 |
Where To From Here?
- Track packages when they fail, so we can find out why they
become broken in the first place:
- fetch problems (not necessarily the maintainer's fault; applications change rapidly)
- dependency changed
- base system changed
- insufficiently tested commits
This will help portmgr to spend a lot less time trying to notify maintainers of individual errors.
- Better comparisons of package sets (from one buildenv to another) to try to isolate the chokepoints
- Integrate with the updated ports survey
- Integrate with fenner's fetch failure survey
- Add ability to subscribe to some reports by email (RSS?)