>Aren't the snapshots running fairly well vetted code anyway - only >using code that's been accepted into the source tree? Obviously not >as well vetted as the -STABLE and -RELEASE, of course.
Snapshots are generated as fast as we can, from what is commited. What gets commited may contain errors. It happens. We keep an eye out to fix those errors and move on. They also contain features. Features mail contain errors. Read the previous sentence. It is best effort, but 20 years of best effort has produced something kind of cool -- and a big piece is that people run snapshots and find bugs early after commit while developers are fresh, so they can be fixed. It is a healthy feedback loop. Not unique to OpenBSD, but it is the fundamental component that drives quality in open source projects -- if both sides care about quality. Not all our users need to be part of this feedback loop, but it is healthy when there are enough people taking part. Enough is an unquantifiable number. If bugs get found by snapshot users, it is working. When they don't find bugs, we don't know if there are too few bugs or too few snapshot users.....