problems
TEST_ALL_IMAGES: "yes"
CONDA_INSTALL_LOCN: "C:\\Miniconda"
# theoretically the CONDA_INSTALL_LOCN could be only two: one for 32bit,
# one for 64bit because we construct envs anyway. But using one for the
# right python version is hopefully making it fast due to package caching.
- TARGET_ARCH: "x64"
CONDA_PY: "27"
CONDA_NPY: "18"
PYTHON_VERSION: "2.7"
TEST_ALL_IMAGES: "no"
TEST_ALL: "no"
CONDA_INSTALL_LOCN: "C:\\Miniconda-x64"
- TARGET_ARCH: "x64"
CONDA_PY: "34"
CONDA_NPY: "110"
PYTHON_VERSION: "3.4"
TEST_ALL_IMAGES: "no"
TEST_ALL: "no"
CONDA_INSTALL_LOCN: "C:\\Miniconda3-x64"
- TARGET_ARCH: "x64"
CONDA_PY: "35"
CONDA_NPY: "110"
PYTHON_VERSION: "3.5"
TEST_ALL_IMAGES: "no"
TEST_ALL: "no"
CONDA_INSTALL_LOCN: "C:\\Miniconda35-x64"
- TARGET_ARCH: "x86"
CONDA_PY: "27"
CONDA_NPY: "18"
PYTHON_VERSION: "2.7"
# this variable influence pdf/svg and most importantly the latex related tests
# which triples the runtime of the tests (7-8min vs 30min).
# pick the one which seems to make the most problems and run it last, so that
# the rest of the tests can give feedback earlier
TEST_ALL: "yes"
CONDA_INSTALL_LOCN: "C:\\Miniconda"

# We always use a 64-bit machine, but can build x86 distributions
# with the PYTHON_ARCH variable (which is used by CMD_IN_ENV).
Expand Down Expand Up @@ -75,9 +77,6 @@ install:
# Fix the appveyor build environment to work with conda build
# workaround for missing vcvars64.bat in py34 64bit
- cmd: copy ci\appveyor\vcvars64.bat "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\bin\amd64"
# workaround for conda build on py27 prefering the normal installed
# VS tools instead of the also installed Py27 VS compiler (which wouldn't need this workarounds...)
- cmd: copy "C:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\bin\vcvars64.bat" "C:\Program Files (x86)\Microsoft Visual Studio 9.0\VC\bin\amd64\vcvarsamd64.bat"

# For building, use a new environment which only includes the requirements for mpl
# same things as the requirements in ci/conda_recipe/meta.yaml
Expand Down Expand Up @@ -116,23 +115,12 @@ test_script:
- cmd: '"%DUMPBIN%" /DEPENDENTS lib\\matplotlib\\_png*.pyd | findstr png.*.dll && exit /b 1 || exit /b 0'

# this are optional dependencies so that we don't skip so many tests...
# WORKAROUND: set miktex to the old version until I figured out whats wrong here...
- cmd: conda install -q pillow miktex=2.9.5857
# autoinstall latex packages (0=no, 1=autoinstall, 2=ask)
# this adds this to the registry!
- cmd: initexmf --set-config-value "[MPM]AutoInstall=1"
# we need to put the real exe files into PATH as subprocess can't call bat files
# this works for the "old" version of the miktex package, then next version should
# not need it
- set "PATH=%PATH%;%CONDA_PREFIX%\Library\miktex\miktex\bin"
# missing packages on conda-forge for ffmpeg avconv mencoder imagemagick inkscape
- cmd: conda install -q -c menpo ffmpeg # a repackaged version
- cmd: if x%TEST_ALL% == xyes; conda install -q pillow miktex inkscape
# missing packages on conda-forge for ffmpeg avconv mencoder imagemagick
- cmd: if x%TEST_ALL% == xyes; conda install -q -c menpo ffmpeg # a repackaged version
# This install sometimes failed randomly :-(
#- cmd: choco install imagemagick
# only install inkscape during one test run as it increases the runtime of one
# test run from 13 min -> 30 min
- cmd: if x%TEST_ALL_IMAGES% == xyes; choco install inkscape
- cmd: if x%TEST_ALL_IMAGES% == xyes; set "PATH=%PATH%;C:\Program Files\Inkscape\inkscape.com"

# Test import of tkagg backend
- python -c "import matplotlib as m; m.use('tkagg'); import matplotlib.pyplot as plt; print(plt.get_backend())"
# tests
Expand All @@ -157,8 +145,7 @@ after_test:
- cmd: path
- cmd: where python
- cmd: '%CMD_IN_ENV% conda config --get channels'
# currently disabled as conda-build errors :-(
#- cmd: '%CMD_IN_ENV% conda build -q .\ci\conda_recipe'
- cmd: '%CMD_IN_ENV% conda build -q .\ci\conda_recipe'

# Move the conda package into the dist directory, to register it
# as an "artifact" for Appveyor.
Expand Down