Creating and updating GitHub issues along with batch changes

July 7, 2022 · View on GitHub

This batch change is an example of how to create tickets alongside changesets. It will:

Find repositories with fmt.Sprintf
Open changesets to replace them with strconv.Itoa
For each open changeset, create a GH issue mentioning the files that have been modified

When reapplying the spec, it will

first check for existing issues matching the changes and update them if they exist
otherwise create new issues

It demonstrates how to use steps.outputs and batch spec templating.

Note: by convention, all the issues tracked by this script must be named batch-change/<batch-change-name>. In this example batch-change/refactor-go-gh-issue

How to

Modify the gh_username flag with your own in gh-issues.batch.yml
Create a GitHub Personal Access Token and set it as the GH_TOKEN environment variable.

export GH_TOKEN=mysecrettoken
src batch apply -f gh.issues.batch.yml

This spec uses batch spec templating to produce text values inserted in the spec at runtime. They are delimited by ${{ and }}. For example, one of the batch changes step passes the name of the batch change, to create a corresponding issue. Using the ${{batch_change.name}} statement allows for it to be dynamically read from the spec name, instead of hard coded.

name: refactor-go-gh-issue

steps:
...
- run: >
      python ../sync-issue.py
      ...
      # no need to hardcode the batch change name twice!
      --batch_change_name ${{batch_change.name}}
      ...

Limitations

For each changeset, the script will load all the issues in the repository affected by the changeset. That is inefficient, and a way better solution would be to maintain an index of all batch changes associated issues.
This does not work in a monorepo as it assumes that there is only one issue matching a given batch change in any given repository.

How to

Getting to know Batch Changes

Limitations