- From: jan-ivar <notifications@github.com>
- Date: Thu, 25 Aug 2016 22:30:04 -0700
- To: w3c/permissions <permissions@noreply.github.com>
- Message-ID: <w3c/permissions/issues/42/242634068@github.com>
@marcoscaceres sorry for the delay. See https://github.com/w3c/webrtc-pc/issues/275#issuecomment-134979978 for relevant WebIDL links and the discussion that follows, for why dictionaries are not polymorphic.
@jyasskin As written, [the WebIDL in the spec](https://w3c.github.io/permissions/#permissions-interface) is incorrect for all three methods. I think we need valid WebIDL here if we're going to have WebIDL at all, otherwise it is misleading.
Specifically: An implementation using a WebIDL compiler would not be able to - as the spec's normative prose dictates - "convert the underlying ECMAScript object to the permission descriptor type dictionary as described in [WEBIDL]", because the underlying ECMAScript object by definition is not available (*permissionDesc* is a secure copy of the base dictionary only), unless we were to do:
```js
[Exposed=(Window,Worker)]
interface Permissions {
Promise<object> query(object permissionDesc);
Promise<object> request(object permissionDesc);
Promise<object> revoke(object permissionDesc);
};
```
or something like it (`object` has terrible security characteristics). The spec needs to be clear here.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/w3c/permissions/issues/42#issuecomment-242634068
Received on Friday, 26 August 2016 05:31:04 UTC