- From: Tobie Langel <notifications@github.com>
- Date: Thu, 31 Aug 2017 15:42:16 +0000 (UTC)
- To: heycam/webidl <webidl@noreply.github.com>
- Cc: Subscribed <subscribed@noreply.github.com>
- Message-ID: <heycam/webidl/pull/423/c326336051@github.com>
> It seems ok to me. Forcing all interfaces to always have extended attributes is a bit weird (leading to the question of why that required bit is not using dedicated syntax), but the general idea does have the benefit of forcing people to think about where their new API should be available. Yeah, I agree on both counts. What's your suggestion, should we stay with extended attributed or move to dedicated syntax? > For example, "The exposure set of the interface is the exposure set of the interface that implements it" is nonsense, because "the interface that implements it" is multiple interfaces with possibly-different exposure sets. That phrasing indeed doesn't work. My intent was to say that if [Exposed] was specified on the mixin itself or on one or more of its members, the mixin members would be exposed in the intersection of the identifiers specified in [Exposed] and the exposure set of the interface implementing it. Otherwise, the mixing members would just be exposed in the globals of the exposure set of the interface implementing it. > And again, there are non-mixin [NoInterfaceObject] interfaces in the platform... Indeed. We made some incorrect assumptions here. Maybe this will be easier to spec once we have mixins? -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/heycam/webidl/pull/423#issuecomment-326336051
Received on Thursday, 31 August 2017 15:42:39 UTC