2015 04 17

Attendees:

• @zpao
• @sebmarkbage
• @spicyj
• @jsfb
• @fkling
• @pieterv
• @jbonta

Agenda

• Week In Review
• Mixins & ES6 classes
• Flow vs PropTypes

Week in Review / Project Updates:

Misc

• react native v0.4 went out today

@zpao

• Shipping 0.13.2 imminently

@jsfb

• continued working on context related changes

@fkling

• react docgen - OVER 1000 installs

@sebmarkbage

• working with jwalke & julien (of Hack) on figuring out next language/rendering tech. Nothing material, still a long way away (years) but some interesting ideas that could influence how we think about React.
• Talked with @vjeux about some promising ideas for animation.

@spicyj

• Working on error boundaries
• Integrating Pull Request review labeling bot that @vjeux wrote

Mixins & ES6 classes:

• There were some discussions about mixins @ TC-39
• We still don’t want our own way of doing them for ES6 classes
• One idea as a potential future forward migration path is to start forcing React.createMixin
  • We still don't really know what official language support would look like, so probably too early to do that.
• Can we get rid of PureRenderMixin?
  • It's getting heavily used in OSS but doesn't work with ES6 classes.
  • It's not quite the right way to actually think about it, the owner is really the pure component (always sending down the same data)
  • Decorators are one way we might deprecate this - other projects are making use of decorators but we don't have a spec for this yet so a lot can change

Flow vs PropTypes

The general idea being discussed was if it would ever be possible to drop propTypes completely from React and rely only on Flow. It might be possible eventually but they are pretty different right now (run time vs static, error levels). Based on some product teams' experience, Flow seems like the better choice in the long term but right now it's a mix of Flow and propTypes, each helping differently.

The discussion here was mostly about uses at FB - we know not everybody will buy into Flow in open source.