1435869 - [META] Decorators proposal

Status: NEW

(Core :: JavaScript Engine, enhancement, P3)

Description

[Till Schneidereit [:till]]

While the Decorators proposal is only at stage 2, we should start getting implementation experience or at least an analysis of any potential performance concerns or other implementation hardships soon. There might be incidental implementation concerns that we could avoid by suggesting specific changes to the proposal but would otherwise have to live with if we wait too long.

Needinfo for Jason and André for the analysis part as discussed after the last tc39 meeting.

Comment 1

[Till Schneidereit [:till]]

André, according to my notes you wanted to specifically look into how decorators interact with static class fields and their access to the class scope.

Comment 2

[André Bargull [:anba]]

https://tc39.github.io/proposal-decorators/ builds upon https://tc39.github.io/proposal-private-methods/ and also includes https://tc39.github.io/proposal-class-fields/ and https://tc39.github.io/proposal-static-class-features/, so there's a bit of reading to do. Hopefully all four proposals were kept in sync, so they don't contradict each other (too much). :-)

But before diving into the decorators proposal, I should probably finish this other larger project (*cough* irregexp update *cough*).

Comment 3

[Till Schneidereit [:till]]

(In reply to André Bargull [:anba] from comment #2)
> https://tc39.github.io/proposal-decorators/ builds upon
> https://tc39.github.io/proposal-private-methods/ and also includes
> https://tc39.github.io/proposal-class-fields/ and
> https://tc39.github.io/proposal-static-class-features/, so there's a bit of
> reading to do. Hopefully all four proposals were kept in sync, so they don't
> contradict each other (too much). :-)

Dan, can you comment on whether this assumption holds? I'm assuming largely yes, but would be good to be sure.

Comment 4

[Daniel Ehrenberg]

I've done my best to keep the proposals in sync, but bug reports are always welcome. Among these, I'd recommend starting with private methods and class fields, which are Stage 3 and more stable. Static class features and decorators are likely to have some tweaks to them on their way to Stage 3.

If any SpiderMonkey engineers are interested, I'd like to go over decorator semantics in detail with JS implementers in a call, in preparation for proposing to Stage 3 at TC39. Who should I send an invitation to?

Comment 6

[Yulia Startsev [:yulia]]

This is conditionally advanced. The primary issue of metadata has been fixed, we are waiting on isPrivate (https://github.com/tc39/proposal-decorators/issues/417) and a few other changes.

Comment 7

[trusktr]

Just curious, has the team started implementation of stage 3 decorators yet? It is going to be incredibly awesome.

Comment 8

[Dan Minor [:dminor]]

Yes, the implementation is being tracked by Bug 1781212.