- From: Ian Hickson <ian@hixie.ch>
- Date: Thu, 6 Apr 2006 00:00:10 +0000 (UTC)
- To: Maciej Stachowiak <mjs@apple.com>
- Cc: Jonas Sicking <jonas@sicking.cc>, Web APIs WG <public-webapi@w3.org>
On Wed, 5 Apr 2006, Maciej Stachowiak wrote: > > I agree that it should be allowed. But I am not sure it should be > required, and I can't imagine how to test an implementation against such > a conformance requirement, as there is no exhaustive list of ways to get > a Document and nothing stops UAs from inventing new ones. No test suite is exhastive. You don't test that the "color" property works for every single color, even though the spec requires a particular behaviour for every single color. > To summarize options: > > 1) Require any non-presented DocumentWindow to have a null defaultView. > 2) Allow a non-presented DocumentWindow to have any AbstractView or null as > the defaultView (essentially, we disclaim stating a requirement on > non-presented documents). > 3) Forbid any non-presented Document to implement DocumentWindow (seems too > restrictive). I am strongly in favour of 3. If we don't do 3, we're going to have to require a whole heck more than 1 -- we're going to have to special case every single API that requires a rendering context. I am extremely against not specifying this (option 2), as I am against not specifying anything that authors could end up relying on. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Thursday, 6 April 2006 00:00:27 UTC