I recently updated the README ¹ on my Github profile to add a new blog post entry, which made me realise this manual process was a good candidate for automation.

I took inspiration from Simon Willison’s blog to utilise Github Actions for this. The Github Action calls a script that:

• Fetch and unmarshall rss feed items from this blog
• Render README template using feed data
• Updates the current README.md if there’s a diff

The core Go script orchestrates fetching the data and rendering the template. Here’s a snippet:

func main() {
	ctx := context.Background()
	var readmeInfo ReadmeInfo

	errGroup, ctx := errgroup.WithContext(ctx)
	errGroup.SetLimit(5)

	errGroup.Go(func() error {
		var err error
		readmeInfo.Posts, err = fetchRSSFeed(ctx, rssSource)
		return err
	})

	if err := errGroup.Wait(); err != nil {
		log.Fatalf("failed to fetch feed data: %+v\n", err)
	}

	if err := renderTemplate(&readmeInfo); err != nil {
		log.Fatalf("failed to render readme, errr: %+v", err)
	}
}

This is currently configured to run every push to the repo, and on schedule for every 6 hours.

What I like about the structure of this script is how I can extend it to pull multiple feed entries. It only pulls my latest posts, and with this new Shorts category I’ve added for TILs and short-form content, I can extend the ReadmeInfo to include a Shorts slice, and add another routine to concurrently fetch the rss info for shorts.