Rule suggestion: prefer object spread to Object.assign

[danvk]

Now that TS supports the object-spread operator, the idea would be to suggest replacing code like this:

Object.assign({}, foo, bar);

with

{...foo, ...bar}

See https://github.com/airbnb/javascript#objects--rest-spread

[ajafff]

@danvk You can already do that with the ban rule and add an appropriate message. This doesn't have automatic fixing though.

On second thought using ban would also disallow cases where you really want to assign properties of one object to another without creating a new one.

Object.assign(obj1, obj2);

[danvk]

Yeah, this has slightly more complex semantics than "ban" would allow. If it's specifically mentioned in the AirBnb guide, I think it warrants a rule.

…

On Thu, Apr 20, 2017 at 5:36 PM Klaus Meinhardt ***@***.***> wrote: @danvk <https://github.com/danvk> You can already do that with the ban rule and add an appropriate message. This doesn't have automatic fixing though. On second thought using ban would also disallow cases where you really want to assign properties of one object to another without creating a new one. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2618 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAF__VR083RY0zlElN-Wpeo4qACh3J6Cks5rx8_kgaJpZM4NDmPs> .

[etsuo]

For what it's worth, an Object Spread does not appear to be the exact equivalent of Object.assign:

value = Object.assign(new Model(), value)

In this case, value instanceOf Model is true.

value = {...new Model(), ...value};

Here, value is an instance of Object.

This auto-fix just broke a bunch of my unit-tests... no biggie... just turned it off, but I figured I'd mention it in case someone thought the two were exactly equivalent.

[adidahiya]

that's a good point @etsuo... the rule should probably make an exception for when the first argument to Object.assign is a new SomeConstructor(...) expression because spread is not functionally equivalent

edit: filed #2824

[danvk]

Is the issue just with new? Or with any object with a prototype other than Object.prototype?

…

On Wed, May 24, 2017 at 7:26 PM Adi Dahiya ***@***.***> wrote: that's a good point @etsuo <https://github.com/etsuo>... the rule should probably make an exception for when the first argument to Object.assign is a new SomeConstructor(...) expression because spread is not functionally equivalent — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2618 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAF__QwX3DonKQotOqfSKKEh3kxZHPEfks5r9Ly9gaJpZM4NDmPs> .