Currently, we do not apply maintenance releases (e.g., 2.0.5) during a semester. We are considering changing that policy. If we do, is there any QA reason for preferring an official maintenance release (e.g., 2.0.5) over a weekly release (e.g., 2.0.5+)? For example, does the official maintenance release get more regression testing than a weekly release?
Thanks.
Normally, there is wait of a few days between a weekly release like the final 2.1.1+, and then re-badging it and releasing it as Moodle 2.1.2. Therefore, if something critical is broken in a weekly release, it will probably be fixed before the real release.
Also, when the integrators know a point release like 2.1.2 is coming up, they are probably a bit more risk-averse when reviewing changes.
So, point releases like 2.1.2 probably are a little bit more stable, but it is marginal. The new development processes (http://docs.moodle.org/dev/Process) seem to be working well, so we get very few regressions in weekly releases. That is, all the official testing happens before the weekly release. All you get before the point release is a few days of community testing (that is, people reporting bugs when they find them).
An equally good approach would be to take any weekly build at a time that suits you; wait for a week while watching to see if any serious new bugs are reported; and if not, use that weekly.
I suppose the other thing you get from official releases is some release notes.
