Proposed change for Focus Navigation #381

TakayoshiKochi · 2016-02-12T11:03:33Z

As discussed in
#375

This adds a clarification about focus navigation order for
nodes distributed under a slot element.

Also added handling tabindex="-1" case.

As discussed in WICG#375 This adds a clarification about focus navigation order for nodes distributed under a slot element. Also added handling tabindex="-1" case.

TakayoshiKochi · 2016-02-12T11:03:44Z

@hayatoito PTAL

hayatoito · 2016-02-12T11:28:17Z

spec/shadow/index.html

@@ -1139,6 +1141,27 @@
          </li>
        </ol>

+        <p>Likewise, the <a>sequential focus navigation order</a> for <a>distributed nodes</a>


Wrong... This should not be distributed nodes. This should be assigned nodes.

In addition to that, you should consider the followings:

Do not forget to include descendants of assigned nodes into this focus navigation group.

Consider the fallback contents of slots, in addition to assigned nodes.

When you try to consider 1 and 2, do not forget the case where a slot is a descendant of another slot.

The traversal order is not defined for sequential focus navigation order for distributed nodes. Define the traversal order of this focus navigation group.

Exclude an assigned node (and more) from its root node's focus navigation group. One node should not join more than one focus navigation group.

Mention how tabindex values are scoped at each focus navigation order.

Thanks.

Probably we want to start using 'control group' (defined in HTML spec) for the scope.
I am working on rewriting with it.

…nt is no longer dispatched

Addresses comments from hayato.

TakayoshiKochi · 2016-02-18T11:17:40Z

@hayatoito PTAL.

I tried to get this upstreamable to HTML spec, section 6.4,
but it is hard to do in a single step, as the HTML spec describes the
navigation order as an algorithm to iteratively find the next focusable node,
while the Shadow DOM spec (before my modification) assumes there is a
(somewhat) pre-computed focus navigation order for each scope and
describes how they are merged together.

So this is my attempt to include the #375 as exactly as possible by
following the current way of describing focus navigation in Shadow DOM.

At this point it seems to me to patch the upstream spec requires lot more work.

hayatoito · 2016-02-18T13:45:51Z

spec/shadow/index.html


        <p>If a <a>node</a> doesn’t <a data-lt="participates">participate</a> in the <a>document flat tree</a>, the <a>node</a> <strong>must</strong> be skipped from the <a>sequential focus navigation</a>.</p>

-        <p>The <a>sequential focus navigation order</a> for a given <a>shadow tree</a> <var>A</var> <strong>must</strong> be inserted into the <a>sequential focus navigation order</a> for the <a>parent tree</a> <var>B</var> as follows:</p>
+        <p>Each <a>shadow root</a> and <a>slot</a> element is a <a>control group owner</a> and


A control group is more upper layer concept, isn't it?

I mean:

There is 1 : N relationship between [a control group] and [focus navigation order in each scope].

There is 1 : 1 relationship between [a control group] and [global focus navigation order], which is the result of merging multiple [focus navigation order in each scope].

My understanding is that frames own its own control group (e.g. <iframe>) and its own sequence, but from user's standpoint, tab navigation sequence look global. Thus sequences for each group is merged.

hayatoito · 2016-02-18T13:50:44Z

I think you will not get a good definition by the current approach. I recommend you to have a more formal approach:

e.g.

algorithm:

Input: a document composed tree
Output: set of owners, where each owner owns a group

algorithm:

Input: node, NODE, in a document composed tree
Output: a group which NODE belongs to

Thus, we can define the disjoint sets clearly. Then,

algorithm:

Input: node A and node B where A nad B joins the same group
Output: A or B, which comes before in the group

Thus, the oder is clearly defined.

TakayoshiKochi · 2016-02-24T08:12:00Z

It looks iterative reviews won't work well if completely different structure is suggested.

