One issue though is the HTML5 doc uses some syntax that relies on a Rubyesque method_missing capability that just isn't present in many environments, including, notably, JavaScript.
E.g. the HTML 5 doc has this example:<label> <input type="checkbox" onchange="sessionStorage.insurance = checked"> I want insurance on this trip. </label>
Now, we can handle this no problem:
Widget.preferences.getItem("insurance")
But:
Widget.preferences.insurance
... poses a bit more of a challenge!
Thoughts?
S
On 16 Mar 2009, at 11:57, Marcos Caceres wrote:
2009/3/9 Arve Bersvendsen <ar...@opera.com>:On Mon, 09 Mar 2009 12:20:25 +0100, Arthur Barstow <art.bars...@nokia.com > wrote:Note that there is a difference between localStorage and Storage. I am assuming that an implementation will place a (different) instance of a Storage interface on the widget.preferences object. This object would not be bound to the same storage instance as window.localStorage, and instead have its origin set to that of the widget (whatever URI scheme or origin we end up with once those issues are resolved), translated: The widget interface would not end up using the same storage area. This already seems to be permitted by the current spec, still residing in HTML5 [1], where there are two separate Storage instances in window.localStorage and window.sessionStorageThis distinction is useful for widgets, as you could basically set one preference that only affects a single widget instance (sessionStorage), while other preferences affect all widgets of a particular class (localStorage). -- Marcos Caceres http://datadriven.com.au
smime.p7s
Description: S/MIME cryptographic signature
