Hi! Topics of this message: 1- Semi-Automatic Merging Period is over 2- Branch 16.x is closed 3- Branching of 18.x will happen mid-November 2017 4- 18.0 will be released around December 10th, 2017 5- What to expect in coming weeks 6- What I ask of you 1- Semi-Automatic Merging Period is over Jonny announced it in another thread, but repeating here to make sure everyone sees it. Please see: https://dev.tiki.org/Where-to-commit 2- Branch 16.x is closed This marks the end of the road for branch 16.x Thank you to the team who released 16.3 recently (the last of the series). The team also packaged 17.1, 15.5 and 12.12 Formal announcements will appear on tiki.org's front page after a bit more testing. Now, looking forward, so everyone knows what to expect and plans accordingly: 3- Branching of 18.x will happen mid-November 2017 I was hoping we could branch at https://tiki.org/TikiFestJapan2017 but that would have given us less than 3 weeks. Much too short. So realistically, we'll branch mid-November. That is less than 2 months away. Still quite short. 4- 18.0 will be released around December 10th, 2017 We definitely want to release 18.0 before the holiday season. Some new people will take on roles in this release to reduce the pressure on Jonny and Bernard (who have been doing so so much for a long long time). More about this in another message later. 5- What to expect in coming weeks A) A decision will be made (either way) on code reformatting. B) Trunk is pretty close to 17.x today, and thus, it's pretty stable. It will go through several weeks of major work and some turbulence, before it stabilizes again. There are at least 20 enhancements and new features that have been waiting for the semi-automatic merging period to end. Beyond the content of experimental branches, there are others in Git branches. But don't worry: no major revamps like Tracker Revamp or converting to Bootstrap. We are keeping the more risky things for Tiki 19, (as it's a post-LTS). Some things to look forward to: https://dev.tiki.org/Tiki18 https://doc.tiki.org/Tiki18 It's hard to measure, but I think Tiki 18 will be the release in which there will be the most brain-hours invested in, ever! It's mostly either: * Refinements of already mature features (ex.: profiles dry-run) * Self-contained additions (Bootstrap “Style Guide”, Media-Alchemyst, etc.) * Code cleanup and reducing technical debt An example of technical debt reduction: The move to Composer / Packagist started around Tiki 11, and has been very beneficial. However, some of the less active libs have been stuck in the "old way". Now, with this latest push, we'll be using Composer / Packagist for almost all external dependencies. Furthermore, Tiki 18.x will be using the latest stable versions for all our major components (Zend Framework, jQuery, Bootstrap, Smarty, etc.). Failing to manage component lifecycle is a risk that bigger and more mature projects incur. Thank you to all who helped with this, in particular Ricardo who recently did a lot of the difficult / crufty ones. Why this is so important: https://blog.versioneye.com/2014/02/18/why-your-software-project-will-slowly-die-without-continuous-updating/ C) A proposal for a continuous integration / testing tool is the works. We will share with the dev list as soon as it's ready. 6 What I ask of you A) Get involved and do stuff B) Make sure your work is stable in trunk by mid-November 2017 C) Use the tiki.org pre-dogfood servers and other feature test servers D) Use trunk or at least a pre-dogfood server for your own projects A) Get involved and do stuff We need more different people doing more things. The more people are doing things, and the more people are reliable, the easier it is for us all. If you want to help, but don't know where / how, contact Bernard or I off-list. B) Make sure your work is stable in trunk by mid-November 2017 I want to branch late, on a stable trunk. I repeat: I want to branch late, on a stable trunk. In recent times, the branching period before the x.0 has been way too long. This is a waste of developer energy. If the .0 happens too long after branching, it means trunk wasn't stable enough. In theory, trunk should be releasable at any time. In practice, it's nice to have a predictable period where we can work in a somewhat messy way, as long as we all know when to converge to stability. More info: https://dev.tiki.org/When-to-branch Let's get trunk stable in time for mid-November 2017, then branch, and proceed to a rapid migration of all 17.x *.tiki.org sites to 18.x If trunk is stable before the branch, this should take about 3 weeks. So do your biggest and riskiest changes sooner rather than later, so things stabilize in time. If you have a riskier change that may not be stable enough in time, please use an experimental branch. And we can look at it together and decide to merge in 18.x or wait for 19.x C) Use the tiki.org pre-dogfood servers and other feature test servers * to thoroughly test all *.tiki.org that are running 17.x * to make sure trunk is stable, before we branch 18.x Fixing a known bug after the branch is just more work and risk. Fix it before the branch, confirm the fix on *.tiki.org pre-dogfood servers and never risk real data. * to test new features so we get the quality high in time for the branch * to document the new features https://tiki.org/Pre-Dogfood-Server For some features, we also have some testing servers which are frequently updated to latest trunk: http://testing.mpdf.suite.wiki/ (Just ask me access off list) I intend to add one more test site for the new Style Guide tool. D) Use trunk or at least a pre-dogfood server for your own projects Use trunk as much as you can. If you have a production site, and can't risk it, set up your own pre-dogfood server via TRIM: https://doc.tiki.org/TRIM#make_clone_and_upgrade Ask regular users of the site to test it, and report issues. Again, finding and fixing a bug on a pre-dogfood server is way way more efficient. When we all do this, we all save a lot of time. With this plan, we can release an awesome 18.0, and get back on schedule and have a more normal timeline for Tiki 19 (for April 2018). Since it's a post-LTS, it will be the time to decide if we upgrade key infrastructure (Ex.: Boostrap 4) and to increase system requirements (ex.: PHP 7.1 as a minimum) Best regards, -- Marc Laporte http://WikiSuite.org http://PluginProblems.com http://Avan.Tech |