Hi all,
There are currently some issues in the 1.8.1 tree regarding filters used in when and with_items loops, that we didn’t have integration tests for – OH NO! PANIC! (Many of you might not be using these features, but some are, and we care about that alot).
Actually, don’t panic, we’re on it - 1.8.2 will be out this week including this fix as well as Windows modules packaged in the distribution. There might be a slight refinement to some undefined variable error handling to follow - TBD - but that part is optionalish.
Future work plans over the next month or so will work like this. As I’ve mentioned before, we have a ‘v2/’ refactoring tree that James is working on very actively. We’re going to press for this tree to quickly become the basis of 1.8. Nothing will change with regard to support, and we’ll enlist everyone to help test in a later date before this tree replaces “/” on devel. It’s not ready for contribution just yet.
While we do this, the rest of the team is going to spend a lot of time on ansible-modules-core and ansible-modules-extras, because we’ve been not spending as much time on these lately. Of course, this is super important to us - but we’ve been spending time on making the core really bug-free in LOTS of ways and attending to that.
Moving forward with the v2/ tree will make the core mantaininable and increase our velocity with regard to how much time we can spend in module land, and also be able to more rapidly bring new language features to the table - including things like block support (try/except/recovery things). We’re already making some nice progress in that area.
So, anyway, expect 1.8.2 soon - as well as a flurry of module activity. Eventually this module activity will probably produce something akin to a 1.8.5 bringing these new modules features in to the fold (as usual, stuff like EC2 is towards the top of the list, but we want to give good attention to everything).
Then, for the 1.9 release, we’ll eventually flip over to the ‘v2/’ code subtree (I’ve mentioned v2 is NOT about a version number), and be ready to go forward with a more understandable central core, recovered from years of absolutely amazing contribution rates.
(And the battery of integration tests help make sure things are still fully compatible - along with a lot of extra unit test coverage).
Anyway, wanted to give an update for those interested in module (and core) progress and how we are addressing this as we grow.
Running a project with 920+ contributors (and that’s likely more now, now that the repos are split) has been a super interesting experience and we thank you all for bearing with us and helping along the way!
Let me know if there are specific questions about the above.
Thanks!