Why not deal with github pull requests efficiently?
Usually due to one of these:
- Lack of manpower (commit touches part of the system nobody is familiar with)
- Lack of interest (PR is too complex or trivial to attract a reviewer)
- PR has various little issues which make it annoying to merge (some post processing is required, Sob missing or wrong, description not ideal, style quirks in code)
- PR has unresolved discussion
I just tried to merge ten trivial random PRs, among these three required a rework, one had a broken commit message which I accidentally committed, one added unrelated stuff to the PR, one required a complete rewording of subject + message + exchange of Sob... overall this task took about half an hour for trivial changes to shell script oriented packages.
You can imagine that reviewing and testing complex PRs with hard to determine consequences (e.g. due to touching kernel, system wide components, versions) takes significantly longer, coupled with an overall rather low ratio of contributors actively merging PRs.