[Question] Why use GITHUB_TOKEN as repo_token?

https://docs.coveralls.io/parallel-build-webhook In the API reference page, the format of API call is like the following

curl -k $COVERALLS_ENDPOINT/webhook?repo_token=$COVERALLS_REPO_TOKEN 
     -d "payload[build_num]=$BUILD_NUMBER&payload[status]=done"

I don’t understand why in this github-action, we can use GITHUB_TOKEN instead of the REPO_TOKEN in the coveralls.io repository page? Could you explain to us? Thank you.

[rdmurphy]

I have the same question and maybe can clarify what (I think) @changhuixu was asking — why does passing the GITHUB_TOKEN provided by GitHub Actions work as the COVERALLS_REPO_TOKEN?

Based on how I’m reading the API reference this token eventually gets passed as repo_token and that should be coming from Coveralls, not GitHub.

[bboe]

I was curious about this question so I tested out coveralls API: https://docs.coveralls.io/api-reference

When the service_name field is passed as github then coveralls expects the repo_token field to be a github token so that it can directly call back to github. In fact, in such a case, using a legit coveralls repo token will result in a failure unless the service_name field is changed to something else.

Here’s an example for how I am using this action for parallel-finished in tandem with the python-coveralls: https://github.com/praw-dev/prawtools/commit/bb0e00b89a899fb2494cf230c0104746ddaeb098