|
| 1 | +## April 28 ([discuss](https://github.com/reactjs/core-notes/pull/10)) |
| 2 | + |
| 3 | +### Attendees |
| 4 | + |
| 5 | +* [Ben](https://twitter.com/soprano) (React) |
| 6 | +* [Christopher](https://twitter.com/vjeux) (React / React Native) |
| 7 | +* [Dan](https://twitter.com/dan_abramov) (React) |
| 8 | +* [Jim](http://github.com/jimfb) (React) |
| 9 | +* [Paul](https://twitter.com/zpao) (React) |
| 10 | +* [Sebastian](https://twitter.com/sebmarkbage) (React) |
| 11 | +* [Shayne](https://github.com/shayne) (React Native) |
| 12 | +* [Tom](https://twitter.com/tomocchino) (React) |
| 13 | + |
| 14 | +### Release Process and React Native |
| 15 | + |
| 16 | +#### We want to release to React Native more frequently |
| 17 | + |
| 18 | +* Can we push more frequently like we do with the Facebook website? |
| 19 | +* What if we ship React minor releases (or patch releases) every two weeks? |
| 20 | +* We should be using a public stable release of React (either a minor or patch) on the website. |
| 21 | +* This would coincide with the React Native release process. |
| 22 | + |
| 23 | +#### React Native users have way more third-party dependencies |
| 24 | + |
| 25 | +* These won’t be compatible with “pre-release” or bleeding edge versions of React. |
| 26 | +* The only people that use RC are folks that have something that was fixed (usually by them) in the RC. |
| 27 | +* Otherwise, upgrading so frequently is incredibly draining. |
| 28 | + |
| 29 | +#### Should we have feature flags, like Ember does? |
| 30 | + |
| 31 | +* Maybe this would help catch regressions earlier. |
| 32 | +* On the other hand, this causes an explosion of possible build configurations. |
| 33 | + |
| 34 | +#### If we’re releasing more frequently we need to be testing React Native more |
| 35 | + |
| 36 | +* We need to be testing the React release in React Native. |
| 37 | +* Everyone on the team needs to have a working React Native setup on their machines. |
| 38 | + |
| 39 | +#### Should we publish nightly releases? |
| 40 | + |
| 41 | +* Can React Native lock itself into whatever nightly it wants? |
| 42 | +* But we don’t want React Native to be pegged to a specific commit. |
| 43 | +* This is pretty much what we started doing in 15.x with `alpha` releases for patches. |
| 44 | + |
| 45 | +#### Current plan |
| 46 | + |
| 47 | +* Going to publish `alpha`s, which can be released as frequently as we want, even like 4 a day. |
| 48 | +* Once something is stable and we’re confident it's working in React Native, we’ll cut an actual release. |
| 49 | +* React Native and Relay always reference a stable release in their dependencies. |
| 50 | + |
| 51 | +#### What will our messaging be about alphas? |
| 52 | + |
| 53 | +* Alpha = use at your own risk. |
| 54 | +* If you depend on an alpha due to an urgent fix you need, pin it to a specific version. |
| 55 | +* Guidance should be “never publish anything to npm with an alpha dependency”. |
| 56 | + |
| 57 | +#### Open question: where does the breaking change development happen? |
| 58 | + |
| 59 | +* Are we on `15.1.0-dev` or `16.0.0-dev`? |
| 60 | +* Do we merge breaking changes to master? |
| 61 | +* Currently we cherry-pick commits to `15-stable` but this is frustrating because 99% changes are not breaking. |
| 62 | +* Need to discuss more, no conclusion yet. |
| 63 | + |
| 64 | +### Experimenting with `StyleSheet.create()` on the web |
| 65 | + |
| 66 | +* [Christopher](http://twitter.com/vjeux) continues experimenting with `StyleSheet.create()` on the Facebook website codebase. |
| 67 | +* He made a proof of concept compiling these calls into CSS sheets compatible with Facebook internal asset pipeline. |
| 68 | +* Maintaining correct specifity with `StyleSheet.create()` is not an easy problem, but it may be solved by throwing errors early during development when there is a potential specificity issue, and providing a clear way for the developer to work around it. |
| 69 | +* Compiling `StyleSheet.create()` to emitting real CSS files through our internal asset pipeline lets us move a bit closer to experimenting with real inline styles without having to bet on them right now. |
| 70 | +* Nothing is certain, and Christopher will continue experimenting this month. |
| 71 | +* If it doesn’t turn out to be valuable, this will not make it into React! Don’t get too excited. |
| 72 | + |
| 73 | +------------ |
| 74 | + |
| 75 | +Please feel free to discuss these notes in the [corresponding pull request](https://github.com/reactjs/core-notes/pull/10). |
0 commit comments