Releases

Releases can be viewed on anaconda.org.

Under normal circumstances, each commit into master should have a build uploaded to Anaconda under the nightly label. If the nightly build is ready to be released to the public, simply edit the labels such that it has both the main and nightly labels. Then, go to Github and draft a new release to match the version on Conda.

vCDAT Manual Releases

In the event that the continuous integration service fails to make a nightly release, the following can be done to create a build and release it manually.

  • Before you can release vCDAT you must be listed as a co-owner of the cdat organization
  • cd into the vcdat directory and make sure that you are using the base conda environment. (Run source deactivate to go from nightly to base for instance)

Once that is done, begin running each command below:

mkdir ~/conda-bld # Creates a directory for our build to go
conda install -q anaconda-client conda-build # Needed in order to build and upload our package
conda config --set anaconda_upload no # Prevents auto upload since we want to do that ourselves
conda install -f six # Not sure this is needed. It was in the old script, and it doesn't hurt to have it
git clone git://github.com/cdat/conda-recipes # Contains the prep_for_build script that we use
cd conda-recipes
ln -s ../conda vcdat # Creates a symbolic link to our conda directory so that prep_for_build can find it
python ./prep_for_build.py -v $NEW_VERSION # Creates our meta.yaml with the given semantic version number (0.0.0)
conda build -c cdat -c conda-forge vcdat # Build the production version of the package
anaconda -t $CONDA_UPLOAD_TOKEN upload -u cdat $BUILT_PACKAGE --force
# $CONDA_UPLOAD_TOKEN: can be retrieved from https://anaconda.org/cdat/settings/access
# $BUILT_PACKAGE: The full path to the newly built .tar.bz2 package. Use the path output by conda build

Make sure to test the release:

conda create -n vcdat-version vcdat -c cdat -c conda-forge
source activate vcdat-version
vcdat --print

Lastly, if the release is in the ‘main’ label, go to Github and draft a new release to match the version on Conda.

VCS-Widgets Release:

cd ~/Projects/CDAT/vcs-widgets
git pull origin master
source activate vcdat

Edit package.json with new version number

npm build
npm publish

https://anaconda.org/cdat

https://www.npmjs.com/package/vcs-widgets