appveyor.yml

environment:
  matrix:
    # For Python versions available on Appveyor, see
    # http://www.appveyor.com/docs/installed-software#python
    # Use Anaconda as SciPy is a headache with Windows.
      # 2.7
    - PYTHON: "C:\\Miniconda"
      # 2.7 64-bit
    - PYTHON: "C:\\Miniconda-x64"
      # 3.4
    - PYTHON: "C:\\Miniconda3"
      # 3.4 64-bit
    - PYTHON: "C:\\Miniconda3-x64"
      # 3.5
    - PYTHON: "C:\\Miniconda35"
      # 3.5 64-bit
    - PYTHON: "C:\\Miniconda35-x64"

install:
  # Prepend chosen Python to the PATH of this build
  - "SET PATH=%PYTHON%;%PYTHON%\\Scripts;%PATH%"
  # Check that we have the expected version and architecture for Python
  - "python --version"
  - "python -c \"import struct; print(struct.calcsize('P') * 8)\""
  # Install the conda supplied packages; msvc_runtime needed for py34
  - "conda update -y conda"
  - "conda install -y pip wheel numpy scipy pyflakes msvc_runtime"
  # install ase into the current python
  - "echo %cd%"
  - "where pip"
  - "pip install ."

build: off

test_script:
  # run tests from temp dir so source tree doesn't interfere
  - "cd %TEMP%"
  - "python -m ase.test"

after_test:
  # This step builds distribution.
  - "cd %APPVEYOR_BUILD_FOLDER%"
  # use pip as distutils doesn't implemet bdist_wheel
  - "pip wheel -w dist ."

artifacts:
  # bdist_wheel puts your built wheel in the dist directory
  - path: dist\*

#on_success:
#  You can use this step to upload your artifacts to a public website.
#  See Appveyor's documentation for more details. Or you can simply
#  access your wheels from the Appveyor "artifacts" tab for your build.