feat(babel-preset-app): pass full config to @babel/preset-env

[lucaswerkmeister]

Pass not just the useBuiltIns and corejs options, but the whole envOptions object, into the @babel/preset-env preset that is used to transform @babel/runtime, just like for the @babel/preset-env that is used for the application source code. This allows users to also specify other options, such as exclude and polyfills, and have them apply here too.

In particular, this can be used to exclude the Promise polyfill, e. g. if Promise is already polyfilled in some other way. Previously,

exclude: ['es.promise'],
polyfills: ['es.array.iterator', 'es.object.assign'],

could be used to configure the preset for the application source code, but the babel runtime would be transformed without those options, and so es.promise would still end up being included.

Closes #5208

What kind of change does this PR introduce? (check at least one)

• Bugfix
• Feature
• Code style update
• Refactor
• Docs
• Underlying tools
• Other, please describe:

Does this PR introduce a breaking change? (check one)

• Yes
• No (I think)

Other information:

[lucaswerkmeister]

(On second thought, polyfills: ['es.object.assign'] might be better than polyfills: ['es.array.iterator', 'es.object.assign'], since according to the corresponding comment, es.array.iterator is only needed to make es.promise work on IE. But I’m not sure.)

[lucaswerkmeister]

If anyone is wondering where this PR is coming from or interested in more context, feel free to look at this task and specifically this comment in our downstream bug tracker :)

[haoqunjiang]

Thanks!
Actually I was thinking about dropping the transpilation of @babel/runtime entirely considering babel/babel#7597 has been fixed.

But I'm not sure if that's safe to do…
Maybe this PR is a better option before the runtime is sure to be bug-free.

[lucaswerkmeister]

@micgro42 I think you were also wondering whether the transpilation was still necessary, did you come to any conclusion?

[micgro42]

"Conclusion"? No. My approach would likely just be to remove it, let the tests run, and test it manually on IE11 and Safari. If all that checks out then I'd probably call it a day.

[haoqunjiang]

After investigation, there's still one edge case: babel/babel#9277

So I think I'll go for this PR instead of removing transpilation at the moment.

[lucaswerkmeister]

Cool, thanks :)