Duplicate "draggableCursor" property in dist/index.js

[mbpreble]

My team was attempting to use the "Circle" component provided by this library but noticed that while it is documented it does not actually exist in the latest version published to NPM (this is 2.0.2, published in 2018).

What we did was install the dependency directly from GitHub, specifying the commit which was currently the tip of master, i.e. in package.json:

"google-maps-react": "github:fullstackreact/google-maps-react#c70a04f668b644a126071c73b392b6094a430a2a"

This sort of approach was has been discussed as a workaround by other users who were trying to consume the "Circle" component: #300 (comment)

We later ran into a severe issue - our application would no longer function in IE11. This also appears to have been identified as an issue, with an outstanding PR which fixed it: #353

The bug still seems to exist in the "dist" directory, which is what we are pulling in by installing from GitHub: https://github.com/fullstackreact/google-maps-react/blob/master/dist/index.js#L270

There's no obvious reason why this bug would still be shipped under "dist", my best guess is that the contents of "dist" are stale and not maintained consistently by a well-defined process. It’s unclear whether the bug was ever committed as source to the repo.

Really this is a report of a couple different problems:

1. The version published to NPM is well behind master, rapidly approaching two years behind
2. The built version of the library committed in this repo under the "dist" directory has an unclear relationship with the source being shipped. It seems to include bugs that don’t exist in source.

It would be most straightforward for us to move on if the last two years of changes to this project were released as a new version in NPM. Failing, that updating the committed files in dist to reflect the correct build output would let us move forward.

Otherwise our clearest path forward is to roll our own "Circle" component and plug it into 2.0.2, but that seems wasteful given there is already an implementation here.

[mbpreble]

Linked PR follows one of the suggestions here, which is re-running the build and just committing the output. This is obviously not an ideal solution for a build system.

If merged this would unblock my team in a straightforward way.