<trackbot> Date: 20 August 2013

<scribe> scribenick: clown

zakim GVoice is Joseph_Scheuhammer

ISSUE-595

ISSUE-595?

<trackbot> ISSUE-595 -- treegrid for UIA has outline roles vs. gridcell roles. IE implemented gridcell - Change it? -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/595

JS: Based on 30 Jul telecon, Cynthia said switch to ROLE_SYSTEM_CELL for treegrid under UIA express
... and DataItem under UIA.

<davidb> DB: for the key press issue… mozilla fix was https://bugzilla.mozilla.org/show_bug.cgi?id=501496

JS: mentioned this change at the weekly call, and wondered if it's really correct.
... then I discovered ISSUE-598 has been raised.

ISSUE-598?

<trackbot> ISSUE-598 -- Mapping for treegrid role is incorrect for IE -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/598

JS: opened this releated issue on the 30th of July.
... it says to map treegrid to control type "tree grid" (not cell).

DB: this needs cynthia.

JS: at the meeting yesterday, Rich suggested things have changed with IE11, and maybe this mapping of treegrid to ROLE_SYSTEM_CELL is no longer applicable.
... just wanted you and mozilla to be aware of this.
... does FF do a UIA implementation?

DB: we are dabbling, but not full bore yet.

ISSUE-596

ISSUE-596?

<trackbot> ISSUE-596 -- Failure of aria-invalid mappings for AT-SPI in test cases: mapping is via text attributes, not object attributes but is not exposed. -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/596

JS: there is a test case with <img role="presentation" aria-invalid="true">
... the aria-invalid is there to check that it counteract the presentation role.
... FF does expose a graphic accessilble in the a11y tree.
... But, because the invalid is given in text attributes (not object attributes), there is nothing exposed because this is a graphic (not text).
... If you look at IA2, it says to expose in text attributes AND to set IA2_STATE_INVALID_ENTRY
... in the above test case, FF does expose IA2_STATE_INVALID_ENTRY for the graphic, but no text attributes.
... there is a STATE_INVALID in AT-SPI.
... My question is: should it be set in this test case?
... here is the test case: https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/presentation-role/504.html

DB: I could look at the code...
... we purposely bypass the object attribute for aria-invalid. But you are curious about state invalid?

JS: yes.

<davidb> http://mxr.mozilla.org/mozilla-central/source/accessible/tests/mochitest/test_aria_token_attrs.html?force=1#72

DB: we do test for ^
... something that is invalid maps to STATE_INVALID.
... you were wondering if the state mapping was accidental. The test I cite shows that it is done on purpose.
... I would expect this to happen for any role we checked.

JS: that means I can change the aria-invalid mappings for AT-SPI to include STATE_INVALID for aria-invalid='true'.
... it comes down to "copying" what is given for IA2.

DB: yes, exactly.
... let me check one more thing.
... there is an ATK_STATE_INVALID_ENTRY.

JS: There is an ATSPI_STATE_INVALID_ENTRY

https://developer.gnome.org/libatspi/unstable/libatspi-atspi-constants.html

<davidb> we want to check that

DB: we want to check that too.

JS: let me run the test case.
... my mistake. accerciser reports "invalid entry"

<davidb> yeah looks like we map invalid to invalid entry for atk "{ ATK_STATE_INVALID_ENTRY, kMapDirectly }, // states::INVALID = 1 << 28"

JS: It's almost identical to IA2.

<davidb> DB: yep

resovled: joseph to update UAIG to state that for aria-invalid='true', for AT-SPI, include a 'set STATE_INVALID_ENTRY'.

resolved: joseph to update UAIG to state that for aria-invalid='true', for AT-SPI, include a 'set STATE_INVALID_ENTRY'.

http://www.w3.org/WAI/PF/aria-implementation/#mapping_state-property_table

ISSUE-607

ISSUE-607?

