Best practice for API integration tests on track repository?

Hi Since Pharo track is dependent on https API (and not CLI), there was recent issue caused by availability of api.exercism.io. Original motivation was to be not dependent on CLI tool as external dependency artefact but rather use Pharo internal HTTP client instead. Mentioned API access works, but recently some developer found request gate by Cloudflare challenge: https://github.com/exercism/pharo-smalltalk/issues/681

Is it a good practice to have CI job just checking API availability (e.g. triggered once per week)?

1 Like

I don’t think that any low rate API access might be a problem. But I’m also not convinced, that such CI job would help with anything but permanent breaking changes on the API. And there are so many more reasons today why HTTP requests fail for users for a day or two. So maybe it would be better to collect failure signals from your Pharo tooling where you would like to know it?