Re: Should restrictions on attribute types look at flattened member types of unions? from Cameron McCormack on 2012-06-21 (public-script-coord@w3.org from April to June 2012)

From: Cameron McCormack <cam@mcc.id.au>
Date: Thu, 21 Jun 2012 12:13:41 +1000
To: Boris Zbarsky <bzbarsky@MIT.EDU>
CC: public-script-coord@w3.org
Message-ID: <4FE28355.8010100@mcc.id.au>

Boris Zbarsky:
> This is not valid WebIDL:
>
>    attribute sequence<long> foo;
>
> but this seems to be valid at the moment:
>
>    attribute (sequence<long> or DOMString) foo;
>
> why should we allow the latter if we don't allow the former?  Seems like
> it would be more consistent to forbid the latter...

I think the latter is already forbidden.  In #idl-attributes it says:

   The type of the attribute MUST NOT be a sequence type or nullable
   sequence type, and it MUST NOT be a union type if one of its member
   types (or one of its member types’ member types, and so on) is a
   sequence type or nullable sequence type.

Received on Thursday, 21 June 2012 02:14:13 UTC