Re-work <dialog>#show*() focus algorithm

Re-work <dialog>#show*() focus algorithm

See the history of this change here:
https://github.com/whatwg/html/issues/1929

R=kochi@chromium.org
BUG=

[esprehn, 2016-12-01 00:39:57]

esprehn@chromium.org changed reviewers:
+ esprehn@chromium.org

[esprehn, 2016-12-01 00:39:58]

This seems really reasonable to me. I do think we probably want to cross shadow
boundaries though? That's a long standing super annoying bug, an <input> inside
a shadow root inside a <dialog> (ex. inside <paper-input>) doesn't work.

https://codereview.chromium.org/2532983004/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/html/HTMLDialogElement.cpp (right):

https://codereview.chromium.org/2532983004/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/html/HTMLDialogElement.cpp:48: ?
NodeTraversal::nextSkippingChildren(*node, dialog)
I think this should probably cross shadow boundaries?

[Dan Beam, 2016-12-01 01:02:21]

Description was changed from

==========
Re-work <dialog>#show*() focus algorithm

R=kochi@chromium.org
BUG=
==========

to

==========
Re-work <dialog>#show*() focus algorithm

R=kochi@chromium.org
BUG=670130
==========

[Dan Beam, 2016-12-01 01:36:32]

Description was changed from

==========
Re-work <dialog>#show*() focus algorithm

R=kochi@chromium.org
BUG=670130
==========

to

==========
Re-work <dialog>#show*() focus algorithm

R=kochi@chromium.org
BUG=
==========

[Dan Beam, 2016-12-01 01:37:27]

adding a layout test now

https://codereview.chromium.org/2532983004/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/html/HTMLDialogElement.cpp (right):

https://codereview.chromium.org/2532983004/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/html/HTMLDialogElement.cpp:48: ?
NodeTraversal::nextSkippingChildren(*node, dialog)
On 2016/12/01 00:39:58, esprehn wrote:
> I think this should probably cross shadow boundaries?

I agree but we talked over chat and you were OK with separating these fixes
(i.e. letting me do a followup CL to address looking through shadow DOM)

[Dan Beam, 2016-12-01 01:37:45]

The CQ bit was checked by dbeam@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-12-01 01:38:07]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-12-01 02:48:10]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-12-01 02:48:11]

Dry run: Try jobs failed on following builders:
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[Dan Beam, 2016-12-02 00:46:55]

Description was changed from

==========
Re-work <dialog>#show*() focus algorithm

R=kochi@chromium.org
BUG=
==========

to

==========
Re-work <dialog>#show*() focus algorithm

See the history of this change here:
https://github.com/whatwg/html/issues/1929

R=kochi@chromium.org
BUG=
==========

[falken, 2016-12-02 03:37:20]

falken@chromium.org changed reviewers:
+ falken@chromium.org

[falken, 2016-12-02 03:37:21]

I made some comments on
https://bugs.chromium.org/p/chromium/issues/detail?id=670130

At a glance it seems reasonable for dialog to be focusable but we should
probably review https://www.w3.org/Bugs/Public/show_bug.cgi?id=23366 for reasons
Hixie decided not to spec it that way.

Implementation-wise I'd recommend running this through JAWS/NVDA and VoiceOver
and also playing with tab behavior to see if anything unexpected changes. Here's
a basic demo site:
http://mattto.github.io/dialog/a11y-test.html

[Dan Beam, 2016-12-15 04:34:38]

btw, for those not watching https://github.com/whatwg/html/issues/1929

I'm no longer convinced that we should be changing the default behavior of
<dialog>.  With a simple layout and average usage, this default focus behavior
is probably fine as it is.  (found a bunch more context to support this)

So instead, we're now talking about adding a different form of usage for showing
without the default behavior (through a parameter).  I think this may be a
better way forward.

Closing this for now, as I haven't heard anybody admit that we want to change
the default behavior in the HTML spec.

[Dan Beam, 2017-01-12 01:08:25]

On 2016/12/15 04:34:38, Dan Beam wrote:
> btw, for those not watching https://github.com/whatwg/html/issues/1929
> 
> I'm no longer convinced that we should be changing the default behavior of
> <dialog>.  With a simple layout and average usage, this default focus behavior
> is probably fine as it is.  (found a bunch more context to support this)

* given that there's a way to turn this off

> 
> So instead, we're now talking about adding a different form of usage for
showing
> without the default behavior (through a parameter).  I think this may be a
> better way forward.
> 
> Closing this for now, as I haven't heard anybody admit that we want to change
> the default behavior in the HTML spec.