currently it is a manual process:
-
Have a git clone of this repo
-
Have python 3 (no additional libraries needed)
-
Github access: The script expects a Github token in the
GITHUB_TOKENenvironment variable. You can create such a token in github.com: "Settings"/"Developer settings"/"Personal access tokens"/"Tokens (classic)". It requires thereposcope to access the releases of the private repository. -
S3 access: For access to our S3-provider "Wasabi" the script uses the AWS CLI tools (https://github.com/aws/aws-cli, Debian package
awscli, command binaryaws, uses~/.aws/credentials). Before using the script the tools must be configured (aws configure). Use the "AWS Access Key ID" and "AWS Secret Access Key" of an authorized user and useeu-central-1as the "Default region name".
- For storing the release assets, the bucket
fylr-releasesis used.
- execute
makein the root directory of this repo.
-
A python3 script reads the release notes from fylr's releases on GitHub.
-
The files (release artifacts/products) are copied to S3 (Wasabi). Files are only added when they are not present yet.
-
Files will be updated:
SUMMARY.md,releases/*.md,releases/*/*.md -
If a release artifact is removed later, it is not removed from s3 but from the releases/...md files.
-
If an artifact is changed later (e.g. release re-done on github), manually delete its file(s) in S3 bucket fylr-releases and
make. -
if you change the relase notes later then it is being recognized by the next run with
make. -
Drafts and PreReleases are ignored.
- Do e.g.
git statusto see what was changed and then add, commit, push