Re: tabindex

On Fri, Aug 15, 2014 at 9:27 PM, Tab Atkins Jr. <jackalmage@gmail.com>
wrote:

> On Fri, Aug 15, 2014 at 8:44 PM, Rik Cabanier <cabanier@gmail.com> wrote:
> > On Fri, Aug 15, 2014 at 5:20 PM, Cameron McCormack <cam@mcc.id.au>
> wrote:
> >> Richard Schwerdtfeger wrote:
> >>> I just updated the spec. to place tabindex on all elements. This is
> >>> where browser implementations are going.
> >>
> >> Yes I think that's what we want, as HTML also allows tabindex="" to be
> set
> >> on any element (and the DOM property tabIndex is also available on any
> >> element).
> >>
> >> There are two possible futures:
> >>
> >> (a) we go ahead with my SVG DOM proposals, in which case all SVG
> elements
> >> will end up inheriting from the HTMLElement interface and gain the
> tabIndex
> >> IDL attribute from there
> >>
> >> (b) we don't go ahead with those proposals, in which case a tabIndex IDL
> >> attribute should live on SVGElement
> >>
> >> Either way, the content attribute tabindex="" is allowed on all
> elements.
> >
> > Even on elements that don't display such as defs, lineargradient, symbol,
> > etc?
>
> <link>, <meta>, etc all accept tabindex in HTML. They get skipped due
> to not being rendered, of course.


That makes sense and I think the spec should call that out.
Didn't we at some point discuss what should happen with tabindex in
symbols? It seems 'tree order' is not enough.