Upgrade to Babel 6

[doug-wade]

This allows us to build lazyload with babel 6

[doug-wade]

As far as I can tell, the default key from this file (export default LazyLoad) is being overwritten by the default key being exported from './decorator'. By importing from react-lazyload/decorator directly, we don't have contention for the key.

My theory is that babel makes exports a single object that is returned, such that if you then re-export default something export default, you overwrite the default key of the returned object.

foo.js

export default () => console.log('foo');

bar.js

export default () => console.log('bar');
// would return { default: () => console.log('bar') }

export foo from foo;
// now returns { default: () => console.log('foo') }

[jasonslyvia]

How about

// src/index.js
import decorator from './decorator';
export const lazyload = decorator;

API changes are always sore spot for module consumers.

[coveralls]

Coverage decreased (-0.1%) to 94.538% when pulling 483f65f on redfin:upgrade-babel into 4923277 on jasonslyvia:master.

[doug-wade]

This is what is recommended by Babel

[doug-wade]

I had to introduce a breaking api change, changing the import for the decorator from import {lazyloader} from 'react-lazyloader' => import {decorator} as lazyloader from 'react-lazyloader/decorator'. I should test with a reduced use case against Babel 6, check the spec, and maybe submit a pr with a broken test to babel.

[coveralls]

Coverage decreased (-0.1%) to 94.538% when pulling 587fc85 on redfin:upgrade-babel into 4923277 on jasonslyvia:master.

[jasonslyvia]

Wow that's a lot of stuff, I need some time to dive into every piece of your comment.

[jasonslyvia]

Can you squash your commit and leave out lib/ changes for better diff?

[doug-wade]

@jasonslyvia Some great points; sorry for the long turnaround on feedback. I think this is much more reasonable pr. I'm also sorry for shoving a bunch of stuff into one pr; we did a lot of iteration before we got this working internally, and I didn't want to open a PR until it was clear we'd actually be relying on it, and our changes would work.

[coveralls]

Coverage decreased (-0.02%) to 94.8% when pulling 6fa2388 on redfin:upgrade-babel into d7e67f5 on jasonslyvia:master.

[jasonslyvia]

Why use transform-react-jsx plugin instead of react preset?

[coveralls]

Coverage decreased (-0.02%) to 94.8% when pulling 0155501 on redfin:upgrade-babel into 91dea19 on jasonslyvia:master.

[coveralls]

Coverage decreased (-0.02%) to 94.8% when pulling 0155501 on redfin:upgrade-babel into 91dea19 on jasonslyvia:master.

[jasonslyvia]

I think we're good to go as per this PR, is there further modification you're going to make, or should I merge it now?

[doug-wade]

I think its ready to merge! Thanks for being so patient.