[whatwg] Iframe dimensions

[Aryeh Gregor]

On Tue, Jul 6, 2010 at 4:41 PM, Ashley Sheridan
<ash at ashleysheridan.co.uk> wrote:
> Could you explicitly call the _self target in links in the frame? I wasn't sure if the target attribute was going or not, but I'd expect target="_self" to override the default seamless action.

It doesn't.  The iframe is considered part of the parent page for the
purposes of navigation, so _self refers to the parent page.  See the
full algorithm:

http://www.whatwg.org/specs/web-apps/current-work/multipage/history.html#navigate

Step 3 there does not take into consideration how the browsing context
to be navigated was chosen.  Anything that tries to navigate the
seamless iframe in any way will navigate the first non-seamless
ancestor instead.  That includes a regular old link, a link with
target="_self" explicit, navigation by JavaScript, and any other
possible way of navigating.  This is intended and correct for the
use-cases that seamless was designed for.

On Wed, Jul 7, 2010 at 10:15 AM, Markus Ernst <derernst at gmx.ch> wrote:
> This looks consistent to me. To clarify it, the first list item in the part
> on @seamless in 4.8.2 could then be extended somehow like:
>
> o The user agent must set the seamless browsing context flag to true for
> that browsing context. This will cause links to open in the parent browsing
> context, unless they contain a target attribute which explicitly specifies
> another link target.
>
> (I assume, "links" also refers to form actions here.)

That doesn't change the conformance requirements that "cause links to
open in the parent browsing context" links to.  The target of that
link is the detailed description of what that requirement means -- the
bullet point in the iframe section is just an informative (and vague,
and inaccurate) summary.  You would have to change the navigation
algorithm to figure out what triggered the navigation and so on.  But
I don't think this is desirable for the primary use-case of seamless.