Re: [webidl] Specify [LegacyUnenumerableNamedProperties] (fixes #82) (#91)

Maybe it's not being clear; let me give a concrete example of how this would play out in practice.

Right now, in Gecko, a class with a named getter implements a function that enables bindings to ask whether a given property name is enumerable.  This function implements the prose definition of which properties are unenumerable.

Say the bindings in Gecko got updated to no longer call this function and instead just use the `[LegacyUnenumerableNamedProperties]` bit from IDL.  Now someone comes along and implements a currently-specced interface that does the prose thing.  They copy the IDL from the spec, implement this "is this name enumerable?" function to return false for all names, and think they're OK.  But of course they're not.  Of course arguably they should have tests that would catch the problem, but I would really rather not rely on that.  Getting the other specs update here shouldn't be terribly complicated; we should just do it.

---
Reply to this email directly or view it on GitHub:
https://github.com/heycam/webidl/pull/91#issuecomment-184703468