Should we publish a version 2? about deepmerge

[TehShrike]

Yeah, that intuitively seems like a change that would bite some people.

I'd be more worried about it, except nobody seems to have ever been bothered by the root object always being freshly created every time, which seems like the case that would matter most. ¯\_(ツ)_/¯

[TehShrike]

If there aren't any reviews, I'll merge/release next week.

[TehShrike]

Version 2.0.0 has been published! Open new issues for anything you notice that seems off.

[macdja38]

So long as we clearly document the change at the top of the npm page + provide an example of how users can easily switch back to the old behavior I don't see too much of a problem. obviously cloning would be a bad way of dealing with certain things because people may not expect it... for example a user using a object based module system probably does not want their modules cloned as calling methods from the cloned vs uncloned objects would have different internal state...

[macdja38]

I could see a lot more people using the library on objects that contain classes vs on classes themselves, but I definitely agree that cloning everything is probably a better way to avoid unexpected behavior in the majority of cases. Along with clear documentation on how to reverse the behavior I see no problem with it.

[z0al]

+1
But, we should of course document the changes

[NoxWings]

I think clone should default to true

Exactly this

[nrkn]

Agree strongly with all of your suggestions, they would make it much better

[TehShrike]

I've opened a pull request - reviews of #77 would be appreciated!

[nrkn]

I'll have the opportunity to put it through its paces pretty intensively soon, merging multiple nodes with complex models from a graph