363617 - Missing support for some ARIA properties

Status: RESOLVED FIXED

(Core :: Disability Access APIs, defect)

Description

[Aaron Leventhal]

The following ARIA properties (from http://www.w3.org/WAI/PF/Group/adaptable/) are not yet supported:

hasparent -- IDREF (relation)
live -- string
atomic -- boolean
relevant -- string
owns -- IDREFS (relations)
level -- numeric
datatype -- string (URL?)
posinset -- numeric
setsize -- numeric
sort -- string

Comment 1

[Aaron Leventhal]

Attachment: Support the 8 new ARIA properties by directly mapping them to AT-SPI object attributes

Comment 2

[Aaron Leventhal]

Spunoff bug 363628 for support for the 2 relations: hasparent and owns.

Comment 3

[Aaron Leventhal]

Comment on attachment 248426 [details] [diff] [review]
Support the 8 new ARIA properties by directly mapping them to AT-SPI object attributes

Needs more tweaks so that level, posinset and setsize map to our MSAA positional description hack.

Comment 4

[Aaron Leventhal]

Attachment: Support the new ARIA properties both in object attributes and in old-style MSAA positional description hack

Comment 5

[[e-mail]]

Comment on attachment 248438 [details] [diff] [review]
Support the new ARIA properties both in object attributes and in old-style MSAA positional description hack

>+    char *ariaProperties[] = { "live", "atomic", "relevant", "datatype", "level", "posinset", "setsize", "sort" };
Couldn't this be made static const?

Sorry, but I'd rather you found an alternate reviewer, if you don't mind.

Comment 6

[Aaron Leventhal]

Comment on attachment 248438 [details] [diff] [review]
Support the new ARIA properties both in object attributes and in old-style MSAA positional description hack

Thanks for the sr=, I'll make it static const char*

Comment 7

[Ginn Chen]

Comment on attachment 248438 [details] [diff] [review]
Support the new ARIA properties both in object attributes and in old-style MSAA positional description hack

I'm confused with posinset description on http://www.w3.org/WAI/PF/Group/adaptable/

It says "The posinset property is used to indicate an item's number or position in the current level. For example, if this element is the third item in a group then posinset is equal to three (integer: 0<=posinset<=n-1.)"
It's not clear whether posinset is zero-based.

If it's zero-based, maybe it's not good to say
"L%d, %d of %d", intLevel, intPosInSet, intSetSize

"L1, 0 of 3." or "L1, 1 of 3." may confuse people.


Also, what about changing 
intLevel >= 0 to intLevel >= 1 ,
intSetSize >= 0 to intSetSize >= 1 ?

Comment 8

[Aaron Leventhal]

Yes, I noticed that too. I'm planning to send in feedback. When we get that cleared up I can make another patch.

Comment 9

[Aaron Leventhal]

Actually I just checked with the WG and they're going to make all 3 of these properties (level, posinset, setsize) >= 1.

Comment 10

[Aaron Leventhal]

I've updated the docs to describe the new object attributes:
http://developer.mozilla.org/en/docs/Accessibility/ATSPI_Support#Supported_AT-SPI_Object_Attributes

Comment 11

[Ginn Chen]

(In reply to comment #9)
> Actually I just checked with the WG and they're going to make all 3 of these
> properties (level, posinset, setsize) >= 1.
> 

That's cool!