Evaluate the introduction of a radashi.js dependency

[dgdavid]

First of all, let me clarify that I'm aware of the disadvantages of excessive reliance on dependencies and did not forget the https://en.wikipedia.org/wiki/Npm_left-pad_incident

Moreover, when it comes to dependencies, I always think of the picture below.

However, I realized recently that we could rely on https://radashi.js.org/ to stop writing the utilities ourselves. It already provides our noop, isEmpty (recently added by myself), isObject, isObjectEmpty through its isEmpty, partition (which they call fork), etc.

We can benefit from its use by writing and maintaining less code, documentation, and tests. Rather than reinventing the wheel, we could focus on Agama features instead :P

Please, note that it is built for TypeScript and fully tree-shakeable.

[imobachgs]

In my opinion, it makes sense. And if something goes wrong, we could replace the helpers with our own implementations again. But I understand that this library is properly tested and it will save us of writing trivial code (with its share of bugs).

[dgdavid]

And if something goes wrong, we could replace the helpers with our own implementations again.

We could even keeps a copy of the library in a automatically syncronized repo, just in case.

But I understand that this library is properly tested

Yes it is.

and it will save us of writing trivial code (with its share of bugs).

And even allow us to contribute with something missing, if needed.

[ancorgs]

I do not oppose.

[dgdavid]

Status update: I've been working a bit on this. I'll put in on hold for a couple of weeks, at least until v 12.2 is published.

[dgdavid]

For the record, today I've created a PR for going ahead with this if finally accepted. #2412

[dgdavid]

And now merged to master branch.