I've written the draft following your suggestion:
https://docs.google.com/document/d/1mNV2Z231SrAZuMJvBVNfVNHfULkMwUZOZ75vZM1K9zY/edit?usp=sharing (anyone can comment)

For the 3rd algorithm (ordering) in the comment,
for document/shadow root scope, the order is probably obvious (tree order),
but for slot scope (not fallback content scope) the traversal order is depth-first
search on result of getAssignedNodes()?

In addition, do we want to describe how to compare 2 focusable areas with tabindex here
to get the scoped order of the sequence?

That would add almost dead-copy of HTML 6.4.3 and
super redundant.

TakayoshiKochi · 2016-02-24T08:13:01Z

@hayatoito comment on the doc linked above is welcome.

Get ready for automatic publication - the HTML Imports spec

Fixes WICG#366. Fixes WICG#165. Fixes WICG#151.

Fixes WICG#409.

See WICG#409 (comment).

Somehow my brain hadn't made the switch so every time I worked on new text I kept reintroducing the old name. Oops.

@annevk

Operating under the assumption that this is going to DOM, @annevk's specs don't use semicolons :-S

Fixes WICG#436.

Fixes WICG#435.

- Figure out where custom elements will go in HTML, and started writing as if they were there - Fleshes out and moved around examples and motivation to form an "introduction" section - Separates out and rearranges concepts a bit better (fixes WICG#437) - Moves the registry to the Window instead of the Document (fixes WICG#369) - Moves to `window.customElements`, an instance of `CustomElementsRegistry`, with one `define` method currently (fixes WICG#431)

TakayoshiKochi · 2016-03-16T08:23:13Z

@hayatoito Updated the draft. PTAL.

…ithm

@rniwa

Re-fixes WICG#431. Closes WICG#446. Also updates acknowledgments to be easier to maintain, and adds @rniwa who was missing (in addition to @rianby64 who spotted this and proposed WICG#446).

- markup for <var> - period or colon at the end of bullet - more links to definition - remove title. - fix autolink

As discussed in WICG#375 This adds a clarification about focus navigation order for nodes distributed under a slot element. Also added handling tabindex="-1" case.

Addresses comments from hayato.

- markup for <var> - period or colon at the end of bullet - more links to definition - remove title. - fix autolink

…into focusScope

TakayoshiKochi · 2016-03-18T04:39:12Z

@hayatoito this branch went into some weird state, so I created another PR #447.
Could you take a look?

Proposed change for Focus Navigation

4a34e72

As discussed in WICG#375 This adds a clarification about focus navigation order for nodes distributed under a slot element. Also added handling tabindex="-1" case.

hayatoito reviewed Feb 12, 2016
View reviewed changes

hayatoito mentioned this pull request Feb 17, 2016

Focus Navigation for slots #375

Closed

hayatoito and others added 5 commits February 18, 2016 13:05

[Fix WICG#277] Add loadedmetadata to the list

346a477

[Fix WICG#373] Make event.deepPath() return an empty array if the eve…

d7bf738

…nt is no longer dispatched

[Fix WICG#386] Fix an editorial issue

c9fe8bf

Fix an editorial issue

315e58a

Updated the draft.

ed43c0a

Addresses comments from hayato.

hayatoito reviewed Feb 18, 2016
View reviewed changes

siusin and others added 13 commits February 24, 2016 17:24

Used ReSpec

4dff036

used ReSpec and added a new WD

4a5a2c7

fixed markup errors

5595315

Merge pull request WICG#394 from siusin/gh-pages

06c4018

Get ready for automatic publication - the HTML Imports spec

fixed broken links

b254946

Switched custom spec to respec

07682f6

oops, removed respec debug

5305ade

Use default ref for ecmacript

8982ecc

Switched DOM ref

e2602bf

Improved the respec markup. Moved image closer to the spec.

bab10ed

Manifest for automatic publication

2af62b3

Fixed broken links per WICG#395

716f462

Use https

818b379

domenic and others added 15 commits March 10, 2016 14:08

Typo fix

29ee805

Specify custom elements parser behavior

fbc5787

Fixes WICG#366. Fixes WICG#165. Fixes WICG#151.

Touchup definition of "list of defined custom element local names"

f464a0d

Flesh out defineElement examples

812e53a

Add checks for non-callable lifecycle callbacks at registration time

a5df449

Add conformance requirements for custom element constructors

00ed64f

Fixes WICG#409.

Allow early return in conformance requirements

8b77fbf

Note some subtleties around constructor vs. attachedCallback

b4420c9

See WICG#409 (comment).

attached/detached → connected/disconnected

5ea3a05

Somehow my brain hadn't made the switch so every time I worked on new text I kept reintroducing the old name. Oops.

Remove some semicolons

51fd7cd

Operating under the assumption that this is going to DOM, @annevk's specs don't use semicolons :-S

[Fix WICG#433] Clarify that relatedTarget property is non-null

3139926

Mention that SVG and MathML won't get new hyphen-containing names either

b362fa0

Fixes WICG#436.

Add a reference to XML's EBNF notation

a44034f

Fixes WICG#435.

Fix typo in API summary for customElements.define

fdcd8e9

hayatoito and others added 10 commits March 16, 2016 17:37

[Issue WICG#441] Add a node about ShadowRoot.innerHTML

7fa281b

[Fix WICG#91] Use DocumentOrShadowRoot Mixin

b12afbb

[Issue WICG#91] Clarify elementFromPoint should use retargeting algor…

e04cb38

…ithm

[Fix WICG#443] Clarify innerHTML should work the same way

1148e56

Update custom element examples to use new define function

d0f7fec

Re-fixes WICG#431. Closes WICG#446. Also updates acknowledgments to be easier to maintain, and adds @rniwa who was missing (in addition to @rianby64 who spotted this and proposed WICG#446).

Update from the draft document.

01e5a71

- markup for <var> - period or colon at the end of bullet - more links to definition - remove title. - fix autolink

Proposed change for Focus Navigation

a49def5

As discussed in WICG#375 This adds a clarification about focus navigation order for nodes distributed under a slot element. Also added handling tabindex="-1" case.

Updated the draft.

3adc0f3

Addresses comments from hayato.

Update from the draft document.

8f880f2

- markup for <var> - period or colon at the end of bullet - more links to definition - remove title. - fix autolink

Merge branch 'focusScope' of github.com:TakayoshiKochi/webcomponents …

5a96e3f

…into focusScope

TakayoshiKochi mentioned this pull request Mar 18, 2016

Proposed change for Focus Navigation #447

Merged

TakayoshiKochi closed this Mar 18, 2016

TakayoshiKochi deleted the focusScope branch May 6, 2016 10:48

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Proposed change for Focus Navigation #381

Proposed change for Focus Navigation #381

TakayoshiKochi commented Feb 12, 2016

TakayoshiKochi commented Feb 12, 2016

hayatoito Feb 12, 2016

TakayoshiKochi Feb 15, 2016

TakayoshiKochi commented Feb 18, 2016

hayatoito Feb 18, 2016

TakayoshiKochi Feb 19, 2016

hayatoito commented Feb 18, 2016

TakayoshiKochi commented Feb 24, 2016

TakayoshiKochi commented Feb 24, 2016

TakayoshiKochi commented Mar 16, 2016

TakayoshiKochi commented Mar 18, 2016

Proposed change for Focus Navigation #381

Proposed change for Focus Navigation #381

Conversation

TakayoshiKochi commented Feb 12, 2016

TakayoshiKochi commented Feb 12, 2016

hayatoito Feb 12, 2016

Choose a reason for hiding this comment

TakayoshiKochi Feb 15, 2016

Choose a reason for hiding this comment

TakayoshiKochi commented Feb 18, 2016

hayatoito Feb 18, 2016

Choose a reason for hiding this comment

TakayoshiKochi Feb 19, 2016

Choose a reason for hiding this comment

hayatoito commented Feb 18, 2016

TakayoshiKochi commented Feb 24, 2016

TakayoshiKochi commented Feb 24, 2016

TakayoshiKochi commented Mar 16, 2016

TakayoshiKochi commented Mar 18, 2016