<trackbot> ISSUE-607 -- Is there a HasKeyboardFocus state in AT-SPI? -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/607

DB: I think that is an error. That the editor put in a 'HasKeyboardFocus' and meant to replace it with the STATE_FOCUSED from AT-SPI.

resolved: change "HasKeyboardFocus" in AT-SPI column to 'STATE_FOCUSED'

http://www.w3.org/WAI/PF/aria-implementation/#focus_state_event_table

ISSUE-604

issue-604?

<trackbot> issue-604 -- Normative requirement on User agent processing of input keys does not belong in UAIG -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/604

<davidb> I think this changed on our source tree late July (https://bugzilla.mozilla.org/show_bug.cgi?id=501496)

http://www.w3.org/WAI/PF/aria-implementation/#keyboard-focus_tabindex

JS: there a list of MUST requirements for tab navigation with tabindex

DB: number 7: When a keydown event is cancelled, also cancel the keypress event
... I think we do this now on our latest but unreleased version of FF.

MC: what is the mechanism? when you preventDefault()? or stopPropagation().

DB: Joseph is doing both.

MC: the cancelbubble — is that necessary?
... preventDefault() for FF and return false for IE.
... but why this requirement?

DB: this doesn't need to be a MUST, but it's nice to have consistency for developers.

MC: why the user agent and not the script?

DB: I think it's a nicety. It's nice to have the consistency to cancel the keypress when the keydown is cancelled.
... it relates to dojo menus.
... some actions are done on keydown, others on keyup.
... what if there is no keydown event, but there is a keypress event.
... there could be a keyup on an element that currently has focus after a keydown event occurred on another element.

MC: that is a very strange situation.
... two questions: (a) is there a defensible requirement? (b) can it be implemented?
... there appears to be a tentative yes to the second question.
... but is it implemented in a wide spread manner.
... it is implemented in the latest FF, and possibly in IE?

DB: I haven't tested in IE. Yes to FF and Chrome.

<davidb> FF nightly is here http://nightly.mozilla.org/

MC: we do not propose to change the requirement in the UAIG.

DB: I'm trying to google the history of this, but I'm not seeing anything.
... you might ping Becky, Joseph. It may have come from dojo land.

http://www.w3.org/TR/2010/WD-wai-aria-implementation-20100916/

Canceling a keydown event must also cancel the keypress event, for purposes of compatibility with other browsers. This is necessary because authors supporting Internet Explorer must use keydown events to process keystrokes, where keydown but not keypress events are fired for non-alphanumeric keys. As web page authors implement script control over the keyboard they need to be able to use keydown but cancel the effect of consumed keystrokes such as arrow keys.

http://www.w3.org/TR/2009/WD-wai-aria-implementation-20090224/

<davidb> that is the reason! ^

JS: do you fully understand it?
... I sort of get it...

<davidb> if you get it, don't give it to me :)

MC: sort of makes sense, but is it still true?
... I found some test pages that let you press various keys, and it shows what each of the various events return.
... sometimes keycodes are returned. sometimes keychar.
... but I'm not sure this is still true today.
... there is a new DOM 3 UI events recent working draft.

<MichaelC> UI Events

JS: Part of the problem is that key events were not standardized in DOM2. Stuck at DOM 1

DB: I am away next week, but I'm probably available on email.

MC: I've got your email with your test file, and work with it. I may send you some questions.

JS: Okay.

This is scribe.perl Revision: 1.138  of Date: 2013-04-25 13:59:11  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/treegird/treegrid/
Found ScribeNick: clown
Inferring Scribes: clown
Default Present: Joseph_Scheuhammer, David_Bolter, Michael_Cooper
Present: Joseph_Scheuhammer David_Bolter Michael_Cooper
Regrets: Cynthia_Shelley
Agenda: http://lists.w3.org/Archives/Public/wai-xtech/2013Aug/0008.html
Found Date: 20 Aug 2013
Guessing minutes URL: http://www.w3.org/2013/08/20-aapi-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.