@@ -85,8 +85,16 @@ Anaconda leverages caching in many places which can cause issues. If you are get
* Run ```conda build recipe/ -c usgs-astrogeology -c conda-forge --no-test```
* The -c options are to give conda a channel priority. Without these, conda will always attempt to download from the default Anaconda channel first. (Unlike the environment.yml files, channel priority cannot be defined within the meta.yaml.)
* Since we do not have testing set-up through conda, the “--no-test” flag must be set in order to avoid errors. (By default, conda looks for a run_test file and will throw an error if it can not be located.)
## Step 4: Test the Conda Build
After the conda build completes, it should be tested by uploading it to your personal anaconda cloud account and conda installing it locally.
* Use the command ```anaconda upload -u <conda-cloud-username> <path-to-the-.tar.bz2-file>``` to upload the conda build to your personal anaconda channel.
* Follow the standard [installation instructions](https://github.com/USGS-Astrogeology/ISIS3#isis3-installation-with-conda) to install this package locally for testing, but at the installation step, instead of running `conda install -c usgs-astrogeology isis`, run `conda install -c <conda-cloud-username> isis`
* Run an ISIS application, like `spiceinit -h` and insure that it runs without error. This is testing whether the conda environment set up during the install is complete enough to run ISIS.
## Step 5: Upload the Build to Anaconda Cloud
@@ -94,10 +102,10 @@ In this step, we will upload the build(s) that we just created into the Anaconda
* If you missed the location of the .tar.bz2 file at the bottom of the build, use ```conda build recipe/ --output``` to reprint. This command does not confirm the file exists - only where it *would* be saved with a successful build.
* For a standard release, use the command ```anaconda upload -u usgs-astrogeology <path-to-the-.tar.bz2-file>```.
* For an Release Candidate, use the command ```anaconda upload -u usgs-astrogeology -l RC <path-to-the-.tar.bz2-file>```.
* For a custom build, specify a custom label and use the command ```anaconda upload -u usgs-astrogeology -l <custom-label> <path-to-the-.tar.bz2-file>```.
* For example, when generating a custom build for the CaSSIS team, we would use the "cassis" label and the command ```anaconda upload -u usgs-astrogeology -l cassis <path-to-the-.tar.bz2-file>```.
* For a standard release, use the command ```anaconda upload -u usgs-astrogeology <path-to-the-.tar.bz2-file>```
* For an Release Candidate, use the command ```anaconda upload -u usgs-astrogeology -l RC <path-to-the-.tar.bz2-file>```
* For a custom build, specify a custom label and use the command ```anaconda upload -u usgs-astrogeology -l <custom-label> <path-to-the-.tar.bz2-file>```
* For example, when generating a custom build for the CaSSIS team, we would use the "cassis" label and the command ```anaconda upload -u usgs-astrogeology -l cassis <path-to-the-.tar.bz2-file>```
If the upload fails or displays a prompt for a username and password, try adding an API token for usgs-astrogeology to your environment by running `export ANACONDA_API_TOKEN=<token>`. Ask another developer for the API token. This approach is recommended over adding `-t <token>` to your anaconda upload command, because of a known bug where `-t` is either interpreted as a package type or a token depending on its position in the `anaconda upload` command.
