PR size doesn't necessarily reflect complexity.
We could set a static code analysis for PRs, and there are many static analysis tools for GitHub.
There seems to be a MIT licensed app for this already using probot and aws lambda serverless. A minor feature really that can be implemented in many ways (ie. just an action with docker runner that pulls the git meta from the repo and calculates the size based on commit diff)
How to filter out 'generated' changes ie. package-lock,json?
PR size doesn't necessarily reflect complexity. We could set a static code analysis for PRs, and there are many static analysis tools for GitHub.
There seems to be a MIT licensed app for this already using probot and aws lambda serverless. A minor feature really that can be implemented in many ways (ie. just an action with docker runner that pulls the git meta from the repo and calculates the size based on commit diff)
How to filter out 'generated' changes ie. package-lock,json?
https://github.com/noqcks/pull-request